After spending a few hours trying to understand Theorem $1.11$ in Rudin's Principles of Mathematical Analysis, I still don't follow the proof.

$1.11$ Theorem Suppose $S$ is an ordered set with the least-upper-bound property, $B \subset S$, $B$ is not empty, and $B$ is bounded below. Let $L$ be the set of all lower bounds of $B$. Then $\alpha = \sup L$ exists in $S$, and $\alpha = \inf B$.

This is what I understand so far:

$B$ is bounded below means that $L$ is not empty and $L = \{ y \ | \ y \leq x \ \forall x \in B \}$. Then every $x \in B$ is an upper bound of $L$, which means that $L$ is bounded above. Since $L \subset S$, $L$ not empty, and $L$ is bounded above that implies that $\sup L = \alpha \in S$. And because $\alpha = \sup L$, $\gamma < \alpha$ implies that $\gamma$ is not an upper bound of $L$ and $\gamma \notin B$ since every element of $B$ is an upper bound of $L$.

This is where I get confused:

Since $B$ is bounded below, there exists an $\omega \in S$ such that $\omega \leq x \ \forall x \in B$. Then Rudin claims, "It follows that $\alpha \leq x$ for every $x \in B$." Can someone explain why that is true or at least give me a hint?

Thanks in advance.

  • 1
    $\begingroup$ Apologies, but I don't understand why L must be a subset of S. As a counter example, what if B = S, wouldn't some elements of L have to be outside of S? Many thanks! $\endgroup$ – user68045 Mar 22 '13 at 18:17
  • $\begingroup$ @user68045 isn't around and commented over 4 years ago, but: Everything is in $S$ in the context of the theorem. There is nothing indicating some larger ordered set of which $S$ is a subset. If $B=S$ and $B$ is bounded below, then $L$ would consist of the single lower bound which must also be the mimimum element of S. $\endgroup$ – Jonas Meyer Sep 5 '17 at 3:38

Rewording it slightly: Every element of $B$ is an upper bound for $L$, so if $x\in B$ is less than $\alpha$, then $x$ is an upper bound for $L$ smaller than the least upper bound. This contradicts the definition of least upper bound, so no such $x$ exist. In other words, $\alpha\leq x$ for every $x\in B$.

I do not know why you included a restatement of the fact that $B$ is bounded below. The part you put in quotes follows directly from the last sentence of the previous paragraph in your post:

And because $\alpha = \sup L$, $\gamma < \alpha$ implies that $\gamma$ is not an upper bound of $L$ and $\gamma \notin B$ since every element of $B$ is an upper bound of $L$.

From here you could use contraposition:

$$\gamma<\alpha\implies \gamma\not\in B$$

is equivalent to

$$\gamma\in B \implies \gamma\geq \alpha.$$

  • $\begingroup$ Your first paragraph cleared up my confusion. Thanks! $\endgroup$ – Student Jan 27 '12 at 23:04
  • $\begingroup$ I'm a bit confused as to what you are trying to show with the contraposition. Are you trying to show that $\alpha\leq x$ for every $x \in B$? If so how does this achieve that? $\endgroup$ – YellowPillow Nov 18 '17 at 1:21
  • $\begingroup$ @YellowPillow: OP wanted to conclude that $\alpha \leq x$ for all $x\in B$, or equivalently that $x\in B\implies \alpha \leq x$, which is equivalent by contraposition to $\alpha > x\implies x\not\in B$. The latter is what the OP had argued. The way I stated the contraposition aligned with their notation. $\endgroup$ – Jonas Meyer Oct 2 '18 at 20:04

I think that Rudin has complicated the proof a little bit. Here is what I learned from his proof to write my own. I think this is more straight forward.

Let $L$ be the set of all lower bounds of $B$ or more precisely $L=\{l\,|\,\forall x:x\in B\implies x\ge l\}$. Since $B$ is bounded below so $L$ is not empty. Also, for every $x$ in $B$ and every $l$ in $L$ we have $l\le x$. This means that every $x\in B$ is an upper bound for $L$. By the L.U.B. property of $S$, we conclude that $L$ has a supremum denoted by $\sup L$. As $\sup L$ is smaller than any other upper bound of $L$ we conclude that $\sup L \le x$ for every $x\in B$. But this means that $\sup L$ is also a lower bound of $B$ and hence $\sup L \in L$. Furthermore, if $l\in L$ is any lower bound of $B$ then $l \le \sup L$. So far, we proved $\sup L$ is a lower bound of $B$ and is greater than any other lower bound of $B$ so it is the greatest lower bound of $B$, that is $\inf B =\sup L$. Accordingly, the set $S$ has the G.L.B. property.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.