Let $R$ be (not necessarily commutative) ring and $S$ a simple right $R$-module. Let $f\in Ann(S)$ be normalizng and a non-zero divisor. Is it always true that $$ pdim_{R}(S)=pdim_{R/(f)}(S)+1? $$
$\begingroup$
$\endgroup$
4
-
$\begingroup$ What does "normalizng" mean in this context? $\endgroup$– Jason StarrCommented Aug 13, 2012 at 14:42
-
$\begingroup$ We say that $f\in R$ is normalizing if $fR=Rf$ holds. $\endgroup$– M SimonCommented Aug 13, 2012 at 17:17
-
3$\begingroup$ See Theorem 7.3.5(i) of the book by McConnell and Robson. $\endgroup$– user91132Commented Aug 13, 2012 at 19:32
-
$\begingroup$ Thanks, Konstantin! This is what I was looking for. I thought that simplicity of $S$ was required, but it turns out not. $\endgroup$– M SimonCommented Aug 13, 2012 at 21:42
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
1
With the definition of normalizing you give, it is not always the case that the projective dimension of $S$ as an $R$-module equals $1$ more than the projective dimension of $S$ as an $R/\langle f \rangle$-module. Let $R$ be $\mathbb{Z}$, let $f$ be $p^2$ for some prime $p$, and let $S$ be $\mathbb{Z}/p\mathbb{Z}$. The projective dimension of $S$ as an $R$-module is $1$, but the projective dimension of $S$ as an $R/\langle f \rangle$-module is infinite.
-
$\begingroup$ Thanks for the counterexample, Jason! I also realized that my statement was not not true without assumption $pdim_{R/(f)}(S)<\infty$. I was slow to correct my claim. This is an educational counterexaple. $\endgroup$– M SimonCommented Aug 13, 2012 at 21:46