1 Introduction
1.1
In this article we study the degeneration of a Whittaker sheaf on Drinfeld’s compactification to an object of the principal series category (for us ‘sheaf’ will mean D-module, but our results carry over mutatis mutandi to $\ell$ -adic sheaves in characteristic $p>0$ for $\ell \neq p$ ). In [Reference Arinkin and GaitsgoryAG15b] this degeneration is implemented by a gluing functor, meaning $!$ -extension from the general locus followed by $!$ -restriction to the special fiber. This operation produces a complex of sheaves, which the authors of [Reference Arinkin and GaitsgoryAG15b] link to the constant term of Poincaré series.
Nearby cycles provide a different, although closely related, method of degeneration. Notably, if we apply nearby cycles (more precisely, the geometric Jacquet functor of [Reference Emerton, Nadler and VilonenENV04]) to the perverse cohomological shift of a nondegenerate Whittaker sheaf, the result is still a perverse sheaf. It comes equipped with a nilpotent endomorphism, which gives rise to the so-called monodromy filtration. We describe the restrictions of this nearby cycles sheaf to the defect strata in terms of the Langlands dual Lie algebra, showing in particular that these restrictions are perverse, and hence that the nearby cycles sheaf is tilting. We also describe the associated graded sheaf of the monodromy filtration on nearby cycles along with its Lefschetz $\mathfrak{sl}_{2}$ -action, in terms of the Picard–Lefschetz oscillators, which are certain factorizable perverse sheaves with $\mathfrak{sl}_{2}$ -action introduced in [Reference SchiederSch18].
In [Reference CampbellCam17], the author introduced a similar construction in the situation of a finite-dimensional flag variety. Namely, we showed that the nearby cycles of a one-parameter family of nondegenerate Whittaker sheaves on a flag variety is the big projective sheaf, which is isomorphic to the tilting extension of the constant perverse sheaf on the big cell. Thus in both cases taking nearby cycles of Whittaker sheaves produces tilting sheaves.
1.2
Fix an algebraically closed field $k$ of characteristic zero and a smooth connected curve $X$ over $k$ . Let $G$ be a connected reductive group over $k$ . We assume for simplicity that the derived subgroup $[G,G]$ is simply connected. We write $I$ for the set of vertices of the Dynkin diagram of $G$ , and $Z_{G}$ for the center. Fix a Borel subgroup $B$ with unipotent radical $N$ , and put $T:=B/N$ .
For $\mathscr{Y}$ an algebraic stack locally of finite type, we write $D(\mathscr{Y})$ for the unbounded derived category of D-modules on $\mathscr{Y}$ and denote by $D_{\operatorname{hol}}(\mathscr{Y})$ the full subcategory consisting of bounded complexes with holonomic cohomologies. The reader is free to view $D(\mathscr{Y})$ as a triangulated category for most of the paper, with the exception of § 3, where the differential graded (DG) enhancement of $D(\mathscr{Y})$ is used.
We denote by $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ Drinfeld’s compactification of the moduli stack $\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}$ of canonically twisted $N$ -bundles. A choice of isomorphism $N/[N,N]\cong \mathbb{G}_{a}^{I}$ gives rise to a map
constructed for example in [Reference Frenkel, Gaitsgory and VilonenFGV01, § 4.1].
Let $\unicode[STIX]{x1D712}$ be a nontrivial exponential D-module on $\mathbb{G}_{a}$ . As in [Reference Frenkel, Gaitsgory and VilonenFGV01], one shows that $\operatorname{ev}^{\unicode[STIX]{x1D6E5}}\unicode[STIX]{x1D712}$ extends cleanly to a perverse sheaf $\mathscr{W}_{1}$ on $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ , where $\operatorname{ev}^{\unicode[STIX]{x1D6E5}}$ denotes cohomologically normalized inverse image along the smooth morphism $\operatorname{ev}$ . The sheaf $\mathscr{W}_{1}$ has irregular singularities because $\unicode[STIX]{x1D712}$ does, so, to be clear, by ‘perverse sheaf’ we mean any holonomic D-module.
Choose a dominant regular cocharacter $\unicode[STIX]{x1D6FE}:\mathbb{G}_{m}\rightarrow T$ , which determines an action
We will denote by $\mathscr{W}$ the perverse sheaf $a_{\unicode[STIX]{x1D6FE}}^{\unicode[STIX]{x1D6E5}}\mathscr{W}_{1}$ on $\mathbb{G}_{m}\times \overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ . Consider the embeddings
The $\mathbb{G}_{m}$ -equivariant object $i^{!}j_{!}\mathscr{W}$ of $D(\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}})$ is studied in [Reference Arinkin and GaitsgoryAG15b].
We will consider instead a closely related perverse sheaf on $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ , namely the nearby cycles $\unicode[STIX]{x1D6F9}(\mathscr{W})$ with respect to the projection $\mathbb{A}^{1}\times \overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}\rightarrow \mathbb{A}^{1}$ . This sheaf is unipotently $\mathbb{G}_{m}$ -monodromic, and (the logarithm of) the monodromy endomorphism of nearby cycles agrees with the obstruction to $\mathbb{G}_{m}$ -equivariance, a nilpotent endomorphism. The composition $\unicode[STIX]{x1D6F9}\circ a_{\unicode[STIX]{x1D6FE}}^{\unicode[STIX]{x1D6E5}}$ for a $\mathbb{G}_{m}$ -equivariant one-parameter family is studied in [Reference Emerton, Nadler and VilonenENV04] under the name geometric Jacquet functor.
Up to isomorphism $\unicode[STIX]{x1D6F9}(\mathscr{W})$ does not depend on the choice of $\unicode[STIX]{x1D712}$ . In particular $\unicode[STIX]{x1D6F9}(\mathscr{W})$ is Verdier self-dual, since $\unicode[STIX]{x1D6F9}$ commutes with Verdier duality and the Verdier dual of $\mathscr{W}$ is the clean extension of $\operatorname{ev}^{\unicode[STIX]{x1D6E5}}\unicode[STIX]{x1D712}^{-1}$ .
The object $i^{!}j_{!}\mathscr{W}$ can be recovered from $\unicode[STIX]{x1D6F9}(\mathscr{W})$ as the derived invariants of monodromy, and conversely $\unicode[STIX]{x1D6F9}(\mathscr{W})$ is the derived coinvariants of the natural $H^{\bullet }(\mathbb{G}_{m})$ -action on $i^{!}j_{!}\mathscr{W}$ (see Definition 3.1.1 and Proposition 3.2.2). One can summarize by saying that nearby cycles and the ‘gluing functor’ $i^{!}j_{!}$ are related by Koszul duality.
From the standpoint of the geometric Langlands program this Koszul duality motivates the study of $\unicode[STIX]{x1D6F9}(\mathscr{W})$ , since the gluing functors play a crucial role in the proof of the geometric Langlands equivalence sketched in [Reference GaitsgoryGai15]. Namely, they are used to achieve a spectral description of the so-called extended Whittaker category. Nearby cycles has the advantage over the gluing functor of being t-exact, so that $\unicode[STIX]{x1D6F9}(\mathscr{W})$ , unlike $i^{!}j_{!}\mathscr{W}$ , is a perverse sheaf.
1.3
The stack $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ has a stratification by defect of the generalized $N$ -bundle. Our first main theorem is a description of the restrictions of $\unicode[STIX]{x1D6F9}(\mathscr{W})$ to the strata. In particular we will prove that this sheaf is tilting with respect to the defect stratification, meaning its $!$ - and $\ast$ -restrictions to the strata are perverse (although generally not lisse).
Write $\unicode[STIX]{x1D6EC}$ for the lattice of cocharacters of $T$ , and denote by $\unicode[STIX]{x1D6EC}^{\operatorname{pos}}\subset \unicode[STIX]{x1D6EC}$ the positive coweights with respect to $B$ . The defect stratification of $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ is indexed by $\unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ , and for each $\unicode[STIX]{x1D707}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ we denote the locally closed stratum embedding by
Denote by $X^{(n)}$ the $n$ -fold symmetrized power of $X$ . If $\unicode[STIX]{x1D707}=\sum _{i\in I}n_{i}\unicode[STIX]{x1D6FC}_{i}$ for some nonnegative integers $n_{i}$ (here $\unicode[STIX]{x1D6FC}_{i}$ is the simple coroot corresponding to $i$ ) then the corresponding configuration space of points in $X$ is defined by
and there is a smooth surjection
In [Reference Braverman and GaitsgoryBG08] the authors introduced certain factorizable perverse sheaves $\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D707}}$ on the configuration spaces $X^{\unicode[STIX]{x1D707}}$ . The $!$ -fiber of $\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D707}}$ at a point $\sum _{i}\unicode[STIX]{x1D707}_{i}x_{i}$ is
where $\check{\mathfrak{n}}$ is the unipotent radical of a Borel subalgebra in the Langlands dual Lie algebra $\check{\mathfrak{g}}$ , and $C^{\bullet }(\check{\mathfrak{n}})$ is its ${\check{T}}$ -graded cohomological Chevalley complex.
Theorem 1.3.1. For any $\unicode[STIX]{x1D707}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ there is an isomorphism $\mathfrak{j}_{=\unicode[STIX]{x1D707}}^{!}\unicode[STIX]{x1D6F9}(\mathscr{W})\tilde{\rightarrow }\mathfrak{m}_{\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D6E5}}\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D707}}$ .
In fact, $\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D707}}$ is indecomposable and $\mathfrak{m}_{\unicode[STIX]{x1D707}}$ has contractible fibers, so the isomorphism in the theorem is automatically unique up to scaling. Since $\unicode[STIX]{x1D6F9}(\mathscr{W})$ is Verdier self-dual, Theorem 1.3.1 implies that it is tilting with respect to the defect stratification.
1.4
Recall that the monodromy filtration on $\unicode[STIX]{x1D6F9}(\mathscr{W})$ is the unique filtration by perverse sheaves
such that for all $1\leqslant i\leqslant n$ the $i$ th power of the monodromy endomorphism induces an isomorphism $F_{i}/F_{i-1}\tilde{\rightarrow }F_{-i}/F_{-i-1}$ . The associated graded sheaf $\operatorname{gr}\unicode[STIX]{x1D6F9}(\mathscr{W})$ has an action of the so-called Lefschetz $\mathfrak{sl}_{2}$ such that the lowering operator is induced by the monodromy endomorphism and $F_{i}/F_{i-1}$ has weight $i$ for the Cartan operator.
We now formulate a description of $\operatorname{gr}\unicode[STIX]{x1D6F9}(\mathscr{W})$ in terms of certain factorizable perverse sheaves with $\mathfrak{sl}_{2}$ -action, called the Picard–Lefschetz oscillators after [Reference SchiederSch18] (see also § 3.2 of [Reference SchiederSch17]).
First let us define the Picard–Lefschetz oscillators on $X^{(n)}$ . Let $\operatorname{std}$ denote the standard two-dimensional representation of the Lefschetz $\mathfrak{sl}_{2}$ , and write $\operatorname{sgn}$ for the sign character of the symmetric group $\unicode[STIX]{x1D6F4}_{n}$ . The $\unicode[STIX]{x1D6F4}_{n}\times \mathfrak{sl}_{2}$ -representation $\operatorname{sgn}\otimes \operatorname{std}^{\otimes n}$ (here $\unicode[STIX]{x1D6F4}_{n}$ also permutes the $\operatorname{std}$ factors) gives rise to a local system with $\mathfrak{sl}_{2}$ -action on the disjoint locus $X_{\operatorname{disj}}^{(n)}$ , by applying the associated bundle construction to the canonical $\unicode[STIX]{x1D6F4}_{n}$ -torsor $X_{\operatorname{disj}}^{n}\rightarrow X_{\operatorname{disj}}^{(n)}$ . Then $\mathscr{P}_{n}$ is defined as the intermediate extension to $X^{(n)}$ of the (perverse cohomological shift of the) local system attached to $\operatorname{sgn}\otimes \operatorname{std}^{\otimes n}$ . The perverse sheaf $\mathscr{P}_{n}$ carries an $\mathfrak{sl}_{2}$ -action by functoriality, and is evidently semisimple.
Recall that a Kostant partition of $\unicode[STIX]{x1D707}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ is an expression of the form $\unicode[STIX]{x1D707}=\sum _{\unicode[STIX]{x1D6FD}\in R^{+}}n_{\unicode[STIX]{x1D6FD}}\unicode[STIX]{x1D6FD}$ , where the $n_{\unicode[STIX]{x1D6FD}}$ are nonnegative integers and $R^{+}$ denotes the set of positive coroots. Denote by $\operatorname{Kost}(\unicode[STIX]{x1D707})$ the set of Kostant partitions of $\unicode[STIX]{x1D707}$ . To any $\mathfrak{k}\in \operatorname{Kost}(\unicode[STIX]{x1D707})$ given by $\unicode[STIX]{x1D707}=\sum _{\unicode[STIX]{x1D6FD}\in R^{+}}n_{\unicode[STIX]{x1D6FD}}\unicode[STIX]{x1D6FD}$ we attach a partially symmetrized power
which is equipped with a canonical finite map $\unicode[STIX]{x1D704}^{\mathfrak{k}}:X^{\mathfrak{k}}\rightarrow X^{\unicode[STIX]{x1D707}}$ .
The Picard–Lefschetz oscillator on $X^{\unicode[STIX]{x1D707}}$ is defined by the formula
In particular, we have $\mathscr{P}^{n\unicode[STIX]{x1D6FC}}=\mathscr{P}_{n}$ for $\unicode[STIX]{x1D6FC}$ a simple coroot. By construction, $\mathscr{P}^{\unicode[STIX]{x1D707}}$ is a semisimple perverse sheaf with $\mathfrak{sl}_{2}$ -action.
Theorem 1.4.1. There is an $\mathfrak{sl}_{2}$ -equivariant isomorphism
In particular $\operatorname{gr}\unicode[STIX]{x1D6F9}(\mathscr{W})$ is semisimple. Using the theorem we can compute the kernel of the monodromy operator, whose simple subquotients are the lowest weight sheaves for the Lefschetz $\mathfrak{sl}_{2}$ -action.
Corollary 1.4.1.1. The canonical morphism $\mathfrak{j}_{=0,!}\operatorname{IC}_{\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}}\rightarrow \unicode[STIX]{x1D6F9}(\mathscr{W})$ is an isomorphism onto the kernel of the monodromy operator.
1.5
A key step in the proof of Theorem 1.4.1 is establishing that $\operatorname{gr}\unicode[STIX]{x1D6F9}(\mathscr{W})$ is semisimple. We give a standard weight-theoretic argument, justified by Mochizuki’s theory of weights for holonomic D-modules developed in [Reference MochizukiMoc15]. We also present a more direct proof of Theorem 1.4.1 which uses the remarkable action, constructed in [Reference Feigin, Finkelberg, Kuznetsov and MirkovicFFKM99], of $\check{\mathfrak{g}}$ on the intersection cohomology of quasimap spaces. As we now explain, this action can be obtained via a deformation-theoretic argument from the global unramified categorical geometric Langlands conjecture, together with its expected compatibility with geometric Eisenstein series. This leads us to a natural generalization of the action from [Reference Feigin, Finkelberg, Kuznetsov and MirkovicFFKM99], which the author intends to return to in future work.
Write $\overline{\mathfrak{p}}:\overline{\operatorname{Bun}}_{B}\rightarrow \operatorname{Bun}_{G}$ and $\overline{\mathfrak{q}}:\overline{\operatorname{Bun}}_{B}\rightarrow \operatorname{Bun}_{T}$ for the canonical morphisms. The functor of compactified Eisenstein series $\operatorname{Eis}_{!\ast }:D(\operatorname{Bun}_{T})\rightarrow D(\operatorname{Bun}_{G})$ , introduced in [Reference Braverman and GaitsgoryBG02], is defined by
The rest of this section plays a motivational role only, and is otherwise not necessary to understand the contents of this paper.
For any algebraic group $H$ , we denote by $\operatorname{LS}_{H}$ the DG algebraic stack of de Rham $H$ -local systems on $X$ . For any DG stack $\mathscr{Y}$ we denote by $\operatorname{QCoh}(\mathscr{Y})$ the unbounded derived category of quasicoherent sheaves on $\mathscr{Y}$ .
The form of the geometric Langlands conjecture which we now state is a coarse version of the one formulated in [Reference Arinkin and GaitsgoryAG15a], but at this level of precision it goes back to Beilinson and Drinfeld. It is expected to enjoy various compatibilities, but here we only describe the conjectural interaction with Eisenstein series. Let
be the map induced by the inclusion ${\check{T}}\rightarrow {\check{G}}$ .
Conjecture 1.5.1 (Global unramified geometric Langlands).
There is a fully faithful functor
which makes the following square commute up to natural isomorphism.
Here $\unicode[STIX]{x1D70C}(\unicode[STIX]{x1D714})$ denotes the automorphism of $D(\operatorname{Bun}_{T})$ given by translation by the same-named $T$ -bundle.
Kodaira–Spencer theory tells us that the tangent space to $\operatorname{LS}_{{\check{G}}}$ at a ${\check{G}}$ -local system $E_{{\check{G}}}$ is identified with $H_{\operatorname{dR}}^{\bullet }(X,\check{\mathfrak{g}}_{E_{{\check{G}}}})[1]$ , where $\check{\mathfrak{g}}_{E_{{\check{G}}}}$ denotes the local system attached to $E_{{\check{G}}}$ and the adjoint representation $\check{\mathfrak{g}}$ . Moreover, the (derived) endomorphism algebra of the skyscraper sheaf $\unicode[STIX]{x1D6FF}_{E_{{\check{G}}}}$ in $\operatorname{QCoh}(\operatorname{LS}_{{\check{G}}})$ is the enveloping algebra of the DG Lie algebra $H_{\operatorname{dR}}^{\bullet }(X,\check{\mathfrak{g}}_{E_{{\check{G}}}})$ , and in particular $H_{\operatorname{dR}}^{\bullet }(X,\check{\mathfrak{g}}_{E_{{\check{G}}}})$ acts on $\unicode[STIX]{x1D6FF}_{E_{{\check{G}}}}$ .
Applying these principles in the case $E_{{\check{G}}}=\mathfrak{r}^{\operatorname{spec}}(E_{{\check{T}}})$ for a ${\check{T}}$ -local system $E_{{\check{T}}}$ , we derive the following concrete consequence of Conjecture 1.5.1. According to geometric class field theory $\mathbb{L}_{T}(\unicode[STIX]{x1D6FF}_{E_{{\check{T}}}})$ is a multiplicative flat line bundle, and in particular is invariant under translation by any $T$ -bundle.
Conjecture 1.5.2. The DG Lie algebra $H_{\operatorname{dR}}^{\bullet }(X,\check{\mathfrak{g}}_{E_{{\check{T}}}})$ acts on $\operatorname{Eis}_{!\ast }(\mathbb{L}_{T}(\unicode[STIX]{x1D6FF}_{E_{{\check{T}}}}))$ .
Moreover, we expect that the full endomorphism algebra of $\operatorname{Eis}_{!\ast }(\mathbb{L}_{T}(\unicode[STIX]{x1D6FF}_{E_{{\check{T}}}}))$ is the enveloping algebra of $H_{\operatorname{dR}}^{\bullet }(X,\check{\mathfrak{g}}_{E_{{\check{T}}}})$ . We also remark that $\mathbb{L}_{T}(\unicode[STIX]{x1D6FF}_{E_{{\check{T}}}})$ has a simple and nonconjectural description: it is the multiplicative line bundle with flat connection on $\operatorname{Bun}_{T}$ corresponding to $E_{{\check{T}}}$ under geometric class field theory. This means that it is characterized (among multiplicative flat line bundles) by the property that for any coweight $\unicode[STIX]{x1D706}:{\check{T}}\rightarrow \mathbb{G}_{m}$ , its inverse image along the map $X\rightarrow \operatorname{Bun}_{T}$ sending $x\mapsto \mathscr{P}_{T}^{\operatorname{triv}}(\unicode[STIX]{x1D706}\cdot x)$ is $\unicode[STIX]{x1D706}(E_{{\check{T}}})$ .
When $E_{{\check{T}}}$ is trivial, which is the only case we will use in this paper, Conjecture 1.5.2 says that $\check{\mathfrak{g}}\,\otimes \,H_{\operatorname{dR}}^{\bullet }(X)$ acts on $\overline{\mathfrak{p}}_{\ast }\operatorname{IC}_{\overline{\operatorname{Bun}}_{B}}$ . In § 6.3 we specify the action of certain generators of $\check{\mathfrak{g}}\,\otimes \,H_{\operatorname{dR}}^{\bullet }(X)$ , namely the factors in the triangular decomposition. In [Reference Feigin, Finkelberg, Kuznetsov and MirkovicFFKM99] the authors verify the necessary relations for $\check{\mathfrak{g}}\,\otimes \,H^{0}(X)$ , which suffices for our application. This approach seems intractable when approaching Conjecture 1.5.2 in full generality: the derivedness of the Lie algebra $\check{\mathfrak{g}}\,\otimes \,H_{\operatorname{dR}}^{\bullet }(X)$ , or more generally $H_{\operatorname{dR}}^{\bullet }(X,\check{\mathfrak{g}}_{E_{{\check{T}}}})$ , makes the checking of relations difficult or impossible to do ‘by hand.’
2 Drinfeld compactifications and Zastava spaces
2.1
Let $2\unicode[STIX]{x1D70C}:\mathbb{G}_{m}\rightarrow T$ denote the sum of the simple coroots and fix a square root $\unicode[STIX]{x1D714}_{X}^{\otimes 1/2}$ of $\unicode[STIX]{x1D714}_{X}$ . We define
By definition, $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ is the fiber product
(see [Reference Braverman and GaitsgoryBG02, § 0.2.1] for the definition of Drinfeld’s compactification $\overline{\operatorname{Bun}}_{B}$ ). The $T$ -bundle $\unicode[STIX]{x1D70C}(\unicode[STIX]{x1D714}_{X})$ and the action of $T$ on $N$ give rise to a group scheme $N^{\unicode[STIX]{x1D714}}$ over $X$ , and the open stratum $\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}\rightarrow \overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ is identified with the moduli stack of $N^{\unicode[STIX]{x1D714}}$ -bundles, as the notation suggests. By construction, $T$ acts on $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ in such a way that $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}\rightarrow \overline{\operatorname{Bun}}_{B}$ factors through a closed embedding
For each $\unicode[STIX]{x1D707}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ the corresponding stratum $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}},=\unicode[STIX]{x1D707}}$ fits into a fiber square
where the lower horizontal morphism is the twisted Abel–Jacobi map $D\mapsto \unicode[STIX]{x1D70C}(\unicode[STIX]{x1D714}_{X})(D)$ . We write
and $\mathfrak{j}_{{\leqslant}\unicode[STIX]{x1D707}}$ for the corresponding open embedding.
The embedding $\mathfrak{j}_{=\unicode[STIX]{x1D707}}$ of the stratum, which is known to be affine, extends to a finite map
where $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}},\geqslant \unicode[STIX]{x1D707}}:=X^{\unicode[STIX]{x1D707}}\times _{\operatorname{Bun}_{T}}\overline{\operatorname{Bun}}_{B}$ .
2.2
Now we introduce the Zastava spaces, which are factorizable local models for $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ . There are several versions of Zastava space, and notations vary significantly within the literature (ours is similar to [Reference Arinkin and GaitsgoryAG15b]).
Define the Zastava space $Z$ to be the open locus in $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}\times _{\operatorname{Bun}_{G}}\operatorname{Bun}_{B^{-}}$ where the generalized $N$ -reduction and $B^{-}$ -reduction are transverse generically on $X$ . It is well known that $Z$ is a scheme, with connected components
indexed by $\unicode[STIX]{x1D706}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ .
The map $\mathfrak{p}^{-}:\operatorname{Bun}_{B^{-}}\rightarrow \operatorname{Bun}_{G}$ gives rise to $\text{}^{\prime }\mathfrak{p}^{-}:Z\rightarrow \overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ . It is shown in [Reference Braverman and GaitsgoryBG02] that $\mathfrak{p}^{-,\unicode[STIX]{x1D706}}$ and therefore $\text{}^{\prime }\mathfrak{p}^{-,\unicode[STIX]{x1D706}}$ are smooth for $\unicode[STIX]{x1D706}$ sufficiently dominant. Moreover, given a quasicompact open $U\subset \overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ , for $\unicode[STIX]{x1D706}$ sufficiently dominant the image of $Z^{\unicode[STIX]{x1D706}}\rightarrow \overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ contains $U$ .
