I thought this question was easy, but turns out to be a little bit tricky
Let us give $\mathbb{R}$ and $\mathbb{R}^2$ the usual topology. Then, $\mathbb{R}$ and $\mathbb{R}^2$ are:
- both uncountable
- both separable, first countable, second countable, suslin
- metrizable, not compact
- connected
$$\vdots$$ Seems we require a new concept:
Let $x \in X$ be a cut point if $X \backslash \{x\}$ has a non-trivial clopen subset.
We know this is a topological invariant. See my answer here: Show that cuts are preserved under homeomorphism
It seems that the only distinguishing feature is that $\mathbb{R}$ have different amount of cut points than $\mathbb{R}^2$.
But how many cut points?
Obviously it only takes one to cut $\mathbb{R}$, and it takes uncountable many (like a line) to cut $\mathbb{R}^2$.
And how do we know whether taking away uncountable many points from $\mathbb{R}$ doesn't also produce a clopen subset? In that case, $\mathbb{R}$ and $\mathbb{R}^2$ would have same amount of cut points, thus not topologically distinguishable.
Can someone resolve this for me?
The point is that every point of $\mathbb{R}$ is a cut point, while no point of $\mathbb{R}^2$ is; together with the fact that if $f:X\cong Y$ is a homeomorphism and $x$ is a cut point of $X$, then $f(x)$ is a cut point of $Y$. With these two facts in hand, it's clear that $\mathbb{R}\not\cong\mathbb{R}^2$: if $f:\mathbb{R}\cong\mathbb{R}^2$, then $0$ is a cut point of $\mathbb{R}$ but $f(0)$ (whatever that is) isn't a cut point of $\mathbb{R}^2$.
I don't quite understand your second-to-last paragraph, about removing many points at once; can you elaborate?