4
$\begingroup$

I have seen many definitions in two versions: sometimes referred to sets, sometimes to spaces.

Some examples: closed set/space, compact set/space, $F_{\sigma\delta}$ set/space.

I asked one of my lecturers about this and they said that if I introduce some definition with "space", I need to explain it more. So I am thinking, defining something as a set is more general, while defining on a space requires looking also at the space structure? Or does this depend on particular case?

I am particularly interested in the case with $F_{σδ}$ sets vs spaces, but I think it is important question in general.

Thank you for your advice.

$\endgroup$
2
  • 4
    $\begingroup$ A 'space' could mean different things in different contexts e.g. vector space, topological space, metric space. More generally, I would think of a 'space' as a set with some additional (usually geometrical) structure. For example, a vector space has an addition and scalar multiplication, a metric space has a metric and a topological space has a topology; all are sets. $\endgroup$
    – user829347
    Jan 21 at 20:51
  • 1
    $\begingroup$ An absolute $G_\delta$ space is something different from a $G_\delta$ subset, e.g. $\endgroup$
    – Henno Brandsma
    Jan 21 at 22:39

2 Answers 2

Reset to default
4
$\begingroup$

The difference is simply that "space" is a notion that implies some underlying structure to it. A set may not have that structure.

This is particularly key when we begin to talk of subsets versus subspaces. Take $\Bbb R^n$, a vector space, as an example. There are a lot of subsets there: pick any number of elements you want at random, but odds are, they won't satisfy the vector space axioms and thus be a (vector) subspace.

What this "structure" is depends on the context (vector spaces, topological spaces, measure spaces, etc.), however. But regardless, when saying "space", you're implying the existence of some kind of underlying structure -- your instructor is poking at that fact: what is that structure?

It's a bit confusing since, especially colloquially, the terms often are used in about the same way. However, there is that technical and important difference there.

$\endgroup$
2
$\begingroup$

The terminology "a closed set" etc, is sloppy, but, alas, common. The better one is "a closed subset" since it makes sense only relative to an ambient topological space. Similarly, use "compact subset" or "compact space." Ditto $F_\sigma$ and $G_\delta$ subsets. In general, when a topologist says "a space" (without any further adjectives), they mean "a topological space." This is very common, get used to it. There are also $G_\delta$-spaces. This notion is different from the one of $G_\delta$-subsets in a topological space.

$\endgroup$
4
  • $\begingroup$ So for the borel sets, $F_\sigma$ subset of some set is different than $F_\sigma$ space? I am only familiar with the notion of $F_\sigma$ set. But I would assume it can be defined as a space too, if the space has open and closed sets defined. So topological space woud be okay. I dont understand why for $G_\delta$ is is different. $\endgroup$
    – Tereza Tizkova
    Jan 21 at 21:41
  • 1
    $\begingroup$ @TerezaTizkova: Yes, the notion of a $F_\sigma$-subset makes sense only in the context of an ambient topological space. Same for $G_\delta$-subsets. As for a $G_\delta$-space, it is a topological space where every closed subset is a $G_\delta$-subset. See the link above. $\endgroup$
    – Moishe Kohan
    Jan 21 at 22:45
  • $\begingroup$ Thank you, didn´t know that. But for the $F_\sigma$, do the notions "correspond"? E.g. can I define a $F_\sigma$-subset $A$ of a topological space $X$ and then say that $A$ is a $F_\sigma$ space iff it is a $F_\sigma$-subset and a subspace of $X? (While for the $G_\delta$ such definition would not make sense?) $\endgroup$
    – Tereza Tizkova
    Jan 22 at 13:22
  • $\begingroup$ @TerezaTizkova: I cannot read your comment because of a TeX error, but $X$ is an $G_\delta$-space (in the sense of the above definition) if and only if every open subset of $X$ is an $F_\sigma$-subset in $X$. Maybe this is what you asked. For this reason, there is no separate definition of $F_\sigma$-spaces (since it leads to the same class). $\endgroup$
    – Moishe Kohan
    Jan 22 at 15:30

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.