For each $0\leqslant \unicode[STIX]{x1D707}\leqslant \unicode[STIX]{x1D706}$ we have the corresponding stratum $Z_{=\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}:=Z^{\unicode[STIX]{x1D706}}\times _{\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}}\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}},=\unicode[STIX]{x1D707}}$ with locally closed embedding
Define $\mathring{Z}^{\unicode[STIX]{x1D706}}:=Z_{=0}^{\unicode[STIX]{x1D706}}$ . Similarly, we have $Z_{{\leqslant}\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}:=Z^{\unicode[STIX]{x1D706}}\times _{\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}}\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}},\leqslant \unicode[STIX]{x1D707}}$ with the open embedding
Put $Z_{{\geqslant}\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}:=Z\times _{\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}}\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}},\geqslant \unicode[STIX]{x1D707}}$ , so that $\text{}^{\prime }\mathfrak{j}_{=\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}$ extends to the finite map
2.3
Let us recall the factorization structure on $Z$ . For every $\unicode[STIX]{x1D706}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ there is a canonical map $\unicode[STIX]{x1D70B}^{\unicode[STIX]{x1D706}}:Z^{\unicode[STIX]{x1D706}}\rightarrow X^{\unicode[STIX]{x1D706}}$ , which is well known to be affine. We write $(X^{\unicode[STIX]{x1D706}_{1}}\times X^{\unicode[STIX]{x1D706}_{2}})_{\operatorname{disj}}\subset X^{\unicode[STIX]{x1D706}_{1}}\times X^{\unicode[STIX]{x1D706}_{2}}$ for the open locus where the two divisors are disjoint. Similarly, put
The factorization structure is a canonical morphism $(Z^{\unicode[STIX]{x1D706}_{1}}\times Z^{\unicode[STIX]{x1D706}_{2}})_{\operatorname{disj}}\rightarrow Z^{\unicode[STIX]{x1D706}_{1}+\unicode[STIX]{x1D706}_{2}}$ which fits into a fiber square, as follows.
The factorization structure is compatible with the defect stratification in the following sense. The factorization structure on the strata consists of, for each decomposition $\unicode[STIX]{x1D707}_{1}+\unicode[STIX]{x1D707}_{2}=\unicode[STIX]{x1D707}$ satisfying $0\leqslant \unicode[STIX]{x1D707}_{1}\leqslant \unicode[STIX]{x1D706}_{1}$ and $0\leqslant \unicode[STIX]{x1D707}_{2}\leqslant \unicode[STIX]{x1D706}_{2}$ , a morphism $(Z_{=\unicode[STIX]{x1D707}_{1}}^{\unicode[STIX]{x1D706}_{1}}\times Z_{=\unicode[STIX]{x1D707}_{2}}^{\unicode[STIX]{x1D706}_{2}})_{\operatorname{ disj}}\rightarrow Z_{=\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}_{1}+\unicode[STIX]{x1D706}_{2}}$ which fits into a fiber square, as follows.
One has similar factorization structures on $Z_{{\leqslant}\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}$ and $Z_{{\geqslant}\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}$ . Moreover, these factorization structures are compatible with $\text{}^{\prime }\mathfrak{j}_{=\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}$ , $\text{}^{\prime }\mathfrak{j}_{{\geqslant}\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}$ , etc.
2.4
We will also need the compactified Zastava space $\overline{Z}$ , which is the open locus in $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}\times _{\operatorname{Bun}_{G}}\overline{\operatorname{Bun}}_{B^{-}}$ where the generalized $N$ - and $B^{-}$ -reductions are generically transverse. In particular there is an open embedding $\text{}^{\prime }\mathfrak{j}^{-}:Z\rightarrow \overline{Z}$ obtained from $\mathfrak{j}^{-}:\operatorname{Bun}_{B^{-}}\rightarrow \overline{\operatorname{Bun}}_{B^{-}}$ by base change.
For any $\unicode[STIX]{x1D708}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ we put
and similarly for $_{{\leqslant}\unicode[STIX]{x1D708}}\overline{Z}$ and $_{{\geqslant}\unicode[STIX]{x1D708}}\overline{Z}$ .
The projections $\unicode[STIX]{x1D70B}^{\unicode[STIX]{x1D706}}$ extend to proper morphisms $\overline{\unicode[STIX]{x1D70B}}^{\unicode[STIX]{x1D706}}:\overline{Z}^{\unicode[STIX]{x1D706}}\rightarrow X^{\unicode[STIX]{x1D706}}$ . The factorization structure on $Z$ extends to $\overline{Z}$ in a way compatible with both defect stratifications on $\overline{Z}$ .
3 Nearby cycles and adelic invariance
3.1
We now reinterpret Beilinson’s construction of nearby cycles from [Reference BeilinsonBei87], which will help to streamline our first proof of Theorem 1.3.1. In this section we use the language of DG categories in the sense of Lurie, meaning $k$ -linear stable $\infty$ -categories: see [Reference Gaitsgory and RozenblymGR17, ch. I.1] for a summary of the theory.
Let $Y$ be a scheme of finite type equipped with an action of $\mathbb{G}_{m}$ and a $\mathbb{G}_{m}$ -equivariant morphism $f:Y\rightarrow \mathbb{A}^{1}$ . Write $Y_{0}$ for the fiber of $Y$ over $0$ and $\mathring{Y}$ for the preimage of $\mathbb{G}_{m}\subset \mathbb{A}^{1}$ . We temporarily denote the embeddings by
The projection $\mathring{Y}\rightarrow \mathbb{G}_{m}$ gives rise to an action of $D(\mathbb{G}_{m})$ , viewed as a symmetric monoidal DG category under $!$ -tensor product, on the DG category $D(\mathring{Y})$ . We claim that this naturally induces an action of the DG algebra $H^{\bullet }(\mathbb{G}_{m})$ on the identity functor of $D(\mathring{Y})$ . Namely, we view the action of $D(\mathbb{G}_{m})$ on $D(\mathring{Y})$ as a monoidal functor
Since $H^{\bullet }(\mathbb{G}_{m})$ is the algebra of endomorphisms of the monoidal unit in $D(\mathbb{G}_{m})$ , this monoidal functor induces the desired action.
In particular, for any holonomic D-module $\mathscr{F}$ on $\mathring{Y}$ , we have an action of $H^{\bullet }(\mathbb{G}_{m})$ on $i^{!}j_{!}\mathscr{F}$ by functoriality. The point $1\in \mathbb{G}_{m}$ induces an augmentation $H^{\bullet }(\mathbb{G}_{m})\rightarrow k$ , and since $H^{\bullet }(\mathbb{G}_{m})$ is generated by a single element in cohomological degree $1$ , the endomorphism algebra $\operatorname{End}_{H^{\bullet }(\mathbb{G}_{m})\text{-mod}}(k)$ is canonically isomorphic to the polynomial ring $k[t]$ . Here $H^{\bullet }(\mathbb{G}_{m})\text{-mod}$ is the DG category of modules over the DG algebra $H^{\bullet }(\mathbb{G}_{m})$ .
Definition 3.1.1. We define unipotent nearby cycles with respect to $f$ to be the functor
given on objects by the formula
In particular $\operatorname{End}_{H^{\bullet }(\mathbb{G}_{m})\text{-mod}}(k)=k[t]$ acts on $\unicode[STIX]{x1D6F9}$ , and we call the action of the generator the monodromy endomorphism of $\unicode[STIX]{x1D6F9}$ .
More precisely, the action of $H^{\bullet }(\mathbb{G}_{m})$ on the identity functor of $D(\mathring{Y})$ defines a lift of the latter to a functor
and likewise on the holonomic subcategory. Then $\unicode[STIX]{x1D6F9}$ is by definition the composition
where the last functor is tensor product with the augmentation $H^{\bullet }(\mathbb{G}_{m})$ -module. Since $H^{\bullet }(\mathbb{G}_{m})$ is isomorphic to $\operatorname{Sym}(k[-1])$ , an action of $H^{\bullet }(\mathbb{G}_{m})$ is the same as an action of the abelian DG Lie algebra $k[-1]$ . Tensor product with the augmentation $H^{\bullet }(\mathbb{G}_{m})$ -module corresponds to (homotopy) $k[-1]$ -coinvariants.
3.2
The following lemma will be used to compare our construction of $\unicode[STIX]{x1D6F9}$ with Beilinson’s. It says that the homotopy coinvariants of $H^{1}(\mathbb{G}_{m})[-1]$ acting on the constant sheaf $k_{\mathbb{G}_{m}}$ is the ‘infinite Jordan block.’
For any $a\geqslant 1$ let $L_{a}$ be the shifted D-module on $\mathbb{G}_{m}$ corresponding to the local system whose monodromy is a unipotent Jordan block of rank $a$ . There are canonical injections $L_{a}\rightarrow L_{a+1}$ (in the heart of the constructible t-structure), and we put
Observe that $L_{\infty }$ has a canonical injective ‘shift’ endomorphism with cokernel $k_{\mathbb{G}_{m}}$ .
Lemma 3.2.1. There is a canonical isomorphism
in $D(\mathbb{G}_{m})$ , which identifies the action of the generator in $k[t]=\operatorname{End}_{H^{\bullet }(\mathbb{G}_{m})\text{-mod}}(k)$ with the canonical endomorphism of $L_{\infty }$ .
Proof. First, observe that since any self-extension of $L_{\infty }$ splits, the action of $H^{1}(\mathbb{G}_{m})[-1]$ on $L_{\infty }$ is trivial. Thus the canonical inclusion $k_{\mathbb{G}_{m}}=L_{1}\rightarrow L_{\infty }$ factors through a map
To prove that this map is an isomorphism, we use a Koszul-type resolution of the augmentation $H^{\bullet }(\mathbb{G}_{m})$ -module. Namely, put $M_{1}:=H^{\bullet }(\mathbb{G}_{m})$ , and let
be the cofiber (i.e., mapping cone) of the action of a generator of $H^{1}(\mathbb{G}_{m})$ . Proceeding inductively, we define
so there are canonical maps $M_{a}\rightarrow M_{a+1}$ . We define $M_{\infty }$ to be the colimit of the $M_{a}$ . Since the module $M_{a}$ has cohomology $k\oplus k[-a]$ , the canonical map $M_{\infty }\rightarrow k$ to the augmentation $H^{\bullet }(\mathbb{G}_{m})$ -module is an isomorphism. In particular $k[t]$ acts on $M_{\infty }$ , and it is not hard to see that the generator acts by the canonical ‘shift’ map $M_{\infty }\rightarrow M_{\infty }$ with cofiber $M_{1}$ .
Thus we may replace the left-hand side of (3.2.1) with $M_{\infty }\otimes _{H^{\bullet }(\mathbb{G}_{m})}k_{\mathbb{G}_{m}}$ . A straightforward inductive argument shows that $M_{a}\,\otimes _{H^{\bullet }(\mathbb{G}_{m})}\,k_{\mathbb{G}_{m}}$ maps isomorphically onto $L_{a}$ , which proves that (3.2.1) is an isomorphism which moreover preserves the filtrations on both sides. The action of $k[t]$ on both sides agrees by inspection: both endomorphisms shift the filtration by $1$ .◻
Proposition 3.2.2. The functor $\unicode[STIX]{x1D6F9}$ has the following properties.
(i) It coincides with the construction in [Reference BeilinsonBei87] and in particular preserves holonomicity.
(ii) If $\mathscr{F}$ is $\mathbb{G}_{m}$ -equivariant, then $\unicode[STIX]{x1D6F9}(\mathscr{F})$ is unipotently $\mathbb{G}_{m}$ -monodromic, and the monodromy endomorphism agrees with the obstruction to $\mathbb{G}_{m}$ -equivariance.
Proof. Let $\mathscr{F}$ be a holonomic D-module on $\mathring{Y}$ . Then Beilinson’s nearby cycles is given by the formula
Moreover, we have $\tilde{\unicode[STIX]{x1D6F9}}(\mathscr{F})=H^{0}i^{!}j_{!}(\mathscr{F}\stackrel{\ast }{\otimes }f^{\ast }L_{a})$ for large $a$ , which is evidently holonomic. The functor
given on objects by $\mathscr{M}\mapsto i^{!}j_{!}(\mathscr{F}\stackrel{\ast }{\otimes }f^{\ast }\mathscr{M})$ admits a unique colimit-preserving extension to the category of ind-holonomic sheaves on $\mathbb{G}_{m}$ , so one can write
The functor $\tilde{\unicode[STIX]{x1D6F9}}$ carries a canonical endomorphism induced by the shift endomorphism $L_{\infty }\rightarrow L_{\infty }$ . Now it follows from Lemma 3.2.1 that $\unicode[STIX]{x1D6F9}(\mathscr{F})\tilde{\rightarrow }\tilde{\unicode[STIX]{x1D6F9}}(\mathscr{F})$ , preserving the monodromy endomorphisms on both sides.
For (ii), note that the $\mathbb{G}_{m}$ -equivariance of $i^{!}j_{!}\mathscr{F}$ implies that $\unicode[STIX]{x1D6F9}(\mathscr{F})=H^{0}i^{!}j_{!}(\mathscr{F}\stackrel{\ast }{\otimes }f^{\ast }L_{a})$ is $\mathbb{G}_{m}$ -monodromic (here $a$ is large). By construction, the monodromy endomorphism is induced by the canonical endomorphism of $L_{a}$ with one-dimensional kernel and cokernel. But the latter is precisely the obstruction to $\mathbb{G}_{m}$ -equivariance for $L_{a}$ , so the claim follows from the functoriality of this obstruction.◻
It follows from part (i) of Proposition 3.2.2 that $\unicode[STIX]{x1D6F9}$ enjoys the standard properties of the unipotent nearby cycles functor: it is t-exact, commutes with Verdier duality, and commutes with proper direct image and smooth inverse image.
