2
$\begingroup$

In my research I confronted with the following question. I do not have a counterexample, so I ask it here if anyone knows. I would like a positive answer, however a negative answer seems more likely to me.

A complex $C$ is said to be a minimal complex provided every morphism $\beta:C\rightarrow C$ that is homotopic to id$_C$ is an isomorphism.

Original question: Suppose that $(R,\mathfrak{m})$ is a commutative local ring and $\{F_j\}_{j\in\mathbb{N}}$ is a direct system of minimal complexes of finite free modules, each of which a minimal free resolution of $H_0(F_j)$. Then is the direct limit $F=\lim\limits_{\longrightarrow} F_j$ also a minimal complex?

The original question has a couneterexample by the answer below. Later on I modified the question as below.

Modified Question: Suppose that $(R,\mathfrak{m})$ is a commutative local ring and $\{F_j\}_{j\in\mathbb{N}}$ is a direct system of minimal complexes of finite free modules, each of which a minimal free resolution of $H_0(F_j)$, such that $$\lim\limits_{\longrightarrow}H_0(F_j)\neq 0.$$ Then is the direct limit $F=\lim\limits_{\longrightarrow} F_j$ also a minimal complex?

P.S.: The counterexample below definitely can not be changed simply to answer negatively the second modified question, because the main point of the counter-example below is that the resulted direct limit complex is split exact.

$\endgroup$

1 Answer 1

3
$\begingroup$

Here's a "minimal" counterexample. Let $R$ be the localization of $k[x]$ at $(x)$, and let $F_j := R \to^x R$ for every $j$, where $\to^x$ denotes multiplication by $x$. Consider the direct system $$\dots \to F_{j-1} \to^x F_j \to^x F_{j+1}\to \dots$$ Then the colimit is $k(x) \to^x k(x)$ and the $0$ map is homotopic to the identity via $x^{-1}$.

$\endgroup$
7
  • $\begingroup$ Nice! That's correct, thanks. In this counterexample the direct limit morphisms induce zero map on the $0$-th homologies. So I change the question a lit bit to exclude this case. $\endgroup$
    – Aurora
    Jun 12, 2021 at 3:17
  • 1
    $\begingroup$ @Aurora You can modify this example to answer your modified question. $\endgroup$ Jun 12, 2021 at 3:38
  • $\begingroup$ Yes, you are right again. What I am looking for is to put good restrictions to exclude simple counter-examples. Of course I have examples of such direct limits in mind which preserve the minimal property. $\endgroup$
    – Aurora
    Jun 12, 2021 at 3:47
  • $\begingroup$ I'm not sure what a good restriction would be. One philosophical comment is that this counter-example is related to the fact that minimal iff $R/\mathfrak{m} \otimes_R F$ has zero differential breaks down when Nakayama's lemma fails. E.g. $k(x) \otimes_R R/(x) = 0$ even though $k(x) \neq 0$. $\endgroup$ Jun 12, 2021 at 4:02
  • $\begingroup$ Yes, definitely that's the obstruction. Is the new one also easy? $\endgroup$
    – Aurora
    Jun 12, 2021 at 4:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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