2

Does there exist an integral Noetherian domain $R$ such that there is an injective unital ring homomorphism $R[[x]]\rightarrow R$?

I thought of using Krull dimension to answer this but a ring can have lower Krull dimension than its subring (e.g. $\mathbb{Z}\subset \mathbb{Q}$).

  • $R=K[[Y]]$, and define the morphism $f(X,Y) \in R[[X]]=K[[X,Y]] \longmapsto f(Y,Y)$? – Aphelli Jul 25 '19 at 09:28
  • @Mindlack I don't think your morphism is injective. –  Jul 25 '19 at 09:32
  • You’re right, I had missed that part. – Aphelli Jul 25 '19 at 09:44
  • This is not possible if $R$ is countable since $R[[x]]$ is uncountable. I think the same is true if $R$ is uncountable because then $R[[X]]$ is somehow more uncountable than $R$. But I'll refrain from asserting certainty. – Parthiv Basu Jul 25 '19 at 10:35
  • 2
    @Parthiv Basu: I am afraid it is not as simple. $R[[X]]$ has cardinality $|R|^{|\mathbb{N}|}$, which is, if $R$ has cardinality, say, that of $\mathbb{R}$, equipotent to $R$. – Aphelli Jul 25 '19 at 10:40

1 Answers1

2

I think the answer is yes, with $R= \mathbb C$ (which is surely noetherian and integral !)

Indeed consider $K$ the field of fractions of $\mathbb C[[x]]$ and $L$ its algebraic closure.

Note that $\mathbb C [[x]]$ has cardinality $|\mathbb C|^{|\mathbb N|} = (2^{\aleph_0})^{\aleph_0} = 2^{\aleph_0^2}=2^{\aleph_0} = |\mathbb C|$, therefore so does $K$ and therefore so does $L$.

It follows that the transcendence degree of $L$ over $\mathbb Q$ is precisely $2^{\aleph_0}$, just like $\mathbb C$; and $L$ is algebraically closed, therefore $L\cong \mathbb C$ (it's actually quicker to see that $L$ injects into $\mathbb C$ but the stronger result is true as well).

But clearly $ \mathbb C[[x]] \to K \to L \to \mathbb C$ is then an injection.

This is the same trick as to show that $\mathbb C(X)$ injects into $\mathbb C$.

Of course, as was pointed out in the comments, for cardinality reasons this doesn't happen if $|R|^{\aleph_0} > |R|$, which happens e.g. for $|R| = \aleph_0$, i.e. $R$ countable (but also, by König's theorem, for higher cardinalities such as $\aleph_\omega$, though of course you will rarely encounter rings that provably have this cardinality)

user26857
  • 52,094
Maxime Ramzi
  • 43,598
  • 3
  • 29
  • 104