3.3
Before proving Theorem 1.3.1, we will show that $\mathfrak{j}_{=\unicode[STIX]{x1D707}}^{!}\unicode[STIX]{x1D6F9}(\mathscr{W})$ is pulled back from $X^{\unicode[STIX]{x1D707}}$ for any $\unicode[STIX]{x1D707}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ . This property is equivalent to invariance under the ‘adelic $N^{\unicode[STIX]{x1D714}}$ ,’ as we now explain.
For any $x\in X$ , we define the open substack $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}^{x}\subset \overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ to consist of those generalized $N^{\unicode[STIX]{x1D714}}$ -bundles whose defect is disjoint from $x$ . A point of the ind-algebraic stack $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}$ consists of two points of $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}^{x}$ together with an identification over $X\setminus \{x\}$ . Note that $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}$ has the structure of a groupoid acting on $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}^{x}$ . The fibers of $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}$ over $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}^{x}\times \overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}^{x}$ are isomorphic to ind-affine space $\operatorname{colim}_{n}\mathbb{A}^{n}$ , which implies that the functor which forgets $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}$ -equivariance is fully faithful, i.e., $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}$ -equivariance is a property.
We say that an object of $D(\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}})$ is $N^{\unicode[STIX]{x1D714}}(\mathbb{A})$ -equivariant if, for every $x\in X$ , its restriction to $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}^{x}$ is $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}$ -equivariant.
Proposition 3.3.1. An object $\mathscr{F}$ of $D(\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}})$ is $N^{\unicode[STIX]{x1D714}}(\mathbb{A})$ -equivariant if and only if, for every $\unicode[STIX]{x1D707}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ , the canonical morphism
is an isomorphism.
For each $x\in X$ , denote by $\mathscr{O}_{x}$ the completed local ring of $X$ at $x$ , with fraction field $K_{x}$ . If $R$ is a $k$ -algebra, we denote by $R~\hat{\otimes }~\mathscr{O}_{x}$ and $R~\hat{\otimes }~K_{x}$ the respective completed tensor products.
The local Hecke stack $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{\operatorname{loc},x}$ is defined as follows: a $\operatorname{Spec}R$ -point of $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{\operatorname{loc},x}$ consists of two $N^{\unicode[STIX]{x1D714}}$ -bundles over $\operatorname{Spec}(R~\hat{\otimes }~\mathscr{O}_{x})$ equipped with an isomorphism over $\operatorname{Spec}(R~\hat{\otimes }~K_{x})$ . There is a natural restriction map $\operatorname{res}^{x}:\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}\rightarrow \mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{\operatorname{loc},x}$ .
Our choice of isomorphism $N/[N,N]\cong \mathbb{G}_{a}^{\oplus I}$ induces a map $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{\operatorname{loc},x}\rightarrow \mathbb{G}_{a}$ in the following way. The projection $N\rightarrow [N,N]$ induces a morphism
where $\mathbb{G}_{a}^{\unicode[STIX]{x1D714}}$ is the additive group scheme over $X$ attached to the canonical line bundle $\unicode[STIX]{x1D714}_{X}$ . Note that there is a canonical isomorphism of groupoids
so taking residues defines a morphism $\operatorname{rsd}^{x}:\mathscr{H}_{\mathbb{G}_{a}^{\unicode[STIX]{x1D714}}}^{\operatorname{loc},x}\rightarrow \mathbb{G}_{a}$ . The composition
is an additive character, meaning it is a morphism of groupoids.
It follows that $\widetilde{\unicode[STIX]{x1D712}}^{x}:=\operatorname{rsd}_{\unicode[STIX]{x1D713}}^{x,!}\unicode[STIX]{x1D712}$ is a character sheaf on $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}$ , i.e., it is multiplicative for the groupoid structure. In particular we can speak of $(\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x},\widetilde{\unicode[STIX]{x1D712}}^{x})$ -equivariant sheaves on $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}^{x}$ , which form a full subcategory of $D(\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}^{x})$ . Likewise, if a sheaf on $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ is $(\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x},\widetilde{\unicode[STIX]{x1D712}}^{x})$ -equivariant for all $x\in X$ we say that it is $(N^{\unicode[STIX]{x1D714}}(\mathbb{A}),\widetilde{\unicode[STIX]{x1D712}})$ -equivariant. Although we will not use this fact, the category of $(N^{\unicode[STIX]{x1D714}}(\mathbb{A}),\widetilde{\unicode[STIX]{x1D712}})$ -equivariant sheaves on $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ is equivalent to the derived category of vector spaces, being generated by the object $\mathscr{W}_{1}$ introduced above. Moreover $(\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x},\widetilde{\unicode[STIX]{x1D712}}^{x})$ -equivariance for a single $x\in X$ implies $(N^{\unicode[STIX]{x1D714}}(\mathbb{A}),\widetilde{\unicode[STIX]{x1D712}})$ -equivariance.
Observe that there is a natural $T$ -action on $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{\operatorname{loc},x}$ . Using the chosen dominant regular cocharacter $\unicode[STIX]{x1D6FE}:\mathbb{G}_{m}\rightarrow T$ , the resulting $\mathbb{G}_{m}$ -action on $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{\operatorname{loc},x}$ is contracting. In particular, it extends to an action $\mathbb{A}^{1}\times \mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{\operatorname{loc},x}\rightarrow \mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{\operatorname{loc},x}$ of the multiplicative monoid $\mathbb{A}^{1}$ . The $!$ -pullback of $\unicode[STIX]{x1D712}$ along the composition
defines an $\mathbb{A}^{1}$ -family $\widetilde{\unicode[STIX]{x1D712}}_{\operatorname{ext}}^{x}$ of character sheaves on $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}$ . Its $!$ -restriction to $\{1\}\times \mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}$ is $\widetilde{\unicode[STIX]{x1D712}}_{x}$ , and it is trivial along $\{0\}\times \mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}$ .
Lemma 3.3.2. The sheaf $\unicode[STIX]{x1D6F9}(\mathscr{W})$ is $N^{\unicode[STIX]{x1D714}}(\mathbb{A})$ -equivariant.
Proof. Fix $x\in X$ ; we omit restriction to $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}^{x}$ from the notation in what follows. By construction $\mathscr{W}$ is $\widetilde{\unicode[STIX]{x1D712}}_{\operatorname{ext}}^{x}|_{\mathbb{G}_{m}\times \mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}}$ -equivariant. Since $\widetilde{\unicode[STIX]{x1D712}}_{\operatorname{ext}}^{x}|_{\{0\}\times \mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}}^{!}$ is the trivial character sheaf, it follows from our construction of $\unicode[STIX]{x1D6F9}$ (or equivalently, Beilinson’s) that $\unicode[STIX]{x1D6F9}(\mathscr{W})$ is $\mathscr{H}_{N^{\unicode[STIX]{x1D714}}}^{x}$ -equivariant as desired.◻
4 Restriction to the strata
4.1
Now we give the first proof of Theorem 1.3.1 by deducing it from [Reference Arinkin and GaitsgoryAG15b, Theorem 1.3.6], which describes the restrictions to the strata of $i!j_{!}\mathscr{W}$ in terms of the perverse sheaf $\unicode[STIX]{x1D6FA}$ . Since we work with a fixed dominant regular coweight $\unicode[STIX]{x1D6FE}$ rather than the entire torus $T$ , it will be necessary to prove a slightly different formulation of the latter theorem.
As in [Reference Arinkin and GaitsgoryAG15b, § 7.2], the inclusion
of $N^{\unicode[STIX]{x1D714}}(\mathbb{A})$ -equivariant sheaves on $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ admits a right adjoint, which we denote by
Let $i$ and $j$ be as in § 1.2, and write $p:\mathbb{A}^{1}\times \overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}\rightarrow \overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ for the projection. We will also abusively denote $p\circ j$ by $p$ .
Proposition 4.1.1. There is a canonical isomorphism $i^{!}j_{!}\mathscr{W}\tilde{\rightarrow }\operatorname{Av}_{\ast }^{N^{\unicode[STIX]{x1D714}}(\mathbb{A})}p_{!}\mathscr{W}$ .
Proof. See [Reference Arinkin and GaitsgoryAG15b, § 10.3], where the claim is proved for the action of the entire torus $T$ . The same proof applies mutatis mutandi to our claim, which involves only the $\mathbb{G}_{m}$ -action.◻
First proof of Theorem 1.3.1.
Since $\mathfrak{j}_{=\unicode[STIX]{x1D707}}^{!}$ preserves (homotopy) colimits, we see that
By Proposition 4.1.1 we have
and Proposition 3.3.1 implies that
Now Theorem 1.3.6 of [Reference Arinkin and GaitsgoryAG15b] yields
It remains to show that under the composed isomorphism
the action of $H^{\bullet }(\mathbb{G}_{m})$ on the left-hand side corresponds to the natural action on $H_{c}^{\bullet }(\mathbb{G}_{m})$ on the right-hand side. Since $H_{c}^{\bullet }(\mathbb{G}_{m})=H^{\bullet }(\mathbb{G}_{m})[-1]$ as $H^{\bullet }(\mathbb{G}_{m})$ -modules, this will finish the proof.
It is clear that
intertwines the actions of $H^{\bullet }(\mathbb{G}_{m})$ , since it is obtained by evaluating a morphism of functors on $\mathscr{W}$ . Tracing through the proof of Theorem 1.3.6 in [Reference Arinkin and GaitsgoryAG15b], we see that the isomorphism (4.1.1) is also obtained by evaluating a morphism of functors on $\mathscr{W}$ , with the appearance of $H_{c}^{\bullet }(\mathbb{G}_{m})$ accounted for by the isomorphism
The latter isomorphism intertwines the actions of $H^{\bullet }(\mathbb{G}_{m})$ as needed.◻
4.2
The rest of this subsection is devoted to the second proof of Theorem 1.3.1. This proof applies [Reference RaskinRas16, Theorem 4.6.1], which says that $\unicode[STIX]{x1D6FA}$ can be realized as the twisted cohomology of Zastava space. Accordingly, we must formulate the analogue of Theorem 1.3.1 on Zastava space. First, the Whittaker sheaf: we claim that
is perverse for any $\unicode[STIX]{x1D706}$ . For $\unicode[STIX]{x1D706}$ sufficiently dominant $\mathfrak{p}^{-,\unicode[STIX]{x1D706}}$ is smooth, so that $\mathscr{W}_{Z^{\unicode[STIX]{x1D706}}}$ is the cohomologically normalized inverse image of the perverse sheaf $\mathscr{W}$ . If $\unicode[STIX]{x1D706}^{\prime }\leqslant \unicode[STIX]{x1D706}$ then we can pull back $\mathscr{W}_{Z^{\unicode[STIX]{x1D706}}}$ along $\operatorname{id}_{\mathbb{G}_{m}}$ times the factorization morphism
and it is not hard to see that we obtain the restriction of $\mathscr{W}_{Z^{\unicode[STIX]{x1D706}^{\prime }}}\boxtimes \,\mathscr{W}_{Z^{\unicode[STIX]{x1D706}-\unicode[STIX]{x1D706}^{\prime }}}$ . Since the factorization map is étale, this implies that $\mathscr{W}_{Z^{\unicode[STIX]{x1D706}^{\prime }}}$ is perverse as desired.
Since the map $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}\rightarrow \operatorname{Bun}_{G}$ is $T$ -equivariant for the trivial action of $T$ on $\operatorname{Bun}_{G}$ , we obtain an action of $T$ on $Z$ which makes $\text{}^{\prime }\mathfrak{p}^{-}$ a $T$ -equivariant map. In particular our fixed dominant regular cocharacter $\unicode[STIX]{x1D6FE}:\mathbb{G}_{m}\rightarrow T$ induces an action of $\mathbb{G}_{m}$ on $Z$ , and since $\mathscr{W}$ was $\mathbb{G}_{m}$ -equivariant, so is $\mathscr{W}_{Z}$ .
Theorem 4.2.1. For any $\unicode[STIX]{x1D707}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ there is an isomorphism
We will need to use the factorization structure on $Z$ in the following way. First, observe that $\mathscr{W}_{Z}$ admits a natural factorization structure. Thus $\unicode[STIX]{x1D6F9}(\mathscr{W}_{Z})$ admits a factorization structure by the Künneth formula for nearby cycles. Although the Künneth formula holds for the total nearby cycles functor, in this case the total nearby cycles equals the unipotent nearby cycles because $\mathscr{W}_{Z}$ is $\mathbb{G}_{m}$ -equivariant.
