I apologize in advance if this is pretty straightforward; I'm a differential geometer and physicist by training so my homological algebra and homotopy theory are a bit weak.

Question: Let $M$ be a smooth manifold, $\Omega^\bullet_M$ the de Rham complex of sheaves, and $\mathbb{R}_M$ the constant sheaf corresponding to $\mathbb{R}$. Does the tensor product of chain complexes of sheaves $(-) \otimes \Omega_M$ with the de Rham complex give a functorial soft resolution of the category $\mathsf{Chain}(\mathbb{R}_M \mathsf{Mod})$ of chain complexes of sheaves of vector spaces on $M$?

If this is true, does anyone happen to know of a reference for this?

If this is not true in general, is it true for some appropriate subcategory, such as bounded chain complexes, or bounded below chain complexes? I suspect that in fact this will be the case for bounded below complexes of sheaves.

Here's my (partial) reasoning for why I suspect the proposition in the question should be true:

I remember reading somewhere (probably in Kashiwara & Schapira, but can't find the result at the moment) that if a sheaf $\mathcal{F}$ is c-soft and flat, then $\mathcal{F} \otimes \mathcal{A}$ is c-soft for any other sheaf $\mathcal{A}$. I assume this should still be true if we replace the words "sheaf" by "chain complex of sheaves", so that both $\mathcal{A}^\bullet$ and $\mathcal{F}^\bullet$ are complexes of sheaves, and $\mathcal{F}^\bullet$ is degree-wise flat and c-soft.

Assuming this is true, this should give the result I want, since sheaves of vector spaces are always flat, and so the de Rham complex is degree-wise c-soft and flat. In detail, we have a quasi-isomorphism $\mathbb{R}_M \xrightarrow{\sim} \Omega_M$ by the usual de Rham resolution. Since tensoring sheaves of vector spaces is exact, then for any complex of sheaves of vector spaces $\mathcal{A}^\bullet \in \mathsf{Chain}(\mathbb{R}_M \mathsf{Mod})$ tensoring with $\mathcal{A}^\bullet$ should preserve the quasi-isomorphism, and hence give us a quasi-isomorphism $\mathcal{A}^\bullet \cong \mathcal{A}^\bullet \otimes \mathbb{R}_M \xrightarrow{\sim} \Omega_M^\bullet \otimes \mathcal{A}^\bullet$.

Is this reasoning correct?

For some motivation about why it would be interesting to know if this is true, in Riehl's Categorical Homotopy Theory text, she shows that given a functorial right deformation $R: \mathsf{M} \to \mathsf{M}$, $r: 1 \overset{\sim}{\Rightarrow} R$ for a functor $F: \mathsf{M} \to \mathsf{N}$ between homotopical categories, we can compute a point-set right derived functor $\mathbb{R}F = F \circ R: \mathsf{M} \to \mathsf{N}$. This point-set derived functor depends on the choice of deformation, but descends to the usual (deformation-independent) total derived functor $\mathbf{R}F: \mathrm{Ho}(\mathsf{M}) \to \mathrm{Ho}(\mathsf{N})$ between the homotopy categories.

In my case we would have $\mathsf{M} = \mathsf{Chain}(\mathbb{R}_M \mathsf{Mod})$ and we would want to compute point-set derived functors for $f_!$ and $\Gamma(M,-)$ using the de Rham complex.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.