My question is:

Has anyone constructed an $(\infty,2)$-category whose objects are (projective, maybe smooth, ...) varieties, and where the 1-morphisms from $X$ to $Y$ are given by $D^b_\infty\text{Coh}(X \times Y)$ (an $(\infty,1)$-enhancement of $D^b\text{Coh}(X \times Y)$)?

By "$(\infty,1)$-enhancement" I mean some $(\infty,1)$-category whose homotopy category is $D^b\text{Coh}$.

I would hope that binary composition would descend to the functor $$D^b\text{Coh}(Y\times Z) \times D^b\text{Coh}(X \times Y) \to D^b\text{Coh}(X \times Z)$$ that sends $(Q,P)$ to $\pi_{02,*}(\pi_{01}^*P \otimes \pi_{12}^*Q)$ a la the Fourier--Mukai transform (where tensor and the projections are the derived functors).

If the answer is "yes", I wonder whether this is special to derived categories of coherent sheaves or whether it's a more general algebraic result about collections of $(\infty,1)$-categories of a certain kind?

I'm new to infinity infinity stuff and I might have misused a technical term, so it might be best not to interpret my words too literally.


For smooth projective varieties it is known that $$D^b_\infty Coh(X\times Y)\simeq Hom(D^b_\infty Coh(X), D^b_\infty Coh(Y))$$ compatibly with the composition you describe (note here $D^b_\infty Coh$ coincides with the $\infty$-category $Perf$ of perfect complexes). Hence your desired $(\infty,2)$-category is a full subcategory of the $(\infty,2)$-category of dg categories (or small idempotent-complete stable $\infty$-categories) where the above Homs take place. The analogous result for $QCoh$ holds in much greater generality (without smoothness or projectivity - eg for quasicompact quasiseparated schemes).

Edit: Without smoothness, you find (here) that $D^b_\infty Coh(X\times Y)$ represents functors from $Perf(X)$ to $D^b_\infty Coh(Y)$, and that functors from $D^b_\infty Coh(X)$ to $D^b_\infty Coh(Y)$ are represented by kernels which are coherent relative to the first factor. Note you have to be careful in the formulation of the question since $D^b_\infty Coh$ is not preserved by tensor product on a singular variety! (e.g. self-tor of a skyscraper at a singular point is not bounded, i.e., not in $D^b_\infty Coh$)

  • $\begingroup$ Thanks! Do you know anything about what happens when the "smoothness" hypothesis is dropped? $\endgroup$ – Nathaniel Bottman Sep 23 '14 at 2:27
  • $\begingroup$ Answered in form of edit.. $\endgroup$ – David Ben-Zvi Sep 23 '14 at 3:54

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.