1
$\begingroup$

Is there some systematic way to reformulate mathematical induction to work to prove propositions on general subsets of the natural numbers?

Own work: I'm thinking one could maybe use modulo arithmetic or divisibility to encode intervals or periodicity:

  1. Prove something true for every number on an interval $i\in [a,b] \subset \mathbb N$, write $i = a+(b-a)k$ and then let $k = 0$ be part of the induction assumption.
  2. Prove something true periodically by $i = a (\mod b)$ or tie the set of numbers to cyclic groups.
$\endgroup$

2 Answers 2

1
$\begingroup$

Let $S\subset\mathbb{Z}$ and let $S$ have a first element $n_0$.

For each $n\in S$ let $n^\prime$ denote the next element of $S$.

Then if the proposition $P(n_0)$ is true and if for $n\in S$, $P(n)$ implies $P(n^\prime)$ then $P(k)$ is true for all $k\in S$.

$\endgroup$
1
$\begingroup$

It's easy using the notion of "next smallest." The key to mathematical induction is that for every natural number the is a well-defined "next" number. For subsets of the natural numbers this is the case as well, although in general it isn't as simple as adding $1$. However, given a set $S\subseteq\mathbb{N}$, if you prove that $P(a)$ holds and you prove that $P(x)\Rightarrow P(n(x))$ holds, where $a$ is the smallest element of $S$ and $n:S\to S$ gives the next smallest element, then you've proven that $P$ holds on all of $S$.

This is related to proof by infinite descent.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .