In an ω-model, every form of induction holds, because the first-order part is the true . In a model with non-standard numbers, induction may fail for complicated formulas: a formula may define a proper initial segment closed under successor, a cut, which contains but not every number. The axioms of only grant induction for formulas, and many principles of ordinary mathematics turn out to be equivalent to forms of induction or of the related collection principle. This lesson sets up the hierarchy of these schemes, following Paris and Kirby.
1.1 The schemes
All formulas are formulas of , with number and set parameters (Definition 1.1 of “The language and models of second-order arithmetic”); the classes and count blocks of quantifiers of the same kind as one.
Let and be or . For every formula , the universal closures of the following formulas form respectively the schemes , and :
- induction: ;
- collection: ;
- least number principle: .
The theory consists of the basic axioms (Definition 1.2 of “The language and models of second-order arithmetic”) and the scheme of induction for bounded formulas.
The formula may contain other free variables, which act as parameters. Collection says that a “finite” family of nonempty sets of witnesses has a common bound; here finite means indexed by the numbers below some , which may be non-standard. In the true , every instance of the three schemes holds. All theories in this course extend , and contains .
proves the elementary properties of addition, multiplication and order: is a discrete linear order with least element , every is for a unique , if and only if for some , which is then unique, polynomials with non-negative coefficients are monotone, and the pairing function is a bijection, definable by a bounded formula, with .
These properties are proved by bounded induction, in the same way as in Peano arithmetic; see Kaye, Models of Peano Arithmetic, Chapter 2. We use them freely.
A formula obtained from formulas by conjunction, disjunction, bounded quantification and existential quantification is not in general, because a bounded universal quantifier cannot be moved across an unbounded existential one. In the true , is equivalent to ; this equivalence is exactly an instance of collection.
Let . Over :
- every formula obtained from formulas by , , bounded quantifiers , and existential quantifiers is provably equivalent to a formula;
- every formula obtained from formulas by , , bounded quantifiers and universal quantifiers is provably equivalent to a formula.
Item 2 follows from item 1 by negation, since the negation of a formula is logically equivalent to a formula. It suffices to treat one operation at a time. By renaming bound variables, is equivalent to , and similarly for ; is equivalent to ; and an existential quantifier merges with the first block. We may also assume that the first block of a formula consists of a single quantifier: is equivalent to , and , with , is bounded if and equivalent to a formula by item 2 for if , in the induction below. The only non-trivial case is the bounded universal quantifier. We proceed by induction on .
For , let with bounded. Then implies by , and the converse is logical. The second formula is .
Let , assume the lemma for , and assume , which contains . A formula has the form with . Conjunctions and disjunctions of formulas, and bounded quantifications of them, are equivalent to formulas by item 2 for ; so the easy cases above stay within . For the bounded universal quantifier, is equivalent to by applied to , which is . By item 2 for , is equivalent to a formula, so the whole formula is equivalent to a formula.
1.2 Σ⁰ₙ and Π⁰ₙ induction
For every , proves that and are equivalent.
() Let be with and , and suppose that . Let
The formula is equivalent to a formula, and the conjunction with the bounded formula can be moved inside its innermost block; so is equivalent to a formula. We have , with . Suppose , and let . Then , and there is with and ; , since , so . Since would imply , we have , and : this is . By , holds, which gives with , so , and : a contradiction.
() Let be with and , and suppose . Let
equivalent to a formula. Then holds since . Suppose and , and let . Then , so , hence . So . By , holds, and with it gives : a contradiction.
The proof reverses the direction of the induction: the property ” fails” is propagated from up to . Both formulas and have the same complexity as the negation of , which is why the trick exchanges and .
1.3 Induction implies collection
For every , proves .
By induction on . Assume , and, if , the theorem for , so that holds and Lemma 1.3 applies to formulas.
