4
$\begingroup$

In the appendix on Dedekind cuts in Rudin's book, he defines for each $r \in \mathbb{Q}$ the Dedekind cut $r^* = \{q \in \mathbb{Q} \mid q < r\}$. He then asserts that $$r^* s^* = (rs)^*.$$ Rudin omits the proof and states that it is similar for the proof that $(r+s)^* = r^* + s^*$. I do not understand how this is the case, in part because I need to consider three cases: both positive, positive negative, and one positive and one negative.

Suppose both are positive, either $r^*, s^* > 0*$. Then $r^* s^* = \{ab \mid a \in r^*, b \in s^*, ab > 0\}$. To show inclusion in the right-hand side, I need to show that given an $ab \in r^* s^*$, I have $ab < rs$. Since $a \in r^*$, $a < r$. Since $b \in s^*$, $b < s$. So $ab < rs$, giving the first inclusion.

I'm having trouble with the second inclusion. If $q \in (rs)^*$, then $q < rs$. Attempting to follow the proof for addition, I can then say that $rs - q > 0$ and define $rs - q = t$ .Then I want to construct, with $t$, two positive rational numbers whose product is $q$, the first term of which is in $r^*$, the second of which is in $s^*$. I can't figure out how to do this. Subtracting $t$, as in the addition proof, doesn't work, and multiplying it gives an extraneous $t^2$ factor.

I haven't had much success with the remaining cases, but I think I could make more progress if I could get this first one. Any help would be appreciated.

$\endgroup$
4
  • $\begingroup$ If $q \lt rs$, then there is some rational number between $q$ and $rs$. For sufficiently small rational $\varepsilon, q \lt (r-\varepsilon)(s-\varepsilon)$. $\endgroup$ Commented Feb 9, 2021 at 4:50
  • $\begingroup$ I agree that I can find say a $t$ such that $q < t <rs$. If $\epsilon$ is positive, then $r - \epsilon < r$ and $s - \epsilon < s$, so provided that $r - \epsilon, s - \epsilon > 0$, I get that $q < rs < (r - \epsilon)(s - \epsilon)$. Then I assume the argument is that $r - \epsilon \in r^*$ and $s - \epsilon \in s^*$, so $(r - \epsilon)(s - \epsilon) \in (rs)^*$, so by left-closure, $q \in (rs)^*$. This is very helpful. Where, though, does finding a rational between $q$ and $rs$ come into play, though? That's the only fact I haven't used here. $\endgroup$
    – JeremyS
    Commented Feb 9, 2021 at 5:00
  • $\begingroup$ I don't understand how you proved the first inclusion, which seems false to me. $a<r,b<s$ does not imply $ab<rs$ -- consider the case when both $a,b$ are large negative rationals. For example, $r=s=2,a=b=-3$. $\endgroup$ Commented Feb 9, 2021 at 5:27
  • $\begingroup$ @JeremyS You may not need that there's a rational between $q$ and $rs$. I'm just riffing off the top of my head. $\endgroup$ Commented Feb 9, 2021 at 5:30

1 Answer 1

2
$\begingroup$

Here is a simple strategy to find rational numbers $a, b$ with $0<a<r,0<b<s$ such that $ab=q$ where $q$ is a given number with $0<q<rs$. Just choose any rational $a$ which lies between $q/s$ and $r$ and then but $b=q/a$ and you are done.


Dedekind's construction of real numbers is the simplest approach possible. It is really a remarkable feat of Rudin to turn it into something seemingly complex.

$\endgroup$

You must log in to answer this question.

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