Does the ring of continuous functions determine $\mathbb R^n$?

485 Views Asked by At

I have two related questions which are just making the question asked in the title more specific:

(a) Is every ring homomorphism (or maybe $\mathbb R$-algebra homorphism) between rings of the form $\mathscr C(\mathbb R^n),$ ie. rings of continuous functions on Euclidean spaces, automatically continuous?

(b) Does the obvious functor $\mathbb R^n\mapsto \mathscr C(\mathbb R^n)$ define a fully faithful embedding of the full subcategory of topological spaces spanned by $\mathbb R^n$ for all $n,$ into the category of (if needs be topological, depending on what the answer to (a) is) rings/algebras?

I am aware that the answer to both questions is positive when attention is restricted to smooth functions or, when $n$ is even and we can identify $\mathbb R^n$ with $\mathbb C^{n/2},$ to holomorphic or even regular (in the sense of algebraic geometry) functions. However, it is the continuous case which interests me.

1

There are 1 best solutions below

2
On BEST ANSWER

Here are two relevant theorems from Gillman and Jerison, Rings of Continuous Functions, Van Nostrand 1960:

Theorem 8.3. Two realcompact spaces $X$ and $Y$ are homeomorphic if and only if $C(X)$ and $C(Y)$ are isomorphic as rings.

Theorem 10.6. Let $t:C(Y) \to C(X)$ be a ring homomorphism with the property that $t\mathbf{1} = \mathbf{1}$. If $Y$ is realcompact, then there exists a unique continuous mapping $\tau:X \to Y$ such that $tf=f\circ \tau$ for all $f \in C(Y)$.

The definition of realcompactness is a little involved, but every Lindelöf space is realcompact, so all subsets of Euclidean spaces are, too. This shows that the answer to both of your questions is "yes".