Toen and Vaquie have constructed for any dg category $\mathcal{C}$ a stack $\mathcal{M}_\mathcal{C}$ parametrising objects in $\mathcal{C}$. Its definition is $$\mathcal{M}_\mathcal{C}(R)\ =\ \text{Maps}_{\text{dgCat}}(\text{Perf} R, \mathcal{C})$$ for $R$ a ring.

  1. It is an Artin stack if $\mathcal{C}$ is of finite type.
  2. The functor $\mathcal{M}_{(-)}$ is right adjoint to the functor taking a 1-stack $X$ to the dg category $\text{Perf}(X)\subseteq \text{QCoh}(X)$.

My question is: if $\mathcal{C}$ is a sheaf of categories* over a space $B$ (scheme, stack, prestack, ... ) can we define a space $$\mathcal{M}_{\mathcal{C}/B}\ \stackrel{?}{\to}\ B,$$ in what generality is it defined (i.e. for an arbitrary dg category $\mathcal{C}$ and prestack $B$?), and are the analogues of 1. and 2. true?

*e.g. if $B$ is affine this is the same thing as a dg category plus an algebra action of $\text{QCoh}(B)$.


1 Answer 1


Let me give the relative construction. We'll say our geometric objects are presheaves on some category $Aff$, and we'll denote sheaves of categories by $2QCoh(-)$.

There's a functor

$$P(Aff)_{/B}^{op} \to 2QCoh(B)$$

$$(f: X \to B) \mapsto Perf(X)$$ sending a presheaf over $B$ to a $Perf(X)$ viewed as a sheaf of categories over $B$, via $f^*$. One shows that this preserves limits, and the right adjoint $M_{(-)/B}$ is a relative moduli of objects, as a presheaf over $B$.

As in the absolute case we have an explicit formula

$$M_{C/B}(Spec(R) \to B) \simeq Map_{2QCoh(B)}(Perf(R), C)$$

where in the first argument on the RHS, $Perf(R)$ is endowed with the structure of a sheaf of categories over $B$ via the map $Spec(R) \to B$. And of course $B= *$ case recovers the absolute case.

I expect arguments of Antieau-Gepner about representability of moduli generalize to this relative setting, though I do not have a precise statement at the moment.

Footnote: As you do in the question statement I'm sweeping some presentability issues under the rug here to give the idea, $2QCoh(-)$ needs to be constructed carefully, but sounds like you know where to look for these constructions.


Your Answer

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

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