1

Suppose that $f$ is defined on an open neighbourhood of a point $a\in\mathbb R$. I wish to prove that if $f$ is "sequentially continuous" at $a$, then $f$ is continuous at $a$. The definition of "sequentially continuous" is that for every sequence $(a_n)$ that approaches $a$ we have $f(a_n)\to f(a)$. The proof is by contraposition.


Suppose that $f$ is not continuous at $a$. Then, there is an $\varepsilon>0$ such that for every $\delta>0$ there is an $x$ for which $|x-a|<\delta$ but $|f(x)-f(a)|\ge\varepsilon$. We will now construct a sequence $(a_n)$ such that $(a_n)\to a$ but it is not true that $f(a_n)\to f(a)$.

We know that there is an $\varepsilon>0$ and an $x$ such that $|x-a|<1$ and $|f(x)-f(a)|\ge\varepsilon$. Let $a_1$ be one such $x$. We know that there is an $\varepsilon>0$ and an $x$ such that $|x-a|<0.1$ and $|f(x)-f(a)|\ge\varepsilon$. Let $a_2$ be one such $x$. Continue in this fashion. More formally, let $$ A_i=\{x:|x-a|<10^{-i}\text{ and }|f(x)-f(a)|\ge\varepsilon\} $$ and consider the set $$ S=\{A_i:i\in\mathbb N\} $$ The discontinuity of $f$ at $a$ means that each $A_i$ is non-empty, and so we can use the axiom of choice to define a function $g:S\to\bigcup S$ such that $g(A_i)\in A_i$ for all $A_i\in S$. Now let $a_i=g(A_i)$, and we are done.


My questions are:

  1. Is this argument correct?
  2. Since $S$ is countable, am I right in thinking that we only need the axiom of countable choice to prove this theorem?
Joe
  • 19,636
  • 3
    Your argument looks good, but in the less formal version, you shouldn't repeat the statement "we know that there is an $\varepsilon > 0$ such that ...", $\varepsilon$ should stay fixed after you first introduce it. You are right that this only needs the axiom of countable choice. – Rob Arthan Oct 04 '21 at 22:25
  • Yes and yes.... – spaceisdarkgreen Oct 04 '21 at 22:25
  • @RobArthan: That answers my question completely. You can post your comment as an answer, if you wish. – Joe Oct 04 '21 at 22:26
  • 1
    By the way, this is covered in Rudin PMA Theorem 4.2. – Adam Rubinson Oct 04 '21 at 22:27
  • Joe - my comment didn't really add anything other than to confirm your ideas (of, course, I am happy to have done that). Why not write up an answer of your own, if the ideas are now clear in your mind. – Rob Arthan Oct 04 '21 at 22:42
  • @RobArthan: Sure, I will do so as soon as I get the chance. – Joe Oct 04 '21 at 22:44
  • https://math.stackexchange.com/questions/126010/continuity-and-the-axiom-of-choice didn't answer your question? – Asaf Karagila Oct 04 '21 at 23:37
  • @AsafKaragila: Since I have only recently started learning about the axiom of choice, I wanted to attempt to construct my own argument, without looking at the linked post in too much detail. (Hence why this post is tagged solution-verification.) I now see that my argument is essentially the same as John Gower's, and so there is no issue in you closing this question as a duplicate. – Joe Oct 05 '21 at 11:24

0 Answers0