4.3
In the second proof, Theorems 1.3.1 and 4.2.1 will be proved simultaneously by an inductive argument. The argument uses the following key lemma.
Let $f:\mathscr{X}\rightarrow \mathscr{Y}$ be a morphism of Artin stacks with $\mathscr{Y}$ smooth, and suppose we are given a function $\mathscr{Y}\rightarrow \mathbb{A}^{1}$ . Let $g:S\rightarrow \mathscr{Y}$ be a morphism where $S$ is an affine scheme and consider the following cartesian square.
Write $i:S_{0}\rightarrow S$ for the inclusion of the vanishing locus of the function $S\rightarrow \mathscr{Y}\rightarrow \mathbb{A}^{1}$ .
Lemma 4.3.1. For any $\mathscr{F}\in D(\mathscr{X})$ which is universally locally acyclic (ULA) over $\mathscr{Y}$ and any $\mathscr{G}\in D(S)$ , there is a canonical isomorphism
Fix $\unicode[STIX]{x1D708}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ . We will apply the lemma in the case $\mathscr{X}=\mathbb{A}^{1}\times \overline{\operatorname{Bun}}_{B^{-},\leqslant \unicode[STIX]{x1D708}}^{\unicode[STIX]{x1D706}}$ , $\mathscr{F}=\operatorname{IC}_{\mathbb{A}^{1}}\boxtimes \,\mathfrak{j}_{!}^{-}(\operatorname{IC}_{\operatorname{Bun}_{B^{-}}^{\unicode[STIX]{x1D706}}})|_{\overline{\operatorname{Bun}}_{B^{-},\leqslant \unicode[STIX]{x1D708}}^{\unicode[STIX]{x1D706}}}$ , and $\mathscr{Y}=\mathbb{A}^{1}\times \operatorname{Bun}_{G}$ . Let us check that the ULA property holds when $\unicode[STIX]{x1D706}$ is sufficiently dominant relative to $\unicode[STIX]{x1D708}$ .
According to [Reference Braverman and GaitsgoryBG08, Corollary 4.5], the perverse sheaf $\mathfrak{j}_{!}^{-}(\operatorname{IC}_{\operatorname{Bun}_{B^{-}}})|_{\overline{\operatorname{Bun}}_{B^{-},\leqslant \unicode[STIX]{x1D708}}^{\unicode[STIX]{x1D706}}}$ has a filtration by $\unicode[STIX]{x1D702}\leqslant \unicode[STIX]{x1D708}$ , with its subquotients being the perverse sheaves $\mathfrak{j}_{{\geqslant}\unicode[STIX]{x1D702},!}^{-}(\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D702}}\boxtimes \,\text{IC}_{\overline{\operatorname{Bun}}_{B^{-},\leqslant \unicode[STIX]{x1D708}-\unicode[STIX]{x1D702}}^{\unicode[STIX]{x1D706}-\unicode[STIX]{x1D702}}})$ . Here $\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D702}}$ is attached to $\check{\mathfrak{n}}^{-}$ rather than $\check{\mathfrak{n}}$ . Since the ULA property is stable under extensions and exterior products, it suffices to show that these subquotients are ULA over $\operatorname{Bun}_{G}$ .
Observe that the diagram
commutes, where the left vertical arrow is projection onto the second factor. Since $\mathfrak{j}_{{\geqslant}\unicode[STIX]{x1D702}}^{-}$ is proper, the functor $\mathfrak{j}_{{\geqslant}\unicode[STIX]{x1D702},!}^{-}$ preserves the ULA property. Thus it suffices to prove that, for $\unicode[STIX]{x1D706}$ sufficiently dominant, $\operatorname{IC}_{\overline{\operatorname{Bun}}_{B^{-},\leqslant \unicode[STIX]{x1D708}-\unicode[STIX]{x1D702}}^{\unicode[STIX]{x1D706}-\unicode[STIX]{x1D702}}}$ is ULA over $\operatorname{Bun}_{G}$ for all $\unicode[STIX]{x1D702}\leqslant \unicode[STIX]{x1D708}$ . This follows immediately from [Reference CampbellCam16, Corollary 4.1.1.1].
Second proof of Theorems 1.3.1 and 4.2.1.
Observe that Theorem 1.3.1 is trivial on the open stratum, since
and similarly for Theorem 4.2.1 on $\mathring{Z}$ .
We begin by proving Theorem 4.2.1 for the deepest strata, i.e., the closed embeddings
Recall that $\unicode[STIX]{x1D6F9}(\mathscr{W}_{Z})$ is $\mathbb{G}_{m}$ -monodromic by construction, so the contraction principle says that
Write ${\unicode[STIX]{x1D70B}\unicode[STIX]{x0030A}}^{\unicode[STIX]{x1D707}}:=\unicode[STIX]{x1D70B}^{\unicode[STIX]{x1D707}}\circ \text{}^{\prime }\mathfrak{j}_{=0}^{\unicode[STIX]{x1D707}}$ . Theorem 4.6.1 in [Reference RaskinRas16] implies that there is an isomorphism
compatible with the factorization structures. Since $\mathscr{W}_{Z}$ is $!$ -extended from $\mathbb{G}_{m}\times \mathring{Z}$ , we obtain
Since $\unicode[STIX]{x1D70B}=\overline{\unicode[STIX]{x1D70B}}\circ \text{}^{\prime }\mathfrak{j}^{-}$ and $\unicode[STIX]{x1D6F9}$ commutes with proper pushforwards, we have
Therefore it suffices to prove that the canonical morphism
is an isomorphism. Since $\unicode[STIX]{x1D6F9}$ commutes with Verdier duality we can replace the $!$ -pushforwards with $\ast$ -pushforwards.
Fix $S\rightarrow \overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ with $S$ an affine scheme and apply Lemma 4.3.1 with $f=\operatorname{id}_{\mathbb{A}^{1}}\times \overline{\mathfrak{p}}^{-}$ ,
$\mathscr{F}=\operatorname{IC}_{\mathbb{A}^{1}}\boxtimes \,\mathfrak{j}_{\ast }^{-}(\unicode[STIX]{x1D714}_{\operatorname{Bun}_{B^{-}}})|_{\overline{\operatorname{Bun}}_{B^{-},\leqslant \unicode[STIX]{x1D708}}^{\unicode[STIX]{x1D706}}}$ , and $\mathscr{G}=\mathscr{W}|_{\mathbb{A}^{1}\times S}^{!}$ . Then the lemma yields an isomorphism
for $\unicode[STIX]{x1D706}$ sufficiently dominant. Changing $\unicode[STIX]{x1D706}$ if necessary so that $\unicode[STIX]{x1D706}\geqslant \unicode[STIX]{x1D707}$ , we can restrict this isomorphism along the map
By factorizability we obtain the desired isomorphism on $_{{\leqslant}\unicode[STIX]{x1D708}}\overline{Z}^{\unicode[STIX]{x1D707}}$ . Since $\unicode[STIX]{x1D708}$ was arbitrary, Theorem 4.2.1 is proved for the deepest strata.
Now we prove Theorem 1.3.1. Fix $\unicode[STIX]{x1D707}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ and choose $\unicode[STIX]{x1D706}\geqslant \unicode[STIX]{x1D707}$ dominant enough that $Z_{=\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}$ surjects smoothly onto $X^{\unicode[STIX]{x1D707}}\times _{\operatorname{Bun}_{T}}\operatorname{Bun}_{B}$ . Note that $(X^{\unicode[STIX]{x1D707}}\times \mathring{Z}^{\unicode[STIX]{x1D706}-\unicode[STIX]{x1D707}})_{\operatorname{disj}}$ is one of the connected components of the fiber product
and that the former surjects onto $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}},=\unicode[STIX]{x1D707}}$ . Theorem 4.2.1 for the deepest and open strata implies that the cohomologically normalized pullback of $\mathfrak{j}_{=\unicode[STIX]{x1D707}}^{!}\unicode[STIX]{x1D6F9}(\mathscr{W})$ to $(X^{\unicode[STIX]{x1D707}}\times \mathring{Z}^{\unicode[STIX]{x1D706}-\unicode[STIX]{x1D707}})_{\operatorname{disj}}$ is $\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D707}}\boxtimes \,\operatorname{IC}_{\mathring{Z}^{\unicode[STIX]{x1D706}-\unicode[STIX]{x1D707}}}$ . Theorem 1.3.1 follows once we observe that the composition
is the projection onto the first factor and apply Lemma 3.3.2.
The previous paragraph implies Theorem 4.2.1 holds on the stratum $Z_{=\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}$ . Let $\unicode[STIX]{x1D706}^{\prime }\geqslant \unicode[STIX]{x1D707}$ and change $\unicode[STIX]{x1D706}$ if necessary so that $\unicode[STIX]{x1D706}\geqslant \unicode[STIX]{x1D706}^{\prime }$ . By restricting along the morphism
and invoking factorization, we obtain Theorem 4.2.1.
The remainder of Theorems 1.3.1 and 4.2.1 follows as in the first proof.◻
5 First proof of Theorem 1.4.1
5.1
Like Theorem 1.3.1, we formulate the analogue of Theorem 1.4.1 on Zastava space.
Theorem 5.1.1. For any $\unicode[STIX]{x1D706}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ , there is an $\mathfrak{sl}_{2}$ -equivariant isomorphism of factorizable sheaves
Now we work out three of the simplest cases of Theorem 5.1.1. For brevity, we will write $\unicode[STIX]{x1D6F9}:=\unicode[STIX]{x1D6F9}(\,\!\mathscr{W}_{Z^{\unicode[STIX]{x1D706}}})$ .
Example 5.1.2. Let $\unicode[STIX]{x1D6FC}$ be a simple coroot. There is an isomorphism $Z^{\unicode[STIX]{x1D6FC}}\cong X\times \mathbb{A}^{1}$ under which $\mathring{Z}^{\unicode[STIX]{x1D6FC}}\cong X\times \mathring{A}^{1}$ , where $\mathring{A}^{1}:=\mathbb{A}^{1}\setminus \{0\}$ . The canonical map $\mathring{Z}^{\unicode[STIX]{x1D6FC}}\rightarrow \mathbb{A}^{1}$ is given in these terms by $(x,t)\mapsto 1/t$ . It follows from [Reference CampbellCam17, Example 4.3] that $\unicode[STIX]{x1D6F9}$ is the cohomologically normalized pullback of the unique indecomposable tilting sheaf on $\mathbb{A}^{1}$ which extends $\operatorname{IC}_{\mathring{A}^{1}}$ . Moreover, the monodromy filtration
satisfies $F_{-1}\cong \operatorname{IC}_{Z_{=\unicode[STIX]{x1D6FC}}^{\unicode[STIX]{x1D6FC}}}$ , $F_{0}/F_{-1}\cong \operatorname{IC}_{Z^{\unicode[STIX]{x1D6FC}}}$ , and $F_{1}/F_{0}\cong \operatorname{IC}_{Z_{=\unicode[STIX]{x1D6FC}}^{\unicode[STIX]{x1D6FC}}}$ , where $Z_{=\unicode[STIX]{x1D6FC}}^{\unicode[STIX]{x1D6FC}}=X\times \{0\}\subset X\times \mathbb{A}^{1}=Z^{\unicode[STIX]{x1D6FC}}$ . This confirms Theorem 5.1.1 in the case $\unicode[STIX]{x1D706}=\unicode[STIX]{x1D6FC}$ .
