# Topologies coinciding at a point or a set.

Consider a set equiped with two topologies. What does it mean to say that the two topologies coincide at a point in the set? Is it meaningful to talk about this concept in general. Is it meaningful in the context of metric spaces?

p.s. I know what the relative topology is for any subset of a topological space and this concept is a different one, as is apparent from this problem.

• Is this not coming from a book you are studying? And the context is not included in that book? – Jonas Meyer Jul 9 '13 at 6:49
• With no context I’d take it to mean that if $p$ is the point and $\tau_1$ and $\tau_2$ are the topologies, then for each $A\subseteq X$, $x\in\operatorname{cl}_{\tau_1}A$ iff $x\in\operatorname{cl}_{\tau_2}A$. That, however, would be a guess. – Brian M. Scott Jul 9 '13 at 6:51
• It really depends on your book, but I have provided two decent guesses below. – Thomas Andrews Jul 9 '13 at 7:11
• I think @BrianM.Scott's definition is equivalent to my first definition. – Thomas Andrews Jul 9 '13 at 7:22
• @Thomas: Yes; they both amount to saying that the point has the same nbhds in both topologies, where by nbhd I mean a set that contains the point in its interior. – Brian M. Scott Jul 9 '13 at 9:11

I would guess that it means the every open neighborhood of $x$ in $\tau_1$ contains an open neighborhood of $x$ in $\tau_2$, and visa versa. This essentially means that "continuity at $x$" is the same in each topology.

It's possible, though, that the stronger statement is intended: the neighborhoods of $x$ are the same in each topology.

There's a lovely way to define the "local topology" at a point. Let $(X,\tau)$ be a topology and $x\in X$. Define a new topology, $\tau_x$ which has as a basis the singletons $\{y\}$ where $y\neq x$ and $U\in\tau$ such that $x\in U$. So the open sets are any set not containing $x$, or any set contain a $\tau$-neighborhood of $x$. It is not hard to see that this is a topology.

In this case, my first definition would be equivalent to saying, for topologies $\tau$ and $\rho$, they coincide at $x$ if and only if $\tau_x=\rho_x$.

Now, in the complete lattice of topologies on $X$, $$\bigcap_{x\in X}\tau_x = \tau$$ So we can recover $\tau$ if we know each $\tau_x$. Then, for a topological space, $Y$, we can define a function $f:X\to Y$ to be $\tau$-continuous at $x$ if it is continuous on $(X,\tau_x)$, and it turns out, due to the lattice meet property above, the function $f$ is continuous on $(X,\tau)$ if and only if it is continuous at each $(X,\tau_x)$.

There's also a simple "local range topology", $\tau^x$ defined as just the $\tau$-open sets containing $x$, and the empty set. Then you can say that $f:Y\to X$ is continuous "to" $x$ if it is continuous when using the topology $\tau^x$. Again, we can recover $\tau$ if we know ecah $\tau^x$. Indeed, $\tau = \bigcup \tau^x$.

The stronger definition above would happen exactly when $\tau^x=\rho^x$. It's true that if $\tau^x=\rho^x\implies\tau_x=\rho_x$, though so maybe this stronger definition is reasonable.