1
$\begingroup$

Having infinitely many open sets in a topology is a topological property

#1: What are example of two topological spaces, preferably simple ones, which this statement show are not homeomorphic.

But, more importantly why does it show this? Having trouble understanding topological properties in connection to homeomorphisms.

#2: What about two spaces where this statement doesn't help deciding whether they are homeomorphic or not?

$\endgroup$
  • 1
    $\begingroup$ Answer to first question: trivial space and discrete space? $\endgroup$ – Alex Vong Dec 15 '17 at 16:33
  • $\begingroup$ Of course the space above is infinite $\endgroup$ – Alex Vong Dec 15 '17 at 16:41
3
$\begingroup$

If $X$ and $Y$ are topological spaces with respective topologies $\tau_X$ and $\tau_Y$, and if $f : X \to Y$ is a homeomorphism, then $f$ induces a bijection $\tau_X \mapsto \tau_Y$: each $U \in \tau_X$ is mapped to its image $f(U) \in \tau_Y$ defined as usual by the formula $$f(U)=\{f(x) \,|\, x \in U\} $$ The proof that this map $\tau_X \to \tau_Y$ is a bijection is simple, and uses only the definition of homeomorphism.

As a consequence, if $X$ and $Y$ are homeomorphic then the two sets $\tau_X$ and $\tau_Y$ have equal cardinalities. So, for example, it cannot happen that one is finite and the other is infinite. It also cannot happen that one is countably infinite and the other is uncountably infinite.

$\endgroup$
1
$\begingroup$

Space 1: the interval $[0, 1]$ with the usual topology. Space 2: the interval $[0, 1]$ with the indiscrete topology, in which the only open sets are the empty set and the whole set.

Note that as sets, these spaces are identical; as topological spaces, they're incredibly distinct

Why is a set with a finite topology different from a set with an infinite topology? A homeomorphism $f$ has to map open sets to open sets (and its inverse, $g$ must do so in the other direction, with $f \circ g = id_X$ and $g \circ f = id_Y$). If $X$ has an infinite topology and $Y$ has a finite one, there must be an open set $U$ in $Y$ that's the image of more than one open set in $X$, say $V_1$ and $V_2$. But what's the preimage $g(U)$ of $U$ in $X$? Under the homeomorphism, it must be both $V_1$ and $V_2$, and that's impossible, because these were distinct open sets of $X$?

$\endgroup$
  • $\begingroup$ Thanks for the answer that make sense now. So, for the econd part of the question.... This doen't help for any two infinite topologies since the second part doesn't apply and we can't say anything about the topologies whether they are or aren't homeomorphic. $\endgroup$ – Overflow2341313 Dec 15 '17 at 16:41
  • $\begingroup$ Does this example still apply if the property is changed to "having finitely many open sets"? Since as I undertood these topologies are infinite and discrete respectively. $\endgroup$ – Overflow2341313 Dec 15 '17 at 16:54
  • $\begingroup$ The second topology is indiscrete --- in the discrete topology, every set is open. (the interval, with that topology, is also not homeomorphic to the interval with either of the other two topologies). For the second question, are you asking for two spaces with infinite topologies that are not homeomorphic? Sure: Take the integers. In topology 1, all sets are open. In topology 2, all sets that contain only even numbers are open (along with the empty set and the whole space). How do I know they're distinct? The second has a set (the odds) that's closed but not open. The first doesn't. $\endgroup$ – John Hughes Dec 15 '17 at 17:34
1
$\begingroup$

Take $\mathbb Z$ with the usual topology and with the topology$$\tau=\{\emptyset,\mathbb{Z}\}\cup\bigl\{\{a,a+1,a+2,\ldots\}\,|\,a\in\mathbb Z\bigr\}.$$This topology is countable, but the usual one isn't (since every subset of $\mathbb Z$ is open).

Note that if $f$ is a homeomorphism from a topological space $(X_1,\tau_1)$ onto a topological space $(X_2,\tau_2)$, then $f$ induces a bijection between $\tau_1$ and $\tau_2$:$$\begin{array}{ccc}\tau_1&\longrightarrow&\tau_2\\X&\mapsto&f(X).\end{array}$$Therefore, $\#\tau_1=\#\tau_2$.

$\endgroup$
  • $\begingroup$ Now the argument does work. $\endgroup$ – Henno Brandsma Dec 16 '17 at 6:27
0
$\begingroup$

Let $X$ be a point and $Y$ the integers, both with the discrete topology. Then $X$ has finitely many open sets, while $Y$ has infinitely many open sets. But since there's no bijection between the two, they definitely can't be homeomorphic.

Topological properties are the same as invariants under homeomorphism. Checking for different topological properties can tell you when spaces are not homeomorphic.

$\endgroup$

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.