5
$\begingroup$

Let $B\leftarrow A\to C$ be a diagram of commutative rings, and let $\mathcal{D}(A)$ be the derived $\infty$-category of $A$-modules (as in Lurie's "Higher Algebra"). Then is there an equivalence $$\mathcal{D}(B\otimes_A^LC):=\operatorname{Mod}_{B\otimes_A^LC}(\mathrm{Sp})\simeq\mathcal{D}(B)\otimes_{\mathcal{D}(A)}\mathcal{D}(C)$$ (with tensor products taken appropriately, perhaps in $\operatorname{Cat}_\infty(\mathcal{K})$ or $\mathrm{Pr}^L_{st}$)?

I am aware that if we omit the base ring $A,$ this holds true. More precisely, for any symmetric monoidal $\infty$-categories $\mathcal{C}$ and $\mathcal{D}$ and commutative algebra objects $B\in\operatorname{CAlg}(\mathcal{C}),$ $C\in\operatorname{CAlg}(\mathcal{D}),$ the natural map $$ \operatorname{Mod}_{B}(\mathcal{C})\otimes\operatorname{Mod}_C(\mathcal{D})\to\operatorname{Mod}_{B\otimes C}(\mathcal{C}\otimes\mathcal{D}) $$ is an equivalence (where the tensor products are taken in $\operatorname{Cat}_\infty(\mathcal{K})$). In the special case $\mathcal{C} = \mathcal{D} = \mathrm{Sp},$ we get $$ \operatorname{Mod}_{B}(\mathrm{Sp})\otimes\operatorname{Mod}_C(\mathrm{Sp})\to\operatorname{Mod}_{B\otimes C}(\mathrm{Sp}\otimes\mathrm{Sp})\simeq\operatorname{Mod}_{B\otimes C}(\mathrm{Sp}). $$ However, I'm not sure how to bootstrap to the relative case from here.

If this is true, I would appreciate a reference or proof, ideally one which is as hands-on as possible so I can understand these objects better. While I want to know if this holds in the generality above, if there's a more intuitive or explicit way to see this at the level of DG-categories, I would also be interested in that.

$\endgroup$
1
  • 6
    $\begingroup$ The functor $A\mapsto\mathrm{Mod}_A(\mathrm{Sp})$ from $E_\infty$-rings to presentable symmetric monoidal $\infty$-categories commutes with all colimits, since it is left adjoint to the functor sending $\mathcal{C}$ to $\operatorname{End}_{\mathcal{C}}(1)$. The problem now is to find a reference for this "well-known" fact... $\endgroup$ Commented Aug 27, 2020 at 11:37

1 Answer 1

14
$\begingroup$

A hands-on explanation: Relative tensor products like $B\otimes_AC$ are computed as the colimit of the simplicial object $B\otimes A^{\otimes \bullet} \otimes C$. The functor $\mathsf{Mod}_{(-)}: \mathsf{Alg}(\mathsf{Sp}) \to \mathsf{Pr}^{L, \mathrm{st}}_{\mathsf{Sp}/}$ preserves all colimits and is symmetric monoidal, so apply $\mathsf{Mod}_{(-)}$ everywhere and we get the formula for the relative tensor product of categories of modules.

A 'reference-heavy' explanation:

The reference that $A \mapsto \mathsf{Mod}_A$ is symmetric monoidal is HA.4.8.5.16. And, the statement about having a right adjoint (when valued in 'presentable stable $\infty$-categories with a distinguished object') is HA.4.8.5.11.

So now we just need a general thing about when $F: \mathcal{C} \to \mathcal{D}$ being a symmetric monoidal left adjoint induces a colimit-preserving functor on $\mathsf{CAlg}(-)$, but that'll be true when the tensor product in $\mathcal{C}$ and $\mathcal{D}$ distribute over colimits- then we get (i) $\mathsf{CAlg}(F)$ preserves coproducts since those are tensor products, and (ii) $\mathsf{CAlg}(F)$ preserves sifted colimits since those are computed on underlying objects (HA.3.2.3.2). Thus $\mathsf{CAlg}(F)$ preserves all colimits, and, in particular, pushouts.

$\endgroup$
2
  • $\begingroup$ Thanks Dylan, this is very helpful! While this certainly answers my question, I must admit I was hoping for something like being able to explicitly show that every module over $B\otimes_A C$ is equivalent in some homotopy appropriate sense to a "compatible pair" $(M,N)\in\operatorname{Mod}_{B}\times\operatorname{Mod}_{C}.$ Maybe this is as hands-on as it gets in the $\infty$-category world, though... $\endgroup$
    – Stahl
    Commented Aug 30, 2020 at 8:11
  • 1
    $\begingroup$ @Stahl you can identify objects of $\mathsf{Mod}_B \otimes_{\mathsf{Mod}_A}\mathsf{Mod}_C$ with $B$-$C$-bimodule objects in the category of $A$-modules. Is that more along the lines of what you were looking for? (This is a special case of HA.4.8.4.6). $\endgroup$ Commented Aug 30, 2020 at 14:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .