# $f$ is continuous at $a$ iff for each subset $A$ of $X$ with $a\in \bar A$, $f(a)\in \overline{ f(A)}$.

Definiton.

$f$ is continuous at $a$ provided that for each open set $V$ in $Y$ containing $f(a)$ there is an open set $U$ in $X$ containing $a$ such that $f(U) \subset V$.

Problem.

$f$ is continuous at $a$ iff for each subset $A$ of $X$ with $a\in \overline{A}$, $f(a)\in \overline{ f(A)}$.

• Welcome to math.SE: since you are new, I wanted to let you know a few things about the site. In order to get the best possible answers, it is helpful if you say in what context you encountered the problem, and what your thoughts on it are; this will prevent people from telling you things you already know, and help them give their answers at the right level. – Julian Kuelshammer May 28 '13 at 17:10
• – Martin Sleziak Dec 3 '18 at 16:53

## 3 Answers

Suppose $f$ is continuous at $a$, according to the definition.

Let $A$ be a subset with $a \in \overline{A}$, we want to show that $f(a) \in \overline{f[A]}$.

We will use that a point is in the closure of a set iff every open neighbourhood of the point intersects that set.

So let $V$ be an open neighbourhood of $f(a)$. Using the definition, we find a neighbourhood $U$ of $a$ such that $f[U] \subset V$. As $a \in \overline{A}$ we know that $V$ intersects $A$, say $a' \in A \cap V$. But then $f(a') \in V \cap f[A]$ (as $f[U] \subset V$), so indeed $V$ intersects $f[A]$, and as $V$ is arbitrary, $f(a) \in \overline{f[A]}$. This shows one implication.

Now let $f$ fulfill the closure condition at $a$; we want to see that $f$ is continuous at $a$. So let $V$ be an open neighbourhood of $f(a)$. Suppose now (striving for a contradiction) that for all open neighbourhoods $U$ of $a$ we have that $f[U] \not\subset V$. This means that every neighbourhood of $a$ contains points of $X \setminus f^{-1}[V]$, so $a \in \overline{X \setminus f^{-1}[V]}$, and so, by the closure condition, $f(a) \in \overline{f[X \setminus f^{-1}[V]} \subset \overline{Y \setminus V}$, but this is false, as $V$ does not intersect $Y \setminus V$. This contradiction shows the continuity of $f$ at $a$.

The condition $f(\bar A)\subseteq \overline{f(A)}$ for each $A\subseteq X$ can be taken as the definition of continuity (provided one defines a closure operator before one defines open sets, of course). Let's assume the

DEF $f:(X,\mathscr T)\to(Y,\mathscr U)$ is continuous if for each open $A\in Y$, $f^{-1}(A)$ is open in $X$.

We know that this is equivalent to

DEF* $f:(X,\mathscr T)\to(Y,\mathscr U)$ is continuous if for each closed $F\in Y$, $f^{-1}(F)$ is closed in $X$.

Then we have the

THM $f:(X,\mathscr T)\to (Y,\mathscr U)$ is continuous if and only if for each $A\subseteq X$, $f(\bar A)\subseteq \overline{f(A)}$

PROOF Suppose that $f$ is continuous. Given a subset $A$, we have $f(A)\subseteq \overline{f(A)}$, which means $A\subseteq f^{-1}f(A)\subseteq f^{-1}\overline{f(A)}$. This last set is closed, so $\bar A\subseteq f^{-1}\overline{f(A)}$, which means $f(\bar A)\subseteq \overline{f(A)}$.

Conversely, suppose that $f(\bar A)\subseteq \overline{f(A)}$ for each $A\subseteq X$. Let $F$ be closed in $Y$. Then $f(\overline{f^{-1}(F)})\subseteq \overline{f({f^{-1}(F)})}\subseteq \bar F=F$. Thus $\overline{f^{-1}(F)}\subseteq f^{-1}(F)$. Since the converse always holds, we have $f^{-1}(F)=\overline {f^{-1}(F)}$, so this set is closed and $f$ is continuous.

$\to)$ Hint: Let $A\subseteq X$ and $x\in \overline{A}$ and $V$ be a neighborhood of $f(x)$. It's enough to show $f(A)\cap U\ne \emptyset$. $f^{-1}(U)$ is a neighborhood of $x$. There's some $$z\in f^{-1}(U)\cap A$$ $$\Rightarrow f(z)\in U\cap f(A)$$

$\leftarrow)$ Hint: Let $x\in X$ and $V$ be an open neighborhood of $f(x)$. It's enough to show $f^{-1}(V)$ is a neighborhood of $x$. Suppose it is not. $x$ is in the boundary of $f^{-1}(V)$. So $$x\in \overline{\big(f^{-1}(V) \big)^c}$$ $$\Rightarrow f(x)\in f\big(\overline{\big(f^{-1}(V) \big)^c}\big)\subseteq \overline{f\big(\big(f^{-1}(V) \big)^c\big)}=\overline{f\big(f^{-1}(V^c) \big)}\subseteq \overline{V^c}$$ So $f(x)$ is in the boundary of $V$. A contradiction.

• "Hint"?! Really? – Sharkos May 28 '13 at 19:02