2
$\begingroup$

Let $X$ be a set, $\tau_1 \leq \tau_2$ two comparable topologies on $X$ ($\tau_1$ is weaker than $\tau_2$) and consider some topological property $\varphi$ that holds for both $\tau_1$ and $\tau_2$. I am interested in a list of properties $\varphi$ that hold for all topologies $\tau$ in between $\tau_1$ and $\tau_2$, i.e. $\tau_1 \leq \tau \leq \tau_2$ and also in a list of properties for which there exists a topology $\tau$ in between $\tau_1$ and $\tau_2$ that does not satisfy $\varphi$.

Examples:

  • if $\varphi$ is Hausdorffness and $\tau_1$ satisfies $\varphi$ (and $\tau_2$ the discrete topology) then all $\tau \geq \tau_1$ satisfy $\varphi$. This is also true for many other separation axioms.
  • if $\varphi$ is (sequential) compactness and $\tau_2$ satisfies $\varphi$ (and $\tau_1$ the trivial topology) then all $\tau \leq \tau_2$ satisfy $\varphi$.
  • if $\varphi$ is first-countability or sequentiality and $X$ is infinite then $\varphi$ is not preserved in between the trivial topology and the discrete topology (e.g. the Arens-Fort space).

Does anyone know of a reference for such a list of "non-boring" properties $\varphi$ and suitable (e.g. maximal) choices for $\tau_1$ and $\tau_2$? (I think also of more advanced questions like: "what property $\psi$ does $\tau_1$ and $\tau_2$ has to satisfy such that if $\tau_1$ and $\tau_2$ satisfy $\varphi$ (e.g. are metrizable) then any topology in between also satisfies $\varphi$.)

| cite | improve this question | | | | |
$\endgroup$
5
$\begingroup$

[Edit: Originally my second point below stated incorrectly that if $\tau_1$ and $\tau_2$ are as described, then every topology in between is metrizable. This is not what the theorem in my paper says, but yesterday, in my haste, I thought that the theorem in my paper should imply this statement. It does not, as I now realize, and I've corrected the error.]

I explore some questions like this in a paper [Preservation and destruction in simple refinements, Top. App. 178 (2014), pp. 236-247]. Some of the main results along these lines are:

  1. If $\tau_1$ is (completely) regular and $\tau_2 \supseteq \tau_1$, then every topology in between is (completely) regular if and only if every $\tau_2$-open set is the intersection of an open set and a closed set in $\tau_1$.

  2. Suppose $\tau_1$ is metrizable and $\tau_2$ has a basis of the form $\tau_1 \cup B$, where $B$ is countable, and every member of $B$ is the intersection of an open set and a closed set in $\tau_1$. Then $\tau_2$ is also metrizable. However, unless $\tau_1 = \tau_2$, there will be non-metrizable topologies strictly in between the two. [The cited paper does not contain a proof of this latter fact, but you can read about it here.]

  3. Suppose $\tau_1$ is Polish. Then the space with basis $\tau_1 \cup \{A\}$ is Polish if and only if $A$ is the intersection of an open set and a closed set in $\tau_1$.

  4. Suppose $\tau_1$ is locally compact. Then the space with basis $\tau_1 \cup \{A\}$ is locally compact if and only if $A$ and its complement are each the intersection of an open set and a closed set in $\tau_1$.

Also, let me say that your comment about separation axioms is true for lower separation axioms ($T_{2\frac{1}{2}}$ and down) but not for higher ones ($T_3$ and up).

| cite | improve this answer | | | | |
$\endgroup$
  • $\begingroup$ Thanks. It is nice to see, that this is current research. Do you also know of further related work regarding such intervals of topologies? $\endgroup$ – yada Jan 14 '16 at 18:52
  • $\begingroup$ There's a fair amount of work on intervals in the lattice of topologies, but of course not all of it is aimed at answering these sorts of questions. I tried to include most of what I'm aware of in the references section, though it's not much. I see it as a relatively unexplored topic. $\endgroup$ – Will Brian Jan 14 '16 at 19:27

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.