0

I am wondering if there is any study of "first order topological properties" that are preserved by continuous maps (between general topological spaces, or spaces satisfying mild conditions like Hausdorffness or whatever, or "nice" spaces).

I don't really have a formalization of "first order topological properties", but probably something along the lines of: in the context of a topological space $(X,\mathscr T)$, properties that $(X,\mathscr T)$ can satisfy or not satisfy, described only using notations like $\in$ (element of), $\subseteq$ (subset of),$\smallsetminus$ (set minus), etc. and also set theoretic things like $|\bullet |$ (cardinality), $\bigcup$ (union), $\times$ (Cartesian product), functions, and references to "standard" sets/spaces like $\mathbb N$ or $[0,1]$ (and of course all the standard first order logic things like $\forall$ (for all), $\exists$ (there exists), $=$ (equals), etc.).

Most famously (Topological properties preserved by continuous maps -> https://en.wikipedia.org/wiki/Continuous_function#Properties_2), we have

  • $\forall \mathscr U \subseteq \mathscr T$ s.t. $X \subseteq \bigcup \scr U$, $\quad\exists \mathscr U' \subseteq \mathscr U,\exists n\in \mathbb N$ with $|\mathscr U'|=n$ and $X \subseteq \bigcup \mathscr U'$.
  • $\exists U \in \mathscr T$ s.t. $(X\smallsetminus U) \in \mathscr T$.
  • $\forall x,y\in X, \exists$ continuous $\gamma:[0,1]\to X$ s.t. $\gamma(0)=x$ and $\gamma(1)=y$.
  • $\exists d:\mathbb N \to X$ s.t. $\forall U \in \mathscr T$, $\quad \exists n\in \mathbb N$ s.t. $d(n) \in U$.

otherwise known as compactness, connectedness, path-connectedness, separability. Of course we can have minor variations of these --- e.g. for compactness we also have

  • Lindelof property: $\forall \mathscr U \subseteq \mathscr T$ s.t. $X \subseteq \bigcup \scr U$, $\quad\exists \mathscr U' \subseteq \mathscr U,\exists$ a bijection $b:\mathbb N\to \mathscr U'$ and $X \subseteq \bigcup \mathscr U'$
  • countable compactness: $\forall \mathscr U \subseteq \mathscr T$ s.t. [$X \subseteq \bigcup \scr U$ and $\exists$ a bijection $b: \mathbb N \to \mathscr U$], $\quad\exists \mathscr U' \subseteq \mathscr U,\exists n\in \mathbb N$ with $|\mathscr U'|=n$ and $X \subseteq \bigcup \mathscr U$
  • $\sigma$-compactness, sequential compactness, etc.

These seem so "hodge-podge"; there surely must be other simple properties we can formulate (that may not have such nice intuitive meanings as the above notions)! And surely someone has made a systematic study of such notions!

Hopefully there are some as-yet undiscovered such properties that are "useful" in the sense that "nice" sets in "nice" spaces fall in both camps: satisfy and does-not-satisfy (e.g. compactness and connectedness are "useful" because we come across some sets that are compact, and some sets that are not compact all the time when working in say $\mathbb R^d$ and these properties allow us to distinguish them topologically).

D.R.
  • 8,691
  • 4
  • 22
  • 52
  • i think your question is essentially a duplicate of the quesiion that you cite. It is too open-ended to have a useful answer. – Rob Arthan Mar 08 '24 at 20:09
  • @RobArthan no, I already gave those examples in the body of my post. I am looking for other properties, that are lesser known. Please don't mark this as a duplicate of that one, because I am asking for new properties. Again, I already know the ones discussed in the link in your comment, and I already discussed them again in the body of my post, hoping to learn something new. I am trying to learn something new here, not rehashing old content. – D.R. Mar 08 '24 at 20:29
  • In that case, your question is far too broad to have a useful answer. You are not even asking for a classification but rather for a survey of an open-ended field. My close vote stands. – Rob Arthan Mar 08 '24 at 20:34
  • @RobArthan I am hoping for a classification of course, but that would be unreasonable to expect one. But still, asking for one example of such a property significantly different to the ones I've listed so far, I think is a fair question. Your close reason is "duplicate", and I don't think this is accurate. None of the close reasons stand: it is not opinion based, it is not multiple questions in one, it is essentially well-defined (I've specified the "first order language"), etc. Sure, it's open ended, but that's not against the rules of MSE. – D.R. Mar 08 '24 at 20:39
  • You haven't given an adequate definition of what you mean by "first order topological property". Your question is far too broad. – Rob Arthan Mar 08 '24 at 20:42

0 Answers0