Let $\mathcal{C}$ be an abelian category and consider the derived category $D(C)$. Suppose that $F: D(\mathcal{C}) \to D(\mathcal{C})$ is an auto-equivalence.

My question: must $F$ preserve the property of being bounded/bounded above/bounded below.

Said differently, if $A$ is an object of $D(\mathcal{C})$ whose cohomology is bounded/bounded above/bounded below, does $F(A)$ also have cohomology which is bounded/bounded above/bounded below?

If the answer to this question is negative, can one impose natural properties on $F$ to make the answer positive?


Not in that level of generality. For example, let $\mathcal{C}$ be the product of copies of the category of abelian groups, indexed by $\mathbb{Z}$, so an object of $\mathbb{C}$ is a family $\{A_i\mid i\in\mathbb{Z}\}$ of abelian groups, and an object of $D(\mathcal{C})$ is a family $\{X_i\mid i\in\mathbb{Z}\}$ of complexes of abelian groups.

Then $\{X_i\}\mapsto\{X_i[i]\}$ is an auto-equivalence that doesn’t preserve boundedness (or boundedness above or below).

However, it is true if $\mathcal{C}$ is the module category of a ring.

This is because it is well-known (see for example, Proposition 15.72.3 here) that the "perfect complexes" (those isomorphic in $D(\mathcal{C})$ to bounded complexes of finitely generated projective modules) can be characterized in the derived category as the "compact objects" (objects $X$ such that the functor $\text{Hom}_{D(\mathcal{C})}(X,-)$ preserves arbitrary coproducts).

But then the bounded (in cohomology) objects $Y$ of $D(\mathcal{C})$ can be characterized as the objects such that, for each perfect complex $X$, $\text{Hom}_{D(\mathcal{C})}(X,Y[i])=0$ for all but finitely many $i\in\mathbb{Z}$, with similar characterizations of bounded above and bounded below objects.

Since there are inherent characterizations of these various classes of objects, they must be preserved by any auto-equivalence.

  • 1
    $\begingroup$ Thanks a lot for this counterexample! Could I ask you to give an explanation or reference for why the statement is true for the module category of a ring - ideally one which generalizes to other "not too large" categories? (For reference, I am particularly interested in "geometric" examples, such as modules over a ring, coherent sheaves over a scheme, ect.) $\endgroup$ – user142700 Apr 30 at 20:59
  • $\begingroup$ @user142700 I've added an explanation to my answer. $\endgroup$ – Jeremy Rickard May 1 at 9:54

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.