I'm trying to understand the proof of a theorem (Auslander-Buchsbaum) which says that given a local ring $(R,m)$, where $m$ is the maximal ideal, and a finitely generated non-zero $R$-module $M$ such that its projective dimension $\operatorname{pd}_R M$ is finite then the following formula holds: $$\operatorname{pd}_R M + \operatorname{depth} M= \operatorname{depth} R.$$

I remind you that the quantity $\operatorname{depth} M$ in general is defined to be the length of a maximal regular sequence $\bar{x}=x_1,\dots,x_n$ on $M$ in $m$ (i.e. $\bar{x}$ is the maximal sequence of elements in $m$ such that (1) $M/(\bar{x})M\neq 0$ and (2) $x_j$ is a non-zero divisor in $M/(x_1,\dots,x_{j-1})M$).

In the proof of this theorem (proven by induction) it is used the fact that the projective dimension $\operatorname{pd}_R M$ is equal to $\operatorname{pd}_{R/xR}M/xM$, where $x$ is not a zerodivisor on $M$ nor on $R$, but I cannot understand why this should be true... Can you help me?


You can again use induction to prove this fact. Base $\operatorname{pd}_R M=0$ is obvious. If $\operatorname{pd}_R M=n>0$, let $F$ be a finitely generated module s.t. we have an epimorphism $F \to M$, denote kernel of this map by $K$, so we have a short exact sequence $$0 \to K \to F \to M \to 0.$$ But $\operatorname{Tor}_1^R(M, R/xR)=\{a \in M : xa=0\}=0$, thus the sequence over $R/xR$ $$0 \to K/xK \to F/xF \to M/xM \to 0$$ is exact.

$\operatorname{pd}_R K=n-1$, and by induction $\operatorname{pd}_{R/xR} {K/xK}=n-1$, module $F/xF$ is free over $R/xR$ thus $\operatorname{pd}_{R/xR} {M/xM}=n$.

  • $\begingroup$ For using the inductionhypothesis $x$ should not be a zerodivisor of $K$. Why is this true? $\endgroup$ – user501184 Feb 12 '18 at 13:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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