I found this in a handwritten note:

Defination : A topological space $X$ is $T_1$ if $\forall x \neq y \in X$ there exist a neighborhood of $y$ such that s.t. $x \not\in V$.

I was almost certain that this was incorrect, because it looked like the definition for $T_0$ space.

However, subsequently this definition was used to prove that all singletons are closed, for example:

$(\Leftarrow)$ Suppose $\{x\}$ is closed, then $\{x\}^c$ is open and $y \in \{x\}^c$ satisfies the definition of a $T_1$ space.

Is this some alternative definition of $T_1$ space?


The definition is correct, albeit a little sloppy. Here’s a slightly more careful version of it:

$X$ is $T_1$ if for each $x\in X$ and each $y\in X\setminus\{x\}$ there is a nbhd $V$ of $y$ such that $x\notin V$.

Let $p$ and $q$ be any two distinct points of $X$. If we set $x=p$ and $y=q$, we see that the definition ensures that $q$ has a nbhd that does not contain $p$. We can just as well set $x=q$ and $y=p$, however, and conclude that $q$ has a nbhd that does not contain $p$. Thus, the definition really does say that $X$ is $T_1$. The point is that since both quantifiers are universal, we can interchange the rôles of $x$ and $y$ in the definition.

The real problem with the quoted version is that it doesn’t make the quantifier on $x$ explicit. Because of this, it’s easy on first reading to get the impression that the definition is asymmetric in $x$ and $y$, like the definition of $T_0$ separation.

  • $\begingroup$ That's interesting. Because what I have is that $(X, \tau)$ is $T_1$ if $\forall x \neq y \in X$, $\exists U, V \in \tau$ such that $x \in U, y \not\in U$ and $y \in V, x \not\in V$. Whereas in the def above $x$ is not enclused in another open set $\endgroup$ – Olórin Jun 14 '16 at 3:54
  • $\begingroup$ @MSEisadatingsite: You don't have to assert the existence of both $U$ and $V$ in that definition. If you assert only the existence of $U$, then you can get a set $V$ by just applying the definition again with $x$ and $y$ swapped. $\endgroup$ – Eric Wofsey Jun 14 '16 at 3:55
  • $\begingroup$ @MSEisadatingsite: That follows immediately from the rephrased version that I gave: it gives an open $V$ such that $y\in V$ and $x\notin V$, and after you interchange $x$ and $y$ it gives you an open $U$ such that $x\in U$ and $y\notin U$. $\endgroup$ – Brian M. Scott Jun 14 '16 at 3:55
  • $\begingroup$ But we cannot do the same for $T_2$ space right? We must show there exists two disjoint open sets $\endgroup$ – Olórin Jun 14 '16 at 3:56
  • 1
    $\begingroup$ @MSEisadatingsite: Correct: this style of definition of the $T_2$ property requires talking about both open sets at the same time, because they must be related in a certain way: specifically, they must be disjoint. Note, though, that there are ways to define $T_1$ and $T_2$ that look very different from these: you should try to prove that $X$ is $T_1$ if and only if $\{x\}$ is the intersection of all open nbhds $x$ for each $x\in X$, and $X$ is $T_2$ iff $\{x\}$ is the intersection of all closed nbhds of $x$ for each $x\in X$. $\endgroup$ – Brian M. Scott Jun 14 '16 at 3:59

Yes, this definition is correct. It is well-known that a space is $T_1$ iff all singleton sets are closed. For the other direction of the proof, suppose $X$ is $T_1$, let $x\in X$, and let $U$ be the union of all open sets which do not contain $x$. Then $U$ is open since it is a union of open sets, and does not contain $x$. Since $X$ is $T_1$, for any $y\neq x$, there is an open $V$ such that $y\in V$ and $x\not\in V$. The set $V$ is then one of the sets whose union is $U$, so $V\subseteq U$. In particular, $y\in V$. Since $y\neq x$ was arbitrary, this means $U=X\setminus\{x\}$. Thus $X\setminus\{x\}$ is open, so $\{x\}$ is closed.

The difference between this definition and the definition of $T_0$ is that in the definition of $T_0$, all you know is that there exists an open set $V$ such that either $y\in V$ and $x\not\in V$ or $x\in V$ and $y\not\in V$. So for example, consider the space $S=\{0,1\}$ with topology $\{\emptyset,\{0\},\{0,1\}\}$. This space is $T_0$: the only pair of points with $x\neq y$ is $0$ and $1$, and $\{0\}$ is a set which contains one but not the other (for some ordering of the two points). However, this space is not $T_1$, since if you set $x=0$ and $y=1$, there is no open set which contains $y$ but not $x$.

  • $\begingroup$ When there are two good answers which one should I accept? $\endgroup$ – Olórin Jun 14 '16 at 4:00
  • $\begingroup$ Whichever one you found more helpful! Personally, I think Brian Scott's answer is a bit better than mine, so I'd encourage you to accept it. $\endgroup$ – Eric Wofsey Jun 14 '16 at 4:01

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.