Proving that for a Dedekind cut $\alpha$ and positive rational $w$, $\exists{n\in\mathbb{Z}}$ such that $nw\in{\alpha}$ but $(n+1)w\notin{\alpha}$

565 Views Asked by At

On page 19 of Principles of Mathematical Analysis by Walter Rudin ("Baby Rudin"), Rudin proves the field axioms of addition for Dedekind cuts ("Step 4" of the Ch.1 Appendix; the entire text can be found here).

In the process of proving that 0*$\subseteq{\alpha+\beta}$, Rudin states that for a positive $w\in{\mathbb{Q}}$, "there is an integer $n$ such that $nw\in{\alpha}$ but $(n+1)w\notin{\alpha}$," and adds "Note that this depends on the fact that $\mathbb{Q}$ has the archimedean property!"

I see that the Archimedean Property implies that $\exists{n_1\in\mathbb{Z}}$ such that $n_1w\in\alpha$, since we can choose $p\in\alpha$ and $n_1\in\mathbb{Z}$ such that $n_1w<{p}$. Similarly, given a $q\notin\alpha$, we can pick $n_2\in\mathbb{Z}$ such that $n_2w>q$, which implies $n_2w\notin\alpha$. But what makes Rudin so sure we can pick $n_1$ and $n_2$ such that $n_1+1=n_2$?

The way I would think to do this (were this not in the middle of a proof about how the Dedekind cuts are a field satisfying the Least Upper Bound Property) would be to use the Least Upper Bound Property of $\mathbb{R}$, of which $\mathbb{Q}$ is a subset. The set $S=\{nw\in\alpha\mid{n\in\mathbb{Z}}\}$ is unempty by the Arch. Prop. and bounded above by, for example, $q$ from the previous paragraph. Therefore, but the L.U.B. Property, $A=\sup{S}$ exists. Letting $\epsilon=\frac{w}{2}$, there is at most one element $s\in{S}$ such that $A-\epsilon<s\leq{A}$. Such an $s$ must exist by the Approximation Property of Sup. If we had $s\neq{A}$, we could let $\epsilon'=|A-s|$. Then the only possible value $s'$ such that $A-\epsilon'<s'\leq{A}$ is A itself, so by the Approximation Property of Sup, we must have that $A\in{S}$. In that case, $A=nw\in\alpha$ for some $n\in\mathbb{Z}$, but $(n+1)w\notin\alpha$. So the claim can be proved this way.

But can we prove it without using the L.U.B. Property of $\mathbb{R}$? Seems a little strange to do so when the point of this whole section of the book is to construct the field $\mathbb{R}$.