Let $A$ be a subring of $B$. Suppose $B$ is a finitely generated module over $A$. Prove that if $A$ is a Noetherian ring (or Artinian ring), then $B$ is a Noetherian (Artinian) ring.

I am quite confused. My idea:

there exists a surjective $A$-module homomorphism $\phi: A^k\rightarrow B$. Suppose $A$ a Noetherian ring but $B$ not a Noetherian ring. Let ideals of $B$ be $J_1\subset J_2\subset\cdots$. Then $\phi_{-1}(J_1)\subset \phi^{-1}(J_2)\subset\cdots$ are $A$-submodules of $A^k$, moreover, they are also ideals of $A^k$. Since $A$ Noetherian, $A^k$ is Noetherian as a $A$-module. (But we cannot obtain $A^k$ Noetherian as a ring.) How to continue?


Here's a nice way to handle both Noetherian and Artinian cases at once: first notice that if $B$ is Noetherian (resp. Artinian) as an $A$-module, then $B$ is Noetherian (resp. Artinian) as a ring (since any chain of $B$-ideals is a chain of $A$-submodules of $B$). It thus suffices to prove the following:

Proposition: If $A$ is Noetherian (resp. Artinian), then any finite $A$-module $B$ is again Noetherian (resp. Artinian).

Proof: Induct on the number of generators of $B$. If $B$ is cyclic, then $B$ is a quotient of $A$, hence is Noetherian (resp. Artinian). In general, if $B = Ax_1 + \ldots + Ax_n$, then there is an exact sequence

$$0 \to Ax_n \to B \to B/Ax_n \to 0$$

Now $B/Ax_n$ needs only $n-1$ generators, so by induction $Ax_n, B/Ax_n$ are Noetherian (resp. Artinian), but then so is $B$.

By the way, the converse is also true: if $A \subseteq B$ as rings, and $B$ is a finite $A$-module, then $A$ is Noetherian iff $B$ is Noetherian (as rings): this is Eakin's theorem.

| cite | improve this answer | |

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.