2
$\begingroup$

I don't know if I'm actually using the right terminology here, to be clear I'm going to state explicitly what I'm trying to figure out to see if I can be pointed in the right direction:

Let $F: \mathcal{A} \rightarrow \mathcal{B}$ be a left exact functor between 2 abelian categories $\mathcal{A}$ and $\mathcal{B}$, let $X^\bullet$ be the complex $0 \rightarrow X^0 \rightarrow X^1 \rightarrow X^2 \rightarrow X^3 \rightarrow \cdots$ of objects in $\mathcal{A}$, let $T^\bullet$ be a complex obtained by removing the FIRST TERM (or maybe the first $q$ terms from the left?), I mean let $T^\bullet$ be $X^0 \rightarrow X^1 \rightarrow X^2 \rightarrow X^3 \rightarrow \cdots$, the complex obtained by removing the $0$ from $X^\bullet$, NO 0's to the left, so $T$ becomes "unbounded". Let's say I have a way of obtaining the derived functor $RF(T^\bullet)$ using the machinery developed for unbounded complexes. Is there any way I can compute some of the terms in $RF(T^\bullet)$ from knowing this?

$\endgroup$
5
  • $\begingroup$ As an extreme case, imagine the initial complex has exactly one term! :-) $\endgroup$ Sep 14, 2012 at 20:16
  • $\begingroup$ If the domain category has finite global dimension $d$, for example, perturbations in a complex only travel $d$ (or maybe $d+1$?) places. If you go through the steps of constructing a projective resolution of your initial complex and then applying the functor to that, you can see this at once. $\endgroup$ Sep 14, 2012 at 20:18
  • $\begingroup$ Sorry I'm a bit lost here, what's a perturbation in this context? Thanks for answering $\endgroup$
    – Louis A
    Sep 14, 2012 at 20:40
  • $\begingroup$ By perturbation I simply mean a change one of the terms in the complex (for example, replacing it with zero) $\endgroup$ Sep 14, 2012 at 20:44
  • $\begingroup$ This is the first time I've heard the term perturbation in this context, so I don't know what it means for a perturbartion to travel $d$ places, apply the functor to what? To the perturbation? To see what at once? Sorry if I sound clueless but I'm totally unfamiliar with what you're saying $\endgroup$
    – Louis A
    Sep 14, 2012 at 21:56

2 Answers 2

1
$\begingroup$

Suppose for simplicity $T$ and $S$ are two bounded complexes supported in negative degrees which differ only in their $0$th term. Construct Cartan-Eilenberg projective resolutions $P_T$ and $P_S$ for them, so that $RF(T)$ and $RF(S)$ are in fact $F(\operatorname{Tot} P_T)$ and $F(\operatorname{Tot} P_S)$. Now, in view of the way one constructs C-E resolutions, $P_T$ and $P_S$ can be chosen in such a way that they only differ in specific places (the terms appearing in the $0$th column and in the differentials which have those terms as codomains, I think). If the global dimension $d$ of $\mathcal A$ is finite, and if one uses finite resolutions for objects of length at most $d$, the complexes $F(\operatorname{Tot} P_T)$ and $F(\operatorname{Tot} P_S)$ therefore differ in at most $d+1$ places.

$\endgroup$
4
  • 1
    $\begingroup$ More generally, if the complex $T$ differs from $S$ only in the terms with degree in $\{-r,\dots,0\}$ and the global dimension of the domain category is $d$, then $RF(T)$ and $RF(S)$ will coincide in all degrees smaller that $-r-d-1$ (plus or minus 1...) $\endgroup$ Sep 14, 2012 at 22:19
  • $\begingroup$ Thank you! Are there any books or references for this? What's the global dimension of a category? I know what's the global dimension of a module, I found this: bit.ly/O4phq7 $\endgroup$
    – Louis A
    Sep 15, 2012 at 0:18
  • $\begingroup$ Do you have a reference for this, I can't find it on Google, please, there has to be a reference, I'm in a hurry, I'm gonna get thrown out of school if I don't hurry up $\endgroup$
    – Louis A
    Sep 15, 2012 at 1:25
  • $\begingroup$ For left exact functors one usually uses injective resolutions. Projective resolutions work for right exact functors. $\endgroup$
    – Sasha
    Sep 15, 2012 at 2:52
1
$\begingroup$

$T$ is called the stupid truncation of $X$ (in a contrast with the canonical or smart truncation).

In the derived category $D(A)$ there is a distinguished triangle $$ T \to X \to R, $$ where $R$ is the complex consisted of the ONE term (or $q$ terms) of $X$. Applying the derived functor $RF$ you get a distinguished triangle $$ RF(T) \to RF(X) \to RF(R) $$ in $D(B)$. So, $RF(R)$ measures the difference between $RF(T)$ and $RF(X)$. In particular there is a long exact sequence $$ \dots \to R^nF(T) \to R^nF(X) \to R^nF(R) \to R^{n+1}F(T) \to \dots $$ and if you know $R^nF(R)$ you can compare $R^nF(X)$ and $R^nF(T)$.

$\endgroup$
2
  • $\begingroup$ Thank you so much! One thing though: The thing is this, complex X is $0 \rightarrow X^1 \rightarrow X^2 \rightarrow X^3 \rightarrow \cdots$ and complex T is $X^1 \rightarrow X^2 \rightarrow X^3 \rightarrow X^4 \rightarrow \cdots$, the same complex starting at $X^1$ so I'm treating it as an unbounded complex, I did it that way because removing the $0$'s from $X$ makes the complex exact and easier to work with $\endgroup$
    – Louis A
    Sep 15, 2012 at 4:34
  • $\begingroup$ Sorry to bother you again, but isn't a stupid truncation the one where you add 0's to the left? I don't know if I'm not understanding this right but my complex $T$ is "unbounded", it doesn't have 0's to the left, $T$ is $X^0 \rightarrow X^1 \rightarrow X^2 \rightarrow X^3 \rightarrow \cdots$, NOT $0 \rightarrow 0 \rightarrow X^0 \rightarrow X^1 \rightarrow X^2 \rightarrow X^3 \rightarrow \cdots$, don't know if I'm confused there $\endgroup$
    – Louis A
    Sep 21, 2012 at 17:34

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.