3
$\begingroup$

In Rudin - Real and complex Analysis we have written:

If $X$ and $Y$ are topological spaces and if $f$ is a mapping of $X$ into $Y$, then $f$ is said to be continuous provided that $f^{-1}(V)$ is an open set in $X$ for every open set $V$ in $Y$.

Can you help clarify what this means? I have thought on, and I don't get it.


This is my attempt to understand, perhaps you can highlight my issues with this:

So $X$ and $Y$ both have a topology over them, firstly, I have no idea if this is the same or a different topology. I imagine it is only the same if $X=Y$, so here is different in all cases other than that.

We have $f:X\to Y$, and we (obviously) have $f^{-1}:Y\to X$. We say that $f$ is continuous if we can go back from $Y$ to $X$, and when we do so from any open set in $Y$, we obtain an open set of $X$.

I think that all of the elements of the topology $X$ are open sets in $X$, and hence we simply need $f^{-1}$ to map to some element in $X$, but that seems pointless, since then any topological space with a mapping would be continuous?

$\endgroup$
  • 1
    $\begingroup$ A map is continuous if and only if the preimage of any open set is open again. Are you looking for examples and non-examples or where do you need clarification? $\endgroup$ – Christoph Apr 7 '15 at 12:58
  • $\begingroup$ Do you understand the definition? If yes, do you have problems developing some intuition about continuous functions? If that's the case, I recommend looking at the equivalent definitions of continuous maps, for example, that one which says a map is continuous if it maps the closure of $A$ into the closure of $f(A)$. This characterization can be interpreted as "points which are close the $A$ stay close to $f(A)$". $\endgroup$ – Stefan Hamcke Apr 7 '15 at 12:59
  • $\begingroup$ I guess the definition uses the (quite standard) notation $f^{-1}(V) = \{x\in X\ :\ f(x)\in V\}$. Of course, the existence of a function $f^{-1}:Y\rightarrow X$ is far from obvious, since it's false whenever $f$ is not bijective. $\endgroup$ – user228113 Apr 7 '15 at 14:21
1
$\begingroup$

One source of confusion is loading $f^{-1}$ with two meanings.

Given a function $f: X \to Y$, you first see $f^{-1}$ defined only when $f$ is a bijection, in which case $f^{-1}: Y \to X$ is the inverse of $f$. Unfortunately, that's not the case here.

In this situation, $f^{-1}(B)$ denotes the preimage of a subset $B \subseteq Y$, so that $f^{-1}(B)=\{x \in Y : f(x) \in B\}$.

To make things perfectly clear, I'd like to stop overloading $f^{-1}$, and instead use the notation $\bar{f}^{-1}$ to denote the preimage function of $f$ To highlight exactly what our preimage function maps between. As it takes as input subsets of our codomain, and returns subsets of our domain, it would be more precise to write

$$ \bar{f} : P(X) \to P(Y) $$

for our image function of $f$, writing $\bar{f}(A) =\{f(x) : x \in A\}$ instead of the usual $f(A)$, and

$$ \bar{f}^{-1} : P(Y) \to P(X)$$

for our preimage function. This viewpoint makes it clear that $\bar{f}$ and $\bar{f}^{-1}$ are really functions between the $P(X)$ and $P(Y)$, the power sets of $X$ and $Y$.

Now, with topological spaces $X$ and $Y$, we're given special collections of subsets $\tau_X \subseteq P(X)$ and $\tau_Y \subseteq P(Y)$ declared to be open, satisfying certain properties. However, there's no reason that if we restrict our preimage function $\bar{f}^{-1}$ to the open subsets of $Y$, that we should only get open subsets of $X$, for every function $f$. This is exactly why continuity is defined; to single out such special functions!

More precisely, the function $f : X \to Y$ is called continuous if the restriction of $\bar{f}^{-1}$ to $\tau_Y$ is a function to $\tau_X$; that is, we have

$$\bar{f}^{-1}: \tau_Y \to \tau_X.$$

I realize this is a (possibly even more) abstract way of thinking about continuity, but I got the impression that you may benefit from seeing it spelled out, very clearly, what all the moving pieces are, and where they're coming from.

$\endgroup$
  • $\begingroup$ Now how do you distinguish the inverse of $\bar f$ (if it exists) from $\bar f^{-1}$? ;-) $\endgroup$ – Christoph Apr 7 '15 at 15:21
  • $\begingroup$ @Christoph Why, just keep climbing down the rabbit hole, of course; you'll need bars upon bars! I was actually thinking about that, but hadn't made a decision :) $\endgroup$ – pjs36 Apr 7 '15 at 15:38
  • $\begingroup$ Using your convention, both things have the same notation! $\endgroup$ – Christoph Apr 7 '15 at 15:39
  • $\begingroup$ @Christoph I did stop overloading $f^{-1}$, as claimed! Of course it was just by pushing around what's overloaded, which is unfortunate and unavoidable, without some agreed-upon notation for 'preimage' that doesn't clash with the inverse notation. I hope it's only a minor annoyance, and that the big idea of $\operatorname{image}$ and $\operatorname{preimage}$ being maps between $P(X)$ and $P(Y)$ isn't overshadowed by it, but such is life. $\endgroup$ – pjs36 Apr 7 '15 at 16:01
3
$\begingroup$

$X$ and $Y$ can have different topologies. Even if $X$ and $Y$ are the same set, they can have different topologies. For example $X=Y=\mathbb{R}$, $X$ has the standard topology on $\mathbb{R}$ and $Y$ has the Lower limit topology. In this example, sets like $[a,b)$ are open in $Y$ and sets like $(c,d)$ are open in $X$ (without getting into arbitrary unions of open sets at least). So if there is a continuous function $f:X \to Y$ then we would expect to see something like $f^{-1}\left([a,b)\right) = (c,d)$ or $f^{-1}\left([a,b)\right)$ being equal to some other open set in $X$.

Was there a proof of the equivalence of topological continuity and $\varepsilon$-$\delta$ continuity in your text? If not, I highly recommend you prove that result. Getting a firm understanding of how the two definitions imply each other should go a long way in improving your understanding of topological continuity. The "inverse-image of open sets is open" definition is not very intuitive, but it is important for success in topology.

$\endgroup$

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.