Let with (bounded if ), and suppose that . Let
It is equivalent to a formula: for it is bounded, and for by Lemma 1.3. Moreover : take larger than a witness and a witness . Let
The formula is equivalent to a formula, by the same argument, so is equivalent to a formula. Then holds with . Suppose and ; let be given by , and with , which exists since . Then witnesses . By , holds: there is such that for every some satisfies , hence some satisfies .
1.4 Least number principles
For every , the schemes , , and are equivalent over .
() Let with and , and suppose . By applied to the formula , there is a least with . Then , so with , hence : a contradiction. The same argument gives .
() Let with , and suppose there is no least with . Let . By Theorem 1.5, holds, so is equivalent to a formula by Lemma 1.3. If , then is least; so . If and , then is least; so implies . By induction, holds, contradicting .
() The same argument, with : the formula is equivalent to a formula, since implies by Theorem 1.4, hence .
Combined with Theorem 1.4, the four schemes are equivalent.
The least number principle gives the model-theoretic meaning of induction. If fails, some formula holds at , is preserved by successor, and fails somewhere; the set of numbers below every failure is then a cut, and its complement has no least element. In a model of , no or formula defines a cut. For instance, in a non-standard model of , the standard numbers form a cut, so they are not definable by such a formula: if a or formula holds for every standard number, it holds for some non-standard number. This is the overspill principle.
1.5 Collection implies induction
For every , proves .
By induction on . Assume ; it contains and . When , the induction hypothesis (the theorem for , applied to ) gives , hence also by Theorem 1.4.
Let with , suppose and , and fix . For every there is with . This formula is equivalent to a formula, hence is , and gives a bound . Let be such that , and . Then for every ,
Indeed, if , the first disjunct fails, so for some . Let
For , is bounded; for , it is equivalent to a formula by Lemma 1.3. We have , since , and by the display. By or , holds, so .
The idea is that collection bounds, uniformly for , the witnesses needed to pass from to . Once the witnesses are bounded, the induction becomes an induction on a formula of lower complexity.
We obtain the hierarchy, for every , over :
Paris and Kirby showed that none of these implications reverses: for every , there are models of in which fails, and models of in which fails. The constructions use non-standard models of arithmetic and are admitted here (Hájek and Pudlák, Chapter IV). Slaman proved that is equivalent to induction for formulas, over with the totality of exponentiation.
1.6 Exercises
Show that, over , is equivalent to for every .
Solution
Every formula is , so implies . Conversely, assume , and let with , such that . Using the pairing function of Fact 1.2, let
Since every is the pair of unique , holds if and only if for the components of . Since exactly one pair satisfies , the prefix amounts to substituting the components of , and commutes with every quantifier of : moving it down to the bounded matrix of , is equivalent to a formula, without any use of collection. Now , so gives with . If , then and . So .
The strong induction scheme for a formula is . Show that proves strong induction for formulas, for every .
Solution
Let satisfy the hypothesis, and fix . Let . By Theorem 1.5 and Lemma 1.3, is equivalent to a formula. We have . If and , then , so by hypothesis, and : this is . By , holds, so , and by hypothesis.
1.7 Notes and references
The hierarchy of induction and collection schemes is due to Paris and Kirby (1978), building on Parsons (1970), who showed that proves the same sentences as primitive recursive arithmetic. Hájek and Pudlák’s monograph is the standard reference; Kaye’s book gives a model-theoretic introduction.
- J. B. Paris and L. A. S. Kirby, Σₙ-collection schemas in arithmetic, in Logic Colloquium ‘77, North-Holland, 1978, 199–209.
- C. Parsons, On a number theoretic choice schema and its relation to induction, in Intuitionism and Proof Theory, North-Holland, 1970, 459–473.
- P. Hájek and P. Pudlák, Metamathematics of First-Order Arithmetic, Springer, 1993, Chapter I.2 and Chapter IV.
- R. Kaye, Models of Peano Arithmetic, Oxford University Press, 1991, Chapters 2 and 10.
- T. A. Slaman, Σₙ-bounding and Δₙ-induction, Proceedings of the American Mathematical Society 132 (2004), 2449–2456.