2
$\begingroup$

Exercise 3.2.5 in Stephen Abbott's Understanding Analysis asks to prove the below theorem. I would like to ask, if my proof is sound. $\newcommand{\absval}[1]{\left\lvert #1 \right\rvert}$

Theorem. A set $F \subseteq \mathbf{R}$ is closed if and only if every Cauchy sequence contained in $F$ has a limit that is also an element of $F$.

My Attempt.

($\Longrightarrow$) Assume $F \subseteq \mathbf{R}$ is closed. By definition, a set is closed, if and only if it contains all its limit points. Let $x \in F$ be an arbitrary limit point of $x$. Thus, $V_\epsilon(x)$ intersects $F$ in some point other than $x$. To produce a Cauchy sequence in $F$, we let $\epsilon = 1/n$. Then, there exists a point $x_n \in F$, where \begin{align*} x_n \in V_\epsilon(x) \cap F \end{align*}

with the stipulation that $x_n \ne x$.

It is easy to see, that $(x_n) \to x$. To see this, choose $N > 1/\epsilon$. Then, for all $n \ge N$, we have, \begin{align*} \absval{x_n - x} < \epsilon \end{align*}

Convergent sequences are Cauchy and Cauchy sequences are convergent. Convergent Sequence $\Longleftrightarrow$ Cauchy sequence.

Since, $F$ contains all its limit points, all Cauchy sequences in $F$ have their limiting value in $F$.

($\Longleftarrow$) Assume that every Cauchy sequence in $F$ has a limit that is also an element of $F$. Therefore, $\lim x_n = x$, $x_n \ne x$. By the definition of convergence, given any $\epsilon > 0$m there exists a term $x_N$ in the sequence satisfying $\absval{x_N - x} < \epsilon$. So, $V_\epsilon(x) \cap F$ contains elements other than $x$.

$\endgroup$
11
  • $\begingroup$ P.S. I have cheated by refreshing my memory with an existing proof given in the text for the statement: A point $a$ is a limit point of the set $A$, if and only if, there exists a sequence $(a_n)$ in $A$, such that $\lim a_n = a$, with $a_n \ne a$ for all $n$. Aside: Is cheating allowed? I could not remember, how to produce a sequence $(a_n)$ in $A$. $\endgroup$ – Quasar Jan 6 at 7:28
  • 1
    $\begingroup$ The first half of the proof goes wrong at the very beginning, when you set out to construct a Cauchy sequence in $F$ with a certain property. What you have to prove is that every Cauchy sequence in $F$ converges to some point of $F$. Thus, you should either (a) start with an arbitrary Cauchy sequence in $F$ and show that it converges to a point of $F$ or (b) assume that that there is a Cauchy sequence in $F$ that does not converge to a point of $F$ and show that $F$ is not compact. $\endgroup$ – Brian M. Scott Jan 6 at 7:32
  • 1
    $\begingroup$ You’ve made a similar error in the second half. You want to show that $F$ is closed, so you should start with a limit point of $F$ and use your hypothesis to show that the point is in $F$. What you did in the first part is useful here. $\endgroup$ – Brian M. Scott Jan 6 at 7:33
  • 1
    $\begingroup$ $p\implies q$ is almost right, but you’ve overlooked a possibility: the Cauchy sequence might be eventually constant. In that case a sufficiently small $V_\epsilon(x)$ won’t necessarily contain a point of $F$ different from $x$, and $x$ need not be a limit point of $F$. (Example: $F=[0,1]\cup\{2\}$, and $x_n=2$ for all $n$.) But that can happen only when $x_n=x$ for all but finitely many $n$, in which case $x\in F$ because the $x_n$ are in $F$. \\ For the other direction you’ve done the right thing, but your explanation is just a little off. For each $n$ you’ve picked a point ... $\endgroup$ – Brian M. Scott Jan 6 at 19:21
  • 1
    $\begingroup$ ... $x_n\in V_{1/n}(x)$; that’s fine. And for any $\epsilon>0$ we have $|x_n-x|<\epsilon$ for all $n\ge\frac1{\epsilon}$. But now you can simply say that the sequence that you’ve constructed converges to $x$, and since every convergent sequence is Cauchy, we’ve shown that $x$ is the limit of a Cauchy sequence in $F$. $\endgroup$ – Brian M. Scott Jan 6 at 19:24
1
$\begingroup$

Your proofs in both directions are wrong. For $\implies$ assume that $F$ is closed and start with any Cauchy sequence $(x_n)$ in $F$. [This is imporatnt]. Since any Cauchy sequence of real numbers converges we see that $x =\lim x_n$ exists as a real number. Since each $x_n \in F$ and $F$ is closed it follows that $x \in F$. For the converse let $(x_n)$ be as sequence in $F$ converging to some $x$. We have to prove that $x \in F$. Now $(x_n)$ is a Cauchy sequence in $F$. By assumption the limit $x$ is in $F$.

$\endgroup$

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.