Example 5.1.3. Now consider the case $\unicode[STIX]{x1D706}=2\unicode[STIX]{x1D6FC}$ . By Theorem 4.2.1 for $\unicode[STIX]{x1D707}=2\unicode[STIX]{x1D6FC}$ , we have a short exact sequence
(recall that $\unicode[STIX]{x1D6FA}^{2\unicode[STIX]{x1D6FC}}$ is the clean extension of the sign local system on $X_{\operatorname{disj}}^{(2)}$ ). Similarly, applying Theorem 4.2.1 for $\unicode[STIX]{x1D707}=\unicode[STIX]{x1D6FC}$ and $0$ we obtain an exact triangle
where we used the fact that $\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D6FC}}\cong \operatorname{IC}_{X}$ . Applying Verdier duality to the equation in [Reference Braverman and GaitsgoryBG08, Corollary 4.5] (or rather the analogous equation on Zastava space), we have
where we identified $\unicode[STIX]{x1D6FA}^{2\unicode[STIX]{x1D6FC}}$ with its Verdier dual. Finally, one computes the simple constituents of $\text{}^{\prime }\mathfrak{j}_{=\unicode[STIX]{x1D6FC},\ast }^{2\unicode[STIX]{x1D6FC}}\operatorname{IC}_{Z_{=\unicode[STIX]{x1D6FC}}^{2\unicode[STIX]{x1D6FC}}}$ as follows. First, consider the short exact sequence
To compute the third term, observe that there is a Cartesian square
and that the $!$ -restriction of $\operatorname{IC}_{Z_{{\geqslant}\unicode[STIX]{x1D6FC}}^{2\unicode[STIX]{x1D6FC}}}$ along the top horizontal morphism is $\operatorname{IC}_{X^{2}}[-1]$ . Since $\text{}^{\prime }\mathfrak{j}_{{\geqslant}\unicode[STIX]{x1D6FC}}^{2\unicode[STIX]{x1D6FC}}$ is finite and birational onto its image, we can use base change to compute
Summarizing, we have
Now we will determine which graded component of $\operatorname{gr}\unicode[STIX]{x1D6F9}$ each simple subquotient lies in. In what follows, ‘weight’ refers to an eigenvalue of the Lefschetz Cartan operator. Since the monodromy filtration is compatible with the factorization structure, when we pull back $\operatorname{gr}\unicode[STIX]{x1D6F9}$ along the factorization map
we get $(\operatorname{gr}\unicode[STIX]{x1D6F9}(\,\!\mathscr{W}_{Z^{\unicode[STIX]{x1D6FC}}}))^{\boxtimes 2}$ restricted to $(Z^{\unicode[STIX]{x1D6FC}}\times Z^{\unicode[STIX]{x1D6FC}})_{\operatorname{disj}}$ . By the previous example, the latter sheaf with $\mathfrak{sl}_{2}$ -action is isomorphic to
It follows immediately that $\operatorname{IC}_{Z^{2\unicode[STIX]{x1D6FC}}}$ has weight $0$ and that the two copies of $\operatorname{IC}_{\overline{Z_{=\unicode[STIX]{x1D6FC}}^{2\unicode[STIX]{x1D6FC}}}}$ have weights $\pm 1$ . Since $\operatorname{std}^{\otimes 2}\cong V_{2}\oplus \operatorname{triv}$ , the three copies of $\unicode[STIX]{x1D6FA}^{2\unicode[STIX]{x1D6FC}}$ have weights $-2$ , $0$ , and $2$ , and $\operatorname{IC}_{Z_{=2\unicode[STIX]{x1D6FC}}^{2\unicode[STIX]{x1D6FC}}}$ has weight $0$ .
In terms of the monodromy filtration
we have $F_{-2}\cong \unicode[STIX]{x1D6FA}^{2\unicode[STIX]{x1D6FC}}\cong F_{2}/F_{0}$ , $F_{-1}/F_{-2}\cong \operatorname{IC}_{\overline{Z_{=\unicode[STIX]{x1D6FC}}^{2\unicode[STIX]{x1D6FC}}}}\cong F_{1}/F_{0}$ , and $F_{0}/F_{-1}$ has simple constituents $\operatorname{IC}_{Z^{2\unicode[STIX]{x1D6FC}}}$ , $\operatorname{IC}_{Z_{=2\unicode[STIX]{x1D6FC}}^{2\unicode[STIX]{x1D6FC}}}$ , and $\unicode[STIX]{x1D6FA}^{2\unicode[STIX]{x1D6FC}}$ . So in order to prove Theorem 5.1.1 in the case $\unicode[STIX]{x1D706}=2\unicode[STIX]{x1D6FC}$ , it remains to show that $F_{0}/F_{-1}$ is semisimple. Its pullback along the factorization map is semisimple, and since semisimplicity is étale local the restriction of $F_{0}/F_{-1}$ to $Z^{2\unicode[STIX]{x1D6FC}}\setminus \unicode[STIX]{x1D70B}^{-1}(X)$ is semisimple. But $\unicode[STIX]{x1D6F9}$ has no simple subquotients supported on $\unicode[STIX]{x1D70B}^{-1}(X)$ , so $F_{0}/F_{-1}$ is the intermediate extension of its restriction to $Z^{2\unicode[STIX]{x1D6FC}}\setminus \unicode[STIX]{x1D70B}^{-1}(X)$ and therefore semisimple.
Example 5.1.4. Suppose $\unicode[STIX]{x1D706}=\unicode[STIX]{x1D6FC}+\unicode[STIX]{x1D6FD}$ is a coroot, where $\unicode[STIX]{x1D6FC}$ and $\unicode[STIX]{x1D6FD}$ are distinct simple coroots. Applying Theorem 4.2.1 for $\unicode[STIX]{x1D707}=\unicode[STIX]{x1D6FC}+\unicode[STIX]{x1D6FD}$ , we obtain the short exact sequence
Similar considerations yield the short exact sequence
Applying [Reference Braverman and GaitsgoryBG08, Corollary 4.5], we have
where $\unicode[STIX]{x1D6F6}^{\unicode[STIX]{x1D706}}$ is the Verdier dual of $\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D706}}$ . According to [Reference Arinkin and GaitsgoryAG15b, § 1.3.2], in this case $\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D706}}$ is the $\ast$ -extension of $\operatorname{IC}_{X_{\operatorname{disj}}^{2}}$ to $X^{\unicode[STIX]{x1D706}}=X^{2}$ , whence $\unicode[STIX]{x1D6F6}^{\unicode[STIX]{x1D706}}$ is the $!$ -extension. In particular $[\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D706}}]=[\unicode[STIX]{x1D6F6}^{\unicode[STIX]{x1D706}}]=[\operatorname{IC}_{X^{2}}]+[\operatorname{IC}_{X}]$ . As for the remaining simple constituents, consider the short exact sequence
and similarly for $\unicode[STIX]{x1D6FD}$ . The third term is $\operatorname{IC}_{X^{2}}$ , so finally we see that
Now we compute the weights of the simple subquotients of $\unicode[STIX]{x1D6F9}$ . We have the factorization morphism
and after pulling back $\operatorname{gr}\unicode[STIX]{x1D6F9}$ the result is (5.1.2), up to relabeling $\unicode[STIX]{x1D6FD}$ as $\unicode[STIX]{x1D6FC}$ . As in Example 5.1.3, it follows that $\operatorname{IC}_{Z^{\unicode[STIX]{x1D6FC}+\unicode[STIX]{x1D6FD}}}$ has weight $0$ , the two copies of $\operatorname{IC}_{\overline{Z_{=\unicode[STIX]{x1D6FC}}^{\unicode[STIX]{x1D706}}}}$ have weights $\pm 1$ and likewise for $\operatorname{IC}_{\overline{Z_{=\unicode[STIX]{x1D6FD}}^{\unicode[STIX]{x1D706}}}}$ , and the four copies of $\operatorname{IC}_{X^{2}}$ have weights $-2$ , $0$ , $0$ , and $2$ . We will see below that in any case where $\unicode[STIX]{x1D706}$ is a coroot, there are two simple subquotients of $\unicode[STIX]{x1D6F9}$ isomorphic to $\operatorname{IC}_{X}$ , with weights $\pm 1$ . In terms of the monodromy filtration, we have $F_{-2}\cong \operatorname{IC}_{X^{2}}\cong F_{2}/F_{1}$ , $F_{-1}/F_{-2}$ and $F_{1}/F_{0}$ each have simple subquotients $\operatorname{IC}_{\overline{Z_{=\unicode[STIX]{x1D6FC}}^{\unicode[STIX]{x1D706}}}}$ , $\operatorname{IC}_{\overline{Z_{=\unicode[STIX]{x1D6FD}}^{\unicode[STIX]{x1D706}}}}$ , and $\operatorname{IC}_{X}$ , and $F_{0}/F_{-1}$ has simple subquotients $\operatorname{IC}_{Z^{\unicode[STIX]{x1D706}}}$ and $\operatorname{IC}_{X^{2}}$ , the latter with multiplicity two. As in Example 5.1.3, one uses factorization to show that $F_{0}/F_{-1}$ is semisimple. To prove the semisimplicity of $F_{-1}/F_{-2}$ and $F_{1}/F_{0}$ , it is enough to show that there are no extensions between $\operatorname{IC}_{X}$ and $\operatorname{IC}_{\overline{Z_{=\unicode[STIX]{x1D6FC}}^{\unicode[STIX]{x1D706}}}}$ . By Verdier duality, it suffices to prove that
We have $\text{}^{\prime }\mathfrak{j}_{=\unicode[STIX]{x1D706}}^{\unicode[STIX]{x1D706},!}\operatorname{IC}_{\overline{Z_{=\unicode[STIX]{x1D6FC}}^{\unicode[STIX]{x1D706}}}}=\operatorname{IC}_{X^{2}}[-1]$ , whence $\unicode[STIX]{x1D6E5}^{!}\operatorname{IC}_{\overline{Z_{=\unicode[STIX]{x1D6FC}}^{\unicode[STIX]{x1D706}}}}=\operatorname{IC}_{X}[-2]$ . Thus
and in particular $\operatorname{Ext}^{1}$ vanishes.
5.2
The following lemma will be used in both proofs of Theorems 1.4.1 and 5.1.1. For any $\unicode[STIX]{x1D707}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ write
for the finite birational map defined as the composition of $\mathfrak{j}_{{\geqslant}\unicode[STIX]{x1D707}}$ and the embedding
induced by the diagonal map $X\rightarrow X^{\unicode[STIX]{x1D707}}$ .
Lemma 5.2.1. If $\unicode[STIX]{x1D707}$ is a coroot, then there is an indecomposable subquotient $\mathscr{M}$ of $\unicode[STIX]{x1D6F9}(\mathscr{W})$ with a filtration
such that $\mathscr{M}_{-1}\cong \unicode[STIX]{x1D6E5}_{\ast }^{\unicode[STIX]{x1D707}}\operatorname{IC}_{X\times _{\operatorname{ Bun}_{T}}\overline{\operatorname{Bun}}_{B}}$ , $\mathscr{M}_{0}/\mathscr{M}_{-1}\cong \operatorname{IC}_{\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}}$ , and $\mathscr{M}/\mathscr{M}_{0}\cong \unicode[STIX]{x1D6E5}_{\ast }^{\unicode[STIX]{x1D707}}\operatorname{IC}_{X\times _{\operatorname{ Bun}_{T}}\overline{\operatorname{Bun}}_{B}}$ .
Proof. First, we claim the subsheaf $\mathfrak{j}_{=0,!}\operatorname{IC}_{\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}}$ of $\unicode[STIX]{x1D6F9}(\mathscr{W})$ has a quotient $\mathscr{M}_{0}$ of the form described above. Recall that $\mathfrak{j}_{=0,!}\operatorname{IC}_{\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}}$ has a descending filtration with subquotients $\mathfrak{j}_{=\unicode[STIX]{x1D708},!\ast }\mathfrak{r}_{\unicode[STIX]{x1D708}}^{\unicode[STIX]{x1D6E5}}\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D708}}$ , and in particular has $\mathfrak{j}_{=\unicode[STIX]{x1D707},!\ast }\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D707}}$ as a subquotient and $\operatorname{IC}_{\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}}$ as a quotient. The former sheaf has $\unicode[STIX]{x1D6E5}_{\ast }^{\unicode[STIX]{x1D707}}\operatorname{IC}_{X\times _{\operatorname{ Bun}_{T}}\overline{\operatorname{Bun}}_{B}}$ as a quotient because $\unicode[STIX]{x1D707}$ is a coroot, so it suffices to show that for any $0<\unicode[STIX]{x1D708}<\unicode[STIX]{x1D707}$ we have
or dually
where $\unicode[STIX]{x1D6F6}^{\unicode[STIX]{x1D708}}$ is the Verdier dual of $\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D708}}$ . One computes using base change that
Since $\unicode[STIX]{x1D6E5}^{\unicode[STIX]{x1D707},!}\operatorname{IC}_{\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}}$ is concentrated in cohomological degrees greater than or equal to $1$ , and both $\unicode[STIX]{x1D6E5}^{!}\unicode[STIX]{x1D6F6}^{\unicode[STIX]{x1D708}}$ and $\unicode[STIX]{x1D6E5}^{\unicode[STIX]{x1D707},!}\operatorname{IC}_{\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}}$ have lisse (actually constant) cohomology sheaves, their $!$ -tensor product is concentrated in cohomological degrees greater than or equal to $2$ . It follows that the $\operatorname{Ext}^{1}$ above vanishes.
