ConceptComplete

Mathematical Induction

Mathematical induction is the fundamental proof technique for statements about natural numbers. In set theory, induction is not an axiom but a theorem derivable from the structure of Ο‰\omega and the axiom of foundation.

TheoremPrinciple of Mathematical Induction

Let P(n)P(n) be a property defined for natural numbers. If:

  1. Base case: P(0)P(0) holds
  2. Inductive step: For all nβˆˆΟ‰n \in \omega, if P(n)P(n) holds then P(n+1)P(n+1) holds

Then P(n)P(n) holds for all nβˆˆΟ‰n \in \omega.

Proof

Suppose P(0)P(0) and βˆ€n(P(n)β‡’P(n+1))\forall n(P(n) \Rightarrow P(n+1)). Let S={nβˆˆΟ‰:P(n)}S = \{n \in \omega : P(n)\}. We must show S=Ο‰S = \omega.

By hypothesis, 0∈S0 \in S (base case). Also, if n∈Sn \in S then P(n)P(n) holds, so P(n+1)P(n+1) holds by the inductive step, hence n+1∈Sn+1 \in S. Thus SS is an inductive set.

Since Ο‰\omega is the smallest inductive set, and SS is an inductive set with SβŠ†Ο‰S \subseteq \omega, we conclude Ο‰βŠ†S\omega \subseteq S. Combined with SβŠ†Ο‰S \subseteq \omega, we have S=Ο‰S = \omega.

β– 
ExampleSum of First n Natural Numbers

Prove that 0+1+2+β‹―+n=n(n+1)20 + 1 + 2 + \cdots + n = \frac{n(n+1)}{2} for all nβˆˆΟ‰n \in \omega.

Base case: n=0n = 0. The sum is 0=0β‹…120 = \frac{0 \cdot 1}{2}. βœ“

Inductive step: Assume 0+1+β‹―+n=n(n+1)20 + 1 + \cdots + n = \frac{n(n+1)}{2}. Then:

0+1+β‹―+n+(n+1)=n(n+1)2+(n+1)=n(n+1)+2(n+1)2=(n+1)(n+2)2\begin{align*} 0 + 1 + \cdots + n + (n+1) &= \frac{n(n+1)}{2} + (n+1) \\ &= \frac{n(n+1) + 2(n+1)}{2} \\ &= \frac{(n+1)(n+2)}{2} \end{align*}

By induction, the formula holds for all nβˆˆΟ‰n \in \omega.

DefinitionStrong Induction

An equivalent form of induction, often called strong induction or complete induction, states:

If for all nβˆˆΟ‰n \in \omega,

[βˆ€m<n P(m)]β‡’P(n)[\forall m < n \, P(m)] \Rightarrow P(n)

then P(n)P(n) holds for all nβˆˆΟ‰n \in \omega.

Here m<nm < n means m∈nm \in n in set-theoretic terms.

Strong induction does not require a separate base case because the implication vacuously holds for n=0n = 0 (there are no m<0m < 0).

TheoremEquivalence of Induction Forms

The principle of mathematical induction and strong induction are equivalent. Each can be derived from the other.

Proof

(Strong β‡’\Rightarrow Weak): Assume strong induction. To prove weak induction, suppose P(0)P(0) and βˆ€n(P(n)β‡’P(n+1))\forall n(P(n) \Rightarrow P(n+1)). For any nn, if βˆ€m<n P(m)\forall m < n \, P(m), then either n=0n = 0 (so P(0)P(0) by hypothesis) or n=k+1n = k+1 for some kk, and P(k)P(k) holds (since k<nk < n), so P(n)=P(k+1)P(n) = P(k+1) by the inductive step. By strong induction, P(n)P(n) for all nn.

(Weak β‡’\Rightarrow Strong): Assume weak induction. Define Q(n)=βˆ€m≀n P(m)Q(n) = \forall m \leq n \, P(m). Then Q(0)⇔P(0)Q(0) \Leftrightarrow P(0). If Q(n)Q(n), then all P(m)P(m) for m≀nm \leq n hold. By the strong induction hypothesis, P(n+1)P(n+1) holds, so Q(n+1)Q(n+1). By weak induction, Q(n)Q(n) for all nn, hence P(n)P(n) for all nn.

β– 
DefinitionRecursion Theorem

The recursion theorem states that for any set AA, any element a∈Aa \in A, and any function h:Aβ†’Ah: A \to A, there exists a unique function f:Ο‰β†’Af: \omega \to A such that:

f(0)=af(n+1)=h(f(n))Β forΒ allΒ nβˆˆΟ‰\begin{align*} f(0) &= a \\ f(n+1) &= h(f(n)) \text{ for all } n \in \omega \end{align*}

This theorem justifies recursive definitions and is proved using transfinite recursion on ordinals.

ExampleFibonacci Numbers

The Fibonacci sequence can be defined by two-place recursion:

F0=0,F1=1Fn+2=Fn+1+FnΒ forΒ nβ‰₯0\begin{align*} F_0 &= 0, \quad F_1 = 1 \\ F_{n+2} &= F_{n+1} + F_n \text{ for } n \geq 0 \end{align*}

To formalize this, we use the recursion theorem with state space NΓ—N\mathbb{N} \times \mathbb{N} and the transition function h(a,b)=(b,a+b)h(a, b) = (b, a+b). Starting from (0,1)(0, 1), this generates pairs (Fn,Fn+1)(F_n, F_{n+1}).

TheoremWell-Ordering Principle

Every non-empty subset of Ο‰\omega has a least element under the natural ordering. This is equivalent to mathematical induction.

Proof

(WO β‡’\Rightarrow Induction): Suppose induction fails for some property PP. Let S={nβˆˆΟ‰:Β¬P(n)}S = \{n \in \omega : \neg P(n)\} be the set of counterexamples. If Sβ‰ βˆ…S \neq \emptyset, it has a least element mm. Since P(0)P(0), we have m>0m > 0, so m=k+1m = k+1 for some kk. Since mm is minimal in SS, we have P(k)P(k). By the inductive step, P(k+1)=P(m)P(k+1) = P(m), contradicting m∈Sm \in S.

(Induction β‡’\Rightarrow WO): Suppose SβŠ†Ο‰S \subseteq \omega has no least element. Define P(n)=βˆ€m≀n(mβˆ‰S)P(n) = \forall m \leq n (m \notin S). Then P(0)P(0) (else 00 would be the least element). If P(n)P(n), then all m≀nm \leq n satisfy mβˆ‰Sm \notin S. If n+1∈Sn+1 \in S, it would be minimal (since no smaller element is in SS), contradiction. So P(n+1)P(n+1). By induction, P(n)P(n) for all nn, so S=βˆ…S = \emptyset.

β– 
Remark

The equivalence of induction, strong induction, well-ordering, and recursion demonstrates that these are different perspectives on the same fundamental property of Ο‰\omega: its well-founded structure. This structure, ultimately derived from the axiom of foundation, is what makes natural numbers suitable for proof by induction and definition by recursion.

Mathematical induction is not merely a proof technique but a deep structural property of the natural numbers, reflecting their construction as the minimal inductive set in ZFC set theory.