I can not understand Remark 12.8.8 in the preprint "SINGULAR SUPPORT OF COHERENT SHEAVES AND THE GEOMETRIC LANGLANDS CONJECTURE". I am somewhat embarrased by the degree of my confusion, hopefully someone knowledgeable could help me.

Authors claim that for any (connected) algebraic stack $Y$ and any compact object $M$ of the DG category of D-modules on $Y$ the map $$ H_{dR}(Y)\otimes M\rightarrow M \qquad (*) $$ vanishes on a sufficiently high power of the augmentation ideal of $H_{dR}(Y)$.

The first difficulty is that $H_{dR}$, I believe, was only defined for the stack $pt/G$ where $G$ is a connected reductive group over an algebraically closed field of characteristic 0. So I am not completely sure what is that supposed to mean for other stacks.

The second difficulty is what would the map (*) be.

The third difficulty is why do we talk about augmentation ideal. Is it true that the image of the map $H_{dR}(pt/G)\rightarrow H_{dR}(Bun_G)$ lies in the augmentation ideal?

The fourth difficulty is how do we actually show that the above claim is true?

  • 1
    $\begingroup$ A title related to the mathematical contents would be better. $\endgroup$
    – YCor
    Dec 12, 2018 at 21:55

1 Answer 1


The answers to your questions can be found in this article: https://arxiv.org/abs/1108.5351. I highly recommend reading it before trying to understand Arinkin-Gaitsgory.

Let me try to resolve your difficulties. I will take for granted the existence of a dg (or equivalent, stable $k$-linear infinity-) category of $\mathcal{D}$-modules on $Y$, satisfying standard functorialities ($!$-pullback and $*$-pushforward). Denote this category by $\mathcal{D}(Y).$

Taking $!$-pullback along the map $Y\rightarrow \operatorname{Spec}k$, we get a dualizing object $\omega_Y$ in the category of $\mathcal{D}$-modules. This is the unit for a natural monoidal $\otimes^!$ structure on $\mathcal{D}(Y)$, defined via the formula $\mathcal{F}\otimes^!\mathcal{G}\cong\Delta^!(\mathcal{F}\boxtimes\mathcal{G}),$ where $\Delta:Y\rightarrow Y\times Y$ is the diagonal map. A warning: For stacks, even one as simple as $Y\cong \operatorname{pt}/\mathbb{G}_m$, $\omega_Y$ often fails to be compact.

Regarding your first difficulty: You can define $H^{\bullet}_{\operatorname{dR}}(Y)$ to be $\operatorname{Hom}^{\bullet}(\omega_Y,\omega_Y).$ To understand why this is a reasonable definition, check that it agrees with the usual one for $Y$ a smooth variety.

Regarding your second difficulty: Because $\omega_Y$ is the unit for the $\otimes^!$ monoidal structure, there is a map $\operatorname{Hom}^{\bullet}(\omega_Y,\omega_Y)\rightarrow\operatorname{Hom}^{\bullet}(\omega_Y\otimes^! M,\omega_Y\otimes^! M)\cong\operatorname{Hom}^{\bullet}(M,M),$ which leads to your map $(*).$

Unfortunately, I don't understand your third difficulty.

Regarding your fourth difficulty: Look at the Drinfeld-Gaitsgory paper I linked above. One of their main theorems is that $\mathcal{D}(Y)$ is compactly generated by induced $\mathcal{D}$-modules. This implies that it suffices to check your claim for $M$ induced, which I believe is a straightforward computation, possibly using their Lemma 2.1.4. (Unfortunately I have to run right now - I'll try and write the details later.)

  • $\begingroup$ regarding the third difficulty, I think I understand now. The augmentation Arinkin-Gaitsgory have in mind is the projection from $Hom(\omega_Y, \omega_Y)$ to its 0-th graded component. So for any positive $n$, $H^{2n}_{dR}(Y)$ lies in the augmentation ideal. Do you think this is correct? Anyway, thank you very much for helping me. $\endgroup$
    – geometer
    Dec 13, 2018 at 6:26
  • $\begingroup$ @geometer: Yes, that is correct. $\endgroup$
    – dhy
    Dec 13, 2018 at 7:25

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.