We have shown that $\mathfrak{j}_{=0,!}\operatorname{IC}_{\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}}$ has a quotient $\mathscr{M}_{0}$ which fits into a short exact sequence
This sequence does not split because $\mathscr{M}_{0}$ is the quotient of the indecomposable sheaf $\mathfrak{j}_{=0,!}\operatorname{IC}_{\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}}$ with simple cosocle $\operatorname{IC}_{\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}}\!$ . Dually, we obtain a subsheaf $\mathscr{M}/\mathscr{M}_{-1}$ of $\mathfrak{j}_{=0,\ast }\operatorname{IC}_{\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}}$ of the desired form, from which follows the existence of $\mathscr{M}$ .◻
We will first give a proof of Theorem 1.4.1 under the assumption that $\operatorname{gr}\unicode[STIX]{x1D6F9}(\mathscr{W})$ is semisimple. The semisimplicity can be proved via Mochizuki’s theory of weights for holonomic $D$ -modules, since $\mathscr{W}$ is pure and (up to shift) the monodromy filtration on nearby cycles of a pure sheaf coincides with the weight filtration (see [Reference MochizukiMoc15, Corollary 9.1.10]).
First proof of Theorems 1.4.1 and 5.1.1.
Both sides of the isomorphism (1.4.1) restrict to $\operatorname{IC}_{\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}}$ . Suppose that we have constructed the isomorphism over $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}},<\unicode[STIX]{x1D707}}$ . Then for $\unicode[STIX]{x1D706}\geqslant \unicode[STIX]{x1D707}$ sufficiently dominant, pulling back along $\text{}^{\prime }\mathfrak{p}^{-}$ yields the isomorphism (5.1.1) over $Z_{{<}\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}$ . One obtains (5.1.1) on $Z_{{<}\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D707}}$ by pullback along the factorization map
since the inverse images of both sides of the isomorphism factorize and are constant along the second component. The same argument yields (5.1.1) on $Z_{{<}\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D707}^{\prime }}=Z^{\unicode[STIX]{x1D707}^{\prime }}$ for $\unicode[STIX]{x1D707}^{\prime }<\unicode[STIX]{x1D707}$ .
On the other hand, one can use factorization to obtain (5.1.1) on $Z^{\unicode[STIX]{x1D707}}\setminus \unicode[STIX]{x1D70B}^{-1}(X)$ . Namely, for $\unicode[STIX]{x1D707}_{1}+\unicode[STIX]{x1D707}_{2}=\unicode[STIX]{x1D707}$ , $\unicode[STIX]{x1D707}_{1},\unicode[STIX]{x1D707}_{2}<\unicode[STIX]{x1D707}$ , the pullback of both sides of the isomorphism along
are identified. Since this factorization map is étale but not necessarily an embedding, we must argue that the isomorphism descends to its image. This immediately reduces to the case that $\unicode[STIX]{x1D707}=n\cdot \unicode[STIX]{x1D6FC}$ for some $\unicode[STIX]{x1D6FC}\in \unicode[STIX]{x1D6E5}$ . In this case, both sides of (5.1.1) are the intermediate extension of their restriction to $\unicode[STIX]{x1D70B}^{-1}(X_{\operatorname{disj}}^{\unicode[STIX]{x1D707}})$ , so it suffices to show that the isomorphism over $(Z^{\unicode[STIX]{x1D6FC}})_{\operatorname{disj}}^{n}$ descends to $\unicode[STIX]{x1D70B}^{-1}(X_{\operatorname{disj}}^{\unicode[STIX]{x1D707}})\subset Z^{\unicode[STIX]{x1D707}}$ . Since $(Z^{\unicode[STIX]{x1D6FC}})_{\operatorname{disj}}^{n}$ is a $\unicode[STIX]{x1D6F4}_{n}$ -torsor over $\unicode[STIX]{x1D70B}^{-1}(X_{\operatorname{disj}}^{\unicode[STIX]{x1D707}})$ and $\unicode[STIX]{x1D6F4}_{n}$ is generated by transpositions, the claim reduces to the case $n=2$ . But this was already done in Example 5.1.3.
Note that $Z_{{<}\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D707}}\cup (Z^{\unicode[STIX]{x1D707}}\setminus \unicode[STIX]{x1D70B}^{-1}(X))=Z^{\unicode[STIX]{x1D707}}\setminus \unicode[STIX]{x1D6E5}(X)$ . The isomorphisms of the previous two paragraphs clearly agree on $Z_{{<}\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D707}}\cap (Z^{\unicode[STIX]{x1D707}}\setminus \unicode[STIX]{x1D70B}^{-1}(X))$ , hence glue to an isomorphism away from the main diagonal.
If $\unicode[STIX]{x1D707}$ is not a coroot, then we claim that $\unicode[STIX]{x1D6F9}(\,\!\mathscr{W}_{Z^{\unicode[STIX]{x1D707}}})$ has no simple subquotients supported on the main diagonal, whence $\operatorname{gr}\unicode[STIX]{x1D6F9}(\mathscr{W}_{Z^{\unicode[STIX]{x1D707}}})$ is the intermediate extension of its restriction to $Z^{\unicode[STIX]{x1D707}}\setminus \unicode[STIX]{x1D6E5}(X)$ . This is true for the right-hand side of (5.1.1) by construction, so the claim implies that the isomorphism extends to $Z^{\unicode[STIX]{x1D707}}$ in this case. By Theorem 4.2.1 there is a filtration of $\unicode[STIX]{x1D6F9}(\,\!\mathscr{W}_{Z^{\unicode[STIX]{x1D707}}})$ by the sheaves $\text{}^{\prime }\mathfrak{j}_{=\unicode[STIX]{x1D708},\ast }^{\unicode[STIX]{x1D707}}\text{}^{\prime }\mathfrak{m}_{\unicode[STIX]{x1D708}}^{\unicode[STIX]{x1D707},\unicode[STIX]{x1D6E5}}\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D708}}$ for $0\leqslant \unicode[STIX]{x1D708}\leqslant \unicode[STIX]{x1D707}$ . Using [Reference Braverman and GaitsgoryBG08, Corollary 4.5], one can show that $\text{}^{\prime }\mathfrak{j}_{=\unicode[STIX]{x1D708},\ast }^{\unicode[STIX]{x1D707}}\text{}^{\prime }\mathfrak{m}_{\unicode[STIX]{x1D708}}^{\unicode[STIX]{x1D707},\unicode[STIX]{x1D6E5}}\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D708}}$ surjects onto $\text{}^{\prime }\mathfrak{j}_{=\unicode[STIX]{x1D707},\ast }^{\unicode[STIX]{x1D707}}\operatorname{add}_{\ast }(\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D708}}\boxtimes \,\unicode[STIX]{x1D6F6}^{\unicode[STIX]{x1D707}-\unicode[STIX]{x1D708}})$ , and that no subquotient of the kernel is supported on $X^{\unicode[STIX]{x1D707}}$ . Now the claim follows, because out of the latter sheaves only $\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D707}}$ and $\unicode[STIX]{x1D6F6}^{\unicode[STIX]{x1D707}}$ could have subquotients supported on the diagonal, and by § 3.3 of [Reference Braverman and GaitsgoryBG08] this occurs if and only if $\unicode[STIX]{x1D707}$ is a coroot.
Suppose that $\unicode[STIX]{x1D707}$ is a coroot. Then $\unicode[STIX]{x1D6E5}_{\ast }\operatorname{IC}_{X}$ appears as a subquotient of $\unicode[STIX]{x1D6FA}^{\unicode[STIX]{x1D707}}$ and of $\unicode[STIX]{x1D6F6}^{\unicode[STIX]{x1D707}}$ with multiplicity one. By the analysis in the previous paragraph $\unicode[STIX]{x1D6E5}_{\ast }\operatorname{IC}_{X}$ appears as a summand of $\operatorname{gr}\unicode[STIX]{x1D6F9}(\,\!\mathscr{W}_{Z^{\unicode[STIX]{x1D707}}})$ with multiplicity two, and there are no other subquotients supported on the main diagonal. Thus the isomorphism (5.1.1) extends to $Z^{\unicode[STIX]{x1D707}}$ , and it remains to show that $\mathfrak{sl}_{2}$ acts on the summand $\operatorname{IC}_{X}^{\oplus 2}$ of $\operatorname{gr}\unicode[STIX]{x1D6F9}(\,\!\mathscr{W}_{Z^{\unicode[STIX]{x1D707}}})$ as the standard representation.
The only other possibility is that $\mathfrak{sl}_{2}$ acts on $\operatorname{IC}_{X}^{\oplus 2}$ trivially, which would imply that the subquotient $\mathscr{M}$ from Lemma 5.2.1 is a subquotient of $F_{0}/F_{-1}$ . But $\mathscr{M}$ is indecomposable and $F_{0}/F_{-1}$ is semisimple, so this is impossible.
Having constructed the isomorphism of Theorem 5.1.1 over $Z^{\unicode[STIX]{x1D707}}$ , we can complete the inductive step of Theorem 1.4.1 by extending the isomorphism from $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}},<\unicode[STIX]{x1D707}}$ to $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}},\leqslant \unicode[STIX]{x1D707}}$ . Choose $\unicode[STIX]{x1D706}\geqslant \unicode[STIX]{x1D707}$ dominant enough that $Z_{{\leqslant}\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}$ surjects smoothly onto $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}},\leqslant \unicode[STIX]{x1D707}}$ . As in the proof of Theorem 1.3.1, note that $(Z^{\unicode[STIX]{x1D707}}\times \mathring{Z}^{\unicode[STIX]{x1D706}-\unicode[STIX]{x1D707}})_{\operatorname{disj}}$ is one of the connected components of the fiber product
and that the former surjects onto $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}},\leqslant \unicode[STIX]{x1D707}}$ . By factorization, the cohomologically normalized pullback of $\operatorname{gr}\unicode[STIX]{x1D6F9}(\mathscr{W})$ to $(Z^{\unicode[STIX]{x1D707}}\times \mathring{Z}^{\unicode[STIX]{x1D706}-\unicode[STIX]{x1D707}})_{\operatorname{disj}}$ is $(\operatorname{gr}\unicode[STIX]{x1D6F9}(\mathscr{W}_{Z^{\unicode[STIX]{x1D707}}}))\boxtimes \operatorname{IC}_{\mathring{Z}^{\unicode[STIX]{x1D706}-\unicode[STIX]{x1D707}}}$ .
A factorization argument as in the proof of Theorem 4.2.1 allows us to construct the isomorphism (5.1.1) over $Z_{{\leqslant}\unicode[STIX]{x1D707}}^{\unicode[STIX]{x1D706}}$ for arbitrary $\unicode[STIX]{x1D706}\geqslant \unicode[STIX]{x1D707}$ , which completes the proof of Theorem 5.1.1.◻
Proof of Corollary 1.4.1.1.
It suffices to prove the corresponding claim on $Z^{\unicode[STIX]{x1D706}}$ for any $\unicode[STIX]{x1D706}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ . The morphism
is injective because $\unicode[STIX]{x1D6F9}(\!\,\mathscr{W})$ is tilting, so it suffices to show that $\text{}^{\prime }\mathfrak{j}_{=0,!}^{\unicode[STIX]{x1D706}}\operatorname{IC}_{\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}}$ and the kernel of monodromy have the same class in the Grothendieck group. By factorization and induction this holds away from the main diagonal, and [Reference Braverman and GaitsgoryBG08, Corollary 4.5] implies that the only subquotient of $\text{}^{\prime }\mathfrak{j}_{=0,!}^{\unicode[STIX]{x1D706}}\operatorname{IC}_{\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}}$ supported on the main diagonal is $\unicode[STIX]{x1D6E5}_{\ast }\operatorname{IC}_{X}$ with multiplicity one. Theorem 5.1.1 implies that the same is true for the kernel of the monodromy operator on $\unicode[STIX]{x1D6F9}(\!\,\mathscr{W})$ .◻
6 Second proof of Theorem 1.4.1
6.1
In this section we will give a proof of Theorems 1.4.1 and 5.1.1 which does not use weights for irregular holonomic $D$ -modules to prove the semisimplicity of $\operatorname{gr}\unicode[STIX]{x1D6F9}(\mathscr{W})$ , but instead depends on Conjecture 1.5.2 (but only the part proved in [Reference Feigin, Finkelberg, Kuznetsov and MirkovicFFKM99]). First we make the statement of the conjecture more precise in the case of a trivial ${\check{T}}$ -local system by specifying the action of generators of $\check{\mathfrak{g}}\otimes H^{\bullet }(X)$ on $\overline{\mathfrak{p}}_{\ast }\operatorname{IC}_{\overline{\operatorname{Bun}}_{B}}$ .
