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

Let $V$ be a free $R$-module of finite rank (or, to start, even a finite-dimensional $K$-vector space), $N,P \in \mathrm{End}(V)$ such that $P,N$ commute, $P$ is idempotent, and $N$ is nilpotent. Consider the endomorphism $\alpha=1+(P+N)(t-1)$ of the $R[t]$-module $V[t]$.

Question. Why is the cokernel of $\alpha$ isomorphic to the image of $P$?

This is used in Rosenberg's proof of the Bass-Heller-Swan Theorem, without proof. Actually one only wants that the cokernel of $\alpha$ is finitely generated as an $R$-module. I know how to deal with the case $N=0$, but it doesn't seem to generalize. This should just be linear algebra, but I don't know how to construct a map $\mathrm{coker}(\alpha) \to \mathrm{im}(P)$ at all.

share|cite|improve this question
What is exactly is $V[t]$? – Manos May 23 '12 at 22:18
$V[t]$ is $V\otimes_R R[t]$. More concretely, you are allowing yourself to multiply elements of $v$ by $t$ in a free way. – Aaron May 23 '12 at 23:30
You can imagine elements of $V[t]$ as polynomials in $t$ with coefficients in $V$. This is a $R[t]$-module in the obvious way. So it is possible to handle this without knowledge of tensor products. – Martin Brandenburg May 24 '12 at 6:15
up vote 1 down vote accepted

It can be showed that cokernel of $\alpha$ is a finitely generated $R$-module.

Let us compute the image of $\alpha$ as following:

$\alpha=(1-P-N)+(P+N)t$, $\alpha P=P(-N+(1+N)t)$, $\alpha (1-P)=(1-P)(1-N)+(1-P)Nt=(1-P)(1-N+Nt)$.

Replacing $N(1+N)^{-1}$ by $N$, then the image equals $\alpha PV[t]+\alpha (1-P)V[t]=(t-N)PV[t]+(1-P)V[t]$.

Suppose $N^k=0$, then $(t-N)V[t]\supset (t-N)(t^{k-1}+t^{k-2}N+\ldots +N^{k-1})V[t]=t^kV[t]$, hence the image contains $t^kV[t]$, and the cokernel is a quotient of $V+Vt+\ldots+Vt^{k-1}$ which is a finitely generated $R$-module.

However, we have not proved the original question yet.

share|cite|improve this answer
Thanks! I accept this because Rosenberg's proof has many gaps and errors anyway. I only need that the cokernel is fg / R. – Martin Brandenburg May 24 '12 at 7:00
Actually it's not hard to show with your method of proof the stronger statement that the cokernel is isomorphic to the image of P. – Martin Brandenburg May 24 '12 at 7:16

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.