Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I want to prove that if $f,g$ are continuous functions from a topological space $(X,\tau)$ to a metric space $(Y,d)$ then the set

$$ A = \{ x \in X : f(x) = g(x) \} $$

is closed. I found a very similar question asked on this site but it referred to Hausdorff spaces, which I haven't quite met. I know I can prove this by proving that the complement of $A$ is open. If I consider $b \in X$ such that $f(b) \neq g(b)$, then I can find two open sets $U,V \subseteq Y$ such that $f(b) \in U $, $\ g(b) \in V$ and $U \cap V = \varnothing$. Then since the functions are continuous we must have that $f^{-1}(U)$ and $g^{-1}(V)$, which both contain $b$, are open. Hence so is their intersection.

Now I'm not sure how to proceed...

EDIT: Whilst there are neater methods in the answers, I've just realised how I can finish the proof above. The intersection

$$ W_b = f^{-1}(U) \cap g^{-1}(V)$$

lies entirely outside $A$ --- one can show this by supposing that there exists some $a \in A$ that is a member of $W_b$ and showing that $f(a) \in U$ and $g(a) \in V$. But $f(a) = g(a)$ by the assumption that $a \in A$, which is impossible since $U \cap V = \varnothing$.

Then the union

$$ \bigcup_{b \in X \setminus A} W_b$$

is a union of open sets and hence open, contains all $b \in X \setminus A$, and does not intersect with $A$ since $W_b \cap A = \varnothing$. Hence the union above is precisely $X \setminus A$, and hence $A$ is closed.

share|cite|improve this question
up vote 8 down vote accepted

If $Y$ is a metric space, the set $\Delta =\{(y,y):y\in Y\}$ is closed in $Y\times Y$ -- this holds more generally iff $Y$ is Hausdorff (can you prove this?). Observe now that $$A=\{x\in X:f(x)=g(x)\}$$ equals the set $$B=\{x\in X:h(x)\in \Delta\}=h^{-1}(\Delta)$$ where $h:X\to Y\times Y$ is defined by $h(x)=(f(x),g(x))$. But $\Delta$ is closed, and $h$ is continuous. This proves the more general case where $Y$ is Hausdorff.

share|cite|improve this answer

Consider the map $F:X\rightarrow \Bbb R$ given by $F(x)=d(f(x),g(x))$. $F$ is continuous (you can check this), and $F^{-1}(\{0\})=A$ so $A$ is closed since $\{0\}$ is closed in $\Bbb R$.

share|cite|improve this answer
What is $0$ in metric space? – Hamou Jul 26 '14 at 22:38
@Hamou $0\in\Bbb R$! – Pedro Tamaroff Jul 26 '14 at 22:38

It's equal to

$$\bigcap \{x\in X: d(f(x),g(x))\le \epsilon\}$$

which is the intersection of closed sets since $Y$ is metric (hence Hausdorff) and $f,g$ are continuous.

We know it's closed because $Y$ is Hausdorff, so that the diagonal, $\Delta Y\subseteq Y\times Y$ is closed. This is so because the complement is open:

Proof: If $(x,y)\ne (y,y)$ then $B(x,{1\over 2}d(x,y))\times B(y,\delta)$ is an open set around $(x,y)$ not containing $(y,y)$, so the complement is open.

share|cite|improve this answer
It doesn't matter if the intersection is finite when you're dealing with closed sets, that's only for open sets. – Adam Hughes Jul 26 '14 at 22:42
@AymanHourieh I've updated my answer to address your question of "why is it closed." Really it holds more generally in all Hausdorff spaces, there you substitute the balls I used in the answer by an open set containing $x$ by not $y$ crossed with any open set around $y$. – Adam Hughes Jul 26 '14 at 22:50
I think it was fine. I just got confused for a second and thought we were talking about a union. – Ayman Hourieh Jul 26 '14 at 22:53

Let $X$ be topological space and Y be a metric space. We show that X\A is open. Let $x \in$ X\A be arbitrary.

Then $f(x)\neq g(x)$ (Otherwise x would be in A). We define $\epsilon:=d(f(x),g(x))$, so the open neighborhoods $U_{\frac{\epsilon}{2}}(f(x))$ and $V_{\frac{\epsilon}{2}}(g(x))$ are disjoint.

Hence $W:=f^{-1}(U_{\frac{\epsilon}{2}}(f(x)))\cap f^{-1}((V_{\frac{\epsilon}{2}}(g(x))) \neq \emptyset $ is an open neighborhood of x with $W \subseteq$ X\A. If this were not true then there were a $y \in W$ such that $f(y)=g(y)$ which can't be happen.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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