MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

For any commutative ring $A$, the set of idempotents of $A$ will be denoted as $E(A)$. This set has a (canonical) ring structure. With addition defined by: $$e+'f=e(1−f)+f(1−e)$$ where $+$ and $−$ are operation in the ring itself. The multiplication operation is the same as the ring itself.

Suppose now I have a commutative unital ring $A$ and let $B$ be another commutative ring that is an integral extension of $A$. Then clearly $E(B)$ is an over-ring of $E(A)$, but is it clear that $E(B)$ is an integral extension of $E(A)$. Are there easy counter examples for this? Would it help if I assumed that $B$ is a finite integral extension of $A$?

Edit: The question above had a trivial answer as Todd pointed out. Now Im curious what happens if I really want $E(B)$ to be a finite integral extension of $E(A)$. Does $B$ being a finite integral extension of $A$ gauarantee that?

share|cite|improve this question
I must not be understanding what you mean. Isn't every element $e \in E(B)$ already an integral element, since it satisfies $e^2 - e = 0$? So that $E(B)$ is automatically integral over $E(A)$ even if $B$ isn't integral over $A$? – Todd Trimble Mar 23 '11 at 9:38
Besides, the sum of two idempotents does not have to be idempotent. Thus $E(A)$ is not a ring. – Martin Brandenburg Mar 23 '11 at 10:17
@Todd: You are absolutely right.. the sums just confused me. Consider this question answered @Martin: In the ring of idempotents the sum is defined by $e +' f = e(1-f) + f(1-e)$ where $+$ and $-$ are operation in the ring itself. Multiplication remains the same as the ring itself. So here each element is the additive inverse of itself. – Jose Capco Mar 23 '11 at 10:28
Ah, thanks ! – Martin Brandenburg Mar 23 '11 at 17:30
Jose, you could probably add the definition of $E(A)$ to the body of the question (as opposed to leaving it hidden in the comments) because it is not that standard, as far as I know. – Mariano Suárez-Alvarez Mar 24 '11 at 7:42
up vote 1 down vote accepted

Every element $e$ of $E(B)$ is already integral since it satisfies $e^2 - e = 0$. (This was effectively in a comment above, but perhaps it's better placed as an answer.)

share|cite|improve this answer
Nice, you pwned it! – darij grinberg Mar 24 '11 at 11:38

Take any $B$ such that $E(B)$ is infinite (e.g. $k^\mathbb{N}$, $k$ a field) and write it as a quotient of a domain $A$ (e.g. a polynomial ring over $k$). Then $E(A)=\mathbb{Z}/2\mathbb{Z}$, so $E(B)$ is not finite over $E(A)$.

share|cite|improve this answer
I think the OP is talking about "integral extensions" in the classical sense, where an extension really extends something (meaning, it is injective). – darij grinberg Mar 24 '11 at 10:19
@darij grinberg: good point, but in my example $A$ injects into $A\times B$. – Laurent Moret-Bailly Mar 24 '11 at 13:09

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.