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

I'm trying to prove that the following statements are equivalent for a commutative ring $R$

A: $_RR=_RN \oplus_RM$ for some submodules $_RN$, $_RM\subseteq_RR$

B: There exists an element $e=e^2\in R$ such that $N=Re$ and $M=R(1-e)$

I have no idea how to show $A \implies B$, but $B\implies A$ seems easy: every $r \in R$ can be written as $re+r(1-e)$ and for $r\in N\cap M$ $r=ae=b(1-e)$ for some $a,b\in R$, so $re=ae^2=b(1-e)e=b(e-e^2)=0$. Now if $R$ is a domain, $e=0$ and $M=R$ or $r=0$ and $N\cap M=\emptyset$. But what if it's not a domain?

share|cite|improve this question
In your work, you show that $ae^2 = 0$. And $ae^2 = ae = r$. – Dylan Moreland Jun 28 '12 at 19:18
You might consider dropping or swapping the commutative algebra tag... it's completely true for noncommutative rings with unity, also. – rschwieb Jun 28 '12 at 19:31
Thank you for this comment, I've changed the tag to abstract algebra. – PhillyPhil Jun 28 '12 at 20:51
up vote 2 down vote accepted

For $B\implies A$, since $(1-e)e=0$, you see that $ae\in R(1-e)$ implies $ae=ae^2=0$. Thus the intersection is zero.

For $A\implies B$, look at $1=m+n\in M\oplus N$. Clearly n=1-m. So $m$ and $n$ are pretty good candidates for idempotents: check!

(Hint: one might start by multiplying $1=m+n$ on the left and right by $n$ to prove something about $mn$ and $nm$.)

share|cite|improve this answer
Thank you! So $m \in M, 1-m \in N \Rightarrow (1-m)m=m(1-m) \in M \cap N=0 \Rightarrow m=m^2$. Is it important, that we look at $R$, $M$ and $N$ as $R$-modules? These arguments work also if we treat $R$, $M$ and $N$ as rings, right? So is it true, that $_RR=_RN \oplus_RM \iff R=N \oplus M$? – PhillyPhil Jun 28 '12 at 22:17
@PhillyPhil Nice, your method is even better! About your generalization: not exactly. A ring decomposition $R=S\oplus T$ corresponds to a pair of central idempotents. The proper generalization of your question is this: for any module $_RM$, $M=N\oplus P$ iff there is an idempotent $e\in End(_RM)$ such that $e(M)=N$ and $(1-e)(M)=P$. Since $End(_RR)\cong R$ acting by right multiplication, you retrieve your question. – rschwieb Jun 28 '12 at 23:49

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.