1

This seems undeniably true to me, but I don't know how to write it down.

Given the non-empty set $S$ containing only natural numbers (starting at 1, not 0). If for every number $x$ greater than 1 there is a number $y$ such that $y<x$, then 1 must be in the set $S$.

I think this must be true, because you can start at any natural number, and take a smaller number, and a smaller number than that, smaller than that and so forth, until you arrive at 1. But how do you actually write this proof down?

Lara
  • 301

3 Answers3

3

Using the well ordering principle of $\mathbb N$, as $S$ is nonempty, there is $y\in S$ so that

$$y\le x\ \ \ \ \ \forall x\in S.$$

If $y\neq 1$, then by the definition of $S$, there is $y_0\in S$ so that $y_0 <y$. But that is impossible by the choice of $y$. Thus $y=1$ and so $1\in S$.

1

I prefer John Ma’s argument, but if you’re more familiar with the induction axiom than with the well-ordering principle, you may prefer this one:

Suppose that $1\notin S$, and let $A=\big\{n\in\Bbb Z^+:S\cap\{1,\ldots,n\}=\varnothing\big\}$; clearly $1\in A$. Suppose that $n\in A$; then $\{1,\ldots,n\}\cap S=\varnothing$, so $n+1\notin S$. But then $\{1,\ldots,n+1\}\cap S=\varnothing$, so $n+1\in A$. Thus, $1\in A$, and for every $n\in\Bbb Z^+$ we have $n\in A\implies n+1\in A$, so by the induction axiom $A=\Bbb Z^+$. But then $S=\Bbb Z^+\cap S=\varnothing$, contradicting the hypothesis that $S$ is non-empty.

Brian M. Scott
  • 616,228
0

$$\begin{align}\text{Let }\mathbb{N}&=\{x\geq1\mid x\in\mathbb{Z}\},\tag{1}\label{defN}\\S&\neq\emptyset\tag{2}\label{defS}\end{align}$$ then we are asked to show that $$\forall_S:\;(S\subseteq\mathbb{N})\land\forall_{x\in{S}}((x>1)\implies\exists_{y\in{S}}:y<x)\implies1\in{S},$$ which is equivalent to $$\forall_S:\;1\notin{S}\implies\lnot\Big((S\subseteq\mathbb{N})\land\forall_{x\in{S}}((x>1)\implies\exists_{y\in{S}}:y<x)\Big)$$ by contraposition and hence to $$\forall_S:\;1\notin{S}\implies(S\not\subseteq\mathbb{N})\lor\lnot\forall_{x\in{S}}((x>1)\implies\exists_{y\in{S}}:y<x).$$ The case of $S\not\subseteq\mathbb{N}$ is uninteresting, so setting $S\subseteq\mathbb{N}$, and by negation of universal quantification we have: $$\forall_{S\subseteq\mathbb{N}}:\;1\notin{S}\implies\exists_{x\in{S}}\,\lnot((x>1)\implies\exists_{y\in{S}}:y<x),$$ which, negating the material implication, yields: $$\forall_{S\subseteq\mathbb{N}}:\;1\notin{S}\implies\exists_{x\in{S}}\,((x>1)\land\lnot\exists_{y\in{S}}:y<x),$$ and by negation of existential quantification we have: $$\forall_{S\subseteq\mathbb{N}}:\;1\notin{S}\implies\exists_{x\in{S}}\,((x>1)\land\forall_{y\in{S}}:y\geq{x}).$$ Indeed, since $1\notin{S}$, we have $x>1$ by the definition of $\mathbb{N}$ in $\eqref{defN}$, so we have: $$\forall_{S\subseteq\mathbb{N}}:\;1\notin{S}\implies\exists_{x\in{S}}\,\forall_{y\in{S}}:y\geq{x},$$ and, as the consequent $\exists_{x\in{S}}\,\forall_{y\in{S}}:y\geq{x}$ (no matter if $1\in{S}$ or $1\notin{S}$) is precisely what is claimed by the well-ordering principle linked by @John, we conclude the implication is true for all subsets $S$ of the set of natural numbers $\mathbb{N}$.

dbanet
  • 1,413