Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

In Hulek's Elementary Algebraic Geometry, Nakayama's lemma is stated as follows: Let $A \neq 0$ be a finite $B$-algebra. Then for all proper ideals $m$ of $B$, we have $mA \neq A$. (Here, $A$ and $B$ are both commutative rings with identity.) My question is: how is that related to the usual statements, e.g. Statement 1 in wikipedia: Specifically, does the wikipedia statement imply Hulek's, and if so, how?

Edit: While the statement is cited correctly, it is misleading, because Hulek's definition of what it means for $A$ to be a finite $B$-algebra is that (1) $B\subset A$ is a subring, and (2) there are $a_i \in A$ such that $A=Ba_1+\cdots+Ba_n$.

share|cite|improve this question
up vote 7 down vote accepted

There is something wrong with the result that you cite. E.g. $A = \mathbb Z/3\mathbb Z$ is a finite algebra over $B = \mathbb Z/6\mathbb Z$ (via the natural surjection), and $m = 2B$ is a proper ideal of $B$. However, $m A = 2 A = A$.

What is true is the following: if A is a non-zero finite $B$-algebra, then $A_{\mathfrak m}$ is non-zero for at least one maximal ideal of $B$ (a general property of non-zero modules over a ring), therefore $A/\mathfrak m A$ is non-zero for at least one maximal ideal of $B$ (Nakayama's lemma, in its usual formulation, applied to the finitely generated $B_{\mathfrak m}$-module $A_{\mathfrak m}$), and so $A/m A \neq 0$ for all ideals $m \subset \mathfrak m$.

Perhaps you meant to ask that $B$ be local, so that all proper ideals $m$ of $B$ lie in just a single maximal ideal $\mathfrak m$?

[Added in light of OP's comment below:] Suppose that $B \to A$ is an embedding (as the author cited by the OP apparently does). Then for any maximal ideal $B_{\mathfrak m} \to A_{\mathfrak m}$ is again an embedding, and thus $A_{\mathfrak m}$ is non-zero. Thus by (usual) Nakayama the quotient $A/\mathfrak m A = A_{\mathfrak m }/\mathfrak m A_{\mathfrak m}$ is non-zero, and hence for any proper ideal $m$ of $B$, we have $A/m A$ is non-zero (because $m$ is contained in some maximal ideal $\mathfrak m$).

share|cite|improve this answer
Oh, I'm sorry! Earlier on, his definition of being a finite $B$-algebra of $A$ includes the stipulation that $B$ is a subring of $A$. On the other hand, everything, including the statement of Nakayama's lemma, comes before any mention of local rings, etc. – Cantor's Paradise May 13 '11 at 3:54
@Cantor: Thanks for the clarification. I've added a proof of the claim under this additional assumption. Regards, – Matt E May 13 '11 at 3:58

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.