We construct the action of $\check{\mathfrak{h}}\otimes H^{\bullet }(X)$ as follows. Pullback along the evaluation map
defines a homomorphism
By adjunction we obtain a morphism $\mathfrak{\{}\otimes H_{\bullet }(X)[-2]\longrightarrow H^{\bullet }(\operatorname{Bun}_{T})$ . Identifying $\mathfrak{h}^{\ast }\cong \check{\mathfrak{h}}$ and $H_{\bullet }(X)[-2]\cong H^{\bullet }(X)$ , the latter using Poincaré duality, we obtain a morphism
Then the action of $H^{\bullet }(\operatorname{Bun}_{T})$ on $\unicode[STIX]{x1D714}_{\operatorname{Bun}_{T}}$ induces by functoriality the desired action of $\check{\mathfrak{h}}\otimes H^{\bullet }(X)$ on $\operatorname{Eis}_{!\ast }\unicode[STIX]{x1D714}_{\operatorname{Bun}_{T}}=\overline{\mathfrak{p}}_{\ast }\operatorname{IC}_{\overline{\operatorname{Bun}}_{B}}$ .
Next we construct the action of $\check{\mathfrak{n}}\otimes H^{\bullet }(X)$ . Denote by $\mathscr{U}(\check{\mathfrak{n}})$ the factorization algebra whose fiber at $\sum _{i}\unicode[STIX]{x1D707}_{i}x_{i}\in X^{\unicode[STIX]{x1D707}}$ is
where the superscript $\unicode[STIX]{x1D707}_{i}$ indicates the corresponding ${\check{T}}$ -graded component. The following is part of [Reference Braverman and GaitsgoryBG08, Theorem 5.6], which itself is a restatement of results from [Reference Feigin, Finkelberg, Kuznetsov and MirkovicFFKM99]: for any $\unicode[STIX]{x1D707}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ there is a canonical morphism
which induces an isomorphism
Here we abuse notation slightly by denoting the maps $\mathfrak{j}_{{\geqslant}\unicode[STIX]{x1D707}}:X^{\unicode[STIX]{x1D707}}\times \overline{\operatorname{Bun}}_{B}\rightarrow \overline{\operatorname{Bun}}_{B}$ and $\mathfrak{j}_{=\unicode[STIX]{x1D707}}:X^{\unicode[STIX]{x1D707}}\times \operatorname{Bun}_{B}\rightarrow \overline{\operatorname{Bun}}_{B}$ by the same symbols we used in the case of $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ . The latter are obtained from the former by base change along $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}\rightarrow \overline{\operatorname{Bun}}_{B}$ .
Pushing forward to $\operatorname{Bun}_{G}$ , we obtain a morphism
The object $\mathscr{U}(\check{\mathfrak{n}})^{\unicode[STIX]{x1D707}}$ is concentrated in (perverse) cohomological degrees greater than or equal to $1$ , and if $\unicode[STIX]{x1D707}$ is a coroot then we have $H^{1}(\mathscr{U}(\check{\mathfrak{n}})^{\unicode[STIX]{x1D707}})=\operatorname{IC}_{X}$ . The resulting morphism $k_{X}\rightarrow \mathscr{U}(\check{\mathfrak{n}})^{\unicode[STIX]{x1D707}}$ induces
which defines the action of $\check{\mathfrak{n}}_{\unicode[STIX]{x1D707}}\otimes H^{\bullet }(X)\cong H^{\bullet }(X)$ on $\overline{\mathfrak{p}}_{\ast }\operatorname{IC}_{\overline{\operatorname{Bun}}_{B}}$ .
Dually, for any $\unicode[STIX]{x1D707}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ there is a canonical morphism
which induces an isomorphism
Here $\mathscr{U}^{\vee }(\check{\mathfrak{n}}^{-})$ is by definition the Verdier dual of $\mathscr{U}(\check{\mathfrak{n}}^{-})$ .
Thus we obtain a morphism
or by adjunction
If $\unicode[STIX]{x1D707}$ is a coroot, then as before we have a morphism $H^{\bullet }(X)\rightarrow H^{\bullet }(X^{\unicode[STIX]{x1D707}},\mathscr{U}(\check{\mathfrak{n}}^{-})^{\unicode[STIX]{x1D707}})$ , which defines the action of $\check{\mathfrak{n}}_{-\unicode[STIX]{x1D707}}^{-}\otimes H^{\bullet }(X)\cong H^{\bullet }(X)$ on $\overline{\mathfrak{p}}_{\ast }\operatorname{IC}_{\overline{\operatorname{Bun}}_{B}}$ .
6.2
Fix a coroot $\unicode[STIX]{x1D707}$ . Recall the subquotient $\mathscr{M}$ of $\unicode[STIX]{x1D6F9}(\mathscr{W})$ from Lemma 5.2.1. The action of $\check{\mathfrak{g}}$ on $\overline{\mathfrak{p}}_{\ast }\operatorname{IC}_{\overline{\operatorname{Bun}}_{B}}$ relates to our problem through the following key lemma, whose proof will occupy this subsection.
Lemma 6.2.1. The sheaf $\mathscr{M}$ does not descend to $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}/\mathbb{G}_{m}$ .
First, observe that $\mathscr{M}_{0}$ and $\mathscr{M}/\mathscr{M}_{-1}$ descend to $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}/T$ and hence to $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}/\mathbb{G}_{m}$ , being subquotients of $\mathfrak{j}_{=0,!}\operatorname{IC}_{\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}}$ and $\mathfrak{j}_{=0,\ast }\operatorname{IC}_{\operatorname{Bun}_{N^{\unicode[STIX]{x1D714}}}}$ respectively. The obstruction to descent of $\mathscr{M}$ to $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}/\mathbb{G}_{m}$ is the resulting composition
Similarly, the obstruction to its descent to $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}/T$ is the composition
Denote by ${\unicode[STIX]{x1D6E5}\unicode[STIX]{x0030A}}^{\unicode[STIX]{x1D707}}:X\times _{\operatorname{Bun}_{T}}\operatorname{Bun}_{B}\rightarrow \overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}$ the locally closed embedding given by composing $\unicode[STIX]{x1D6E5}^{\unicode[STIX]{x1D707}}$ with the open embedding $X\times _{\operatorname{Bun}_{T}}\operatorname{Bun}_{B}\rightarrow X\times _{\operatorname{Bun}_{T}}\overline{\operatorname{Bun}}_{B}$ . Composition with the canonical morphisms
and
gives
since the map $(X\times _{\operatorname{Bun}_{T}}\operatorname{Bun}_{B})/T\rightarrow X\times \operatorname{pt}/T$ induces an isomorphism on cohomology.
Lemma 6.2.2. The image of the endomorphism (6.2.2) under (6.2.3) is
Proof. Theorem 5.1.5 in [Reference Braverman and GaitsgoryBG02] says that $\operatorname{IC}_{\overline{\operatorname{Bun}}_{B}}$ is ULA over $\operatorname{Bun}_{T}$ , which implies that the $!$ -restriction of $\operatorname{IC}_{\overline{\operatorname{Bun}}_{B}}$ to $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}/T$ is $\operatorname{IC}_{\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}/T}[\dim T-\dim \operatorname{Bun}_{T}]$ . It follows that the $!$ -restriction of $\unicode[STIX]{x1D6E5}_{\ast }^{\unicode[STIX]{x1D707}}\operatorname{IC}_{X\times \overline{\operatorname{Bun}}_{B}}$ to $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}/T$ is a shift of $\unicode[STIX]{x1D6E5}_{\ast }^{\unicode[STIX]{x1D707}}\operatorname{IC}_{(X\times _{\operatorname{ Bun}_{T}}\overline{\operatorname{Bun}}_{B})/T}$ , where we abusively write $\unicode[STIX]{x1D6E5}^{\unicode[STIX]{x1D707}}:X\times \overline{\operatorname{Bun}}_{B}\rightarrow \overline{\operatorname{Bun}}_{B}$ for the similarly defined finite map. This gives rise to a commutative square
where the upper horizontal arrow is defined similarly to (6.2.3) and the right vertical arrow is $\operatorname{id}_{X}$ times restriction along $\unicode[STIX]{x1D70C}(\unicode[STIX]{x1D714}):\operatorname{pt}/T\rightarrow \operatorname{Bun}_{T}$ . The previous subsection implies that $\mathscr{M}_{0}$ and $\mathscr{M}/\mathscr{M}_{-1}$ extend to $\overline{\operatorname{Bun}}_{B}$ , giving rise to a morphism
which restricts to (6.2.2) on $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}/T$ . Thus it suffices to show that the image of (6.2.4) in $H^{2}(X\times \operatorname{Bun}_{T})$ restricts to $-1\otimes h_{\unicode[STIX]{x1D707}}$ along $\operatorname{id}_{X}\times \unicode[STIX]{x1D70C}(\unicode[STIX]{x1D714})$ .
Observe that
so $\overline{\mathfrak{p}}_{\ast }$ induces a morphism
Composing with the canonical map $k\rightarrow H^{\bullet }(X)$ and its dual $H_{\bullet }(X)\rightarrow k$ , we obtain
By construction, the image of (6.2.2) under (6.2.5) coincides with the action of $f_{\unicode[STIX]{x1D707}}e_{\unicode[STIX]{x1D707}}\in U(\mathfrak{{\check{g}}})$ . By composing with the morphisms $\mathfrak{p}_{!}\operatorname{IC}_{\operatorname{Bun}_{B}}\rightarrow \overline{\mathfrak{p}}_{\ast }\operatorname{IC}_{\overline{\operatorname{Bun}}_{B}}$ and $\overline{\mathfrak{p}}_{\ast }\operatorname{IC}_{\overline{\operatorname{Bun}}_{B}}\rightarrow \mathfrak{p}_{\ast }\operatorname{IC}_{\operatorname{Bun}_{B}}$ , we obtain
Note that (6.2.6) annihilates the endomorphism of $\overline{\mathfrak{p}}_{\ast }\operatorname{IC}_{\overline{\operatorname{Bun}}_{B}}$ given by the action of $e_{\unicode[STIX]{x1D707}}f_{\unicode[STIX]{x1D707}}$ , since it factors through a sheaf supported on the boundary. Proposition 4.9 in [Reference Feigin, Finkelberg, Kuznetsov and MirkovicFFKM99] says that the relation $[e_{\unicode[STIX]{x1D707}},f_{\unicode[STIX]{x1D707}}]=h_{\unicode[STIX]{x1D707}}$ holds in $\operatorname{End}(\overline{\mathfrak{p}}_{\ast }\operatorname{IC}_{\overline{\operatorname{Bun}}_{B}})$ , which implies that the images of $f_{\unicode[STIX]{x1D707}}e_{\unicode[STIX]{x1D707}}$ and $-h_{\unicode[STIX]{x1D707}}$ under (6.2.6) coincide.
Now consider the commutative square
where the right vertical morphism is the composition
Note that $\operatorname{id}_{X}\times \unicode[STIX]{x1D70C}(\unicode[STIX]{x1D714}):X\times \operatorname{pt}/T\rightarrow X\times \operatorname{Bun}_{T}$ admits a canonical retraction, given by the projection $X\times \operatorname{Bun}_{T}\rightarrow X$ and the evaluation map $X\times \operatorname{Bun}_{T}\rightarrow \operatorname{pt}/T$ . So far we have shown that the image of (6.2.2) under the resulting composition
agrees with the image of $-1\otimes h_{\unicode[STIX]{x1D707}}$ under
so it suffices to show that the latter composition is injective.
We have $H^{2}(X\times \operatorname{pt}/T)=\check{\mathfrak{h}}\oplus H^{2}(X)$ , and we have already used the fact that for each $\unicode[STIX]{x1D706}\in \unicode[STIX]{x1D6EC}$ , an element $h\in \check{\mathfrak{h}}$ maps to $\langle h,\unicode[STIX]{x1D706}\rangle$ times the canonical morphism $\mathfrak{p}_{!}\operatorname{IC}_{\operatorname{Bun}_{B}^{\unicode[STIX]{x1D706}}}\rightarrow \mathfrak{p}_{\ast }\operatorname{IC}_{\operatorname{Bun}_{B}^{\unicode[STIX]{x1D706}}}$ . One checks that the canonical generator of $H^{2}(X)$ maps to canonical map $\mathfrak{p}_{!}\operatorname{IC}_{\operatorname{Bun}_{B}}\rightarrow \mathfrak{p}_{\ast }\operatorname{IC}_{\operatorname{Bun}_{B}}$ itself, which proves the desired injectivity.◻
Note that Lemma 6.2.2 already implies that $\mathscr{M}$ does not descend to $\overline{\operatorname{Bun}}_{N^{\unicode[STIX]{x1D714}}}/T$ , since $h_{\unicode[STIX]{x1D707}}\neq 0$ .
Proof of Lemma 6.2.1.
The morphism (6.2.1) induces an element of $H^{2}(X\times \operatorname{pt}/\mathbb{G}_{m})$ in the same way that (6.2.2) gives rise to $-1\otimes h_{\unicode[STIX]{x1D707}}\in H^{2}(X\times \operatorname{pt}/T)$ . Moreover, these constructions fit into a commutative square
where the vertical morphisms are induced by $\unicode[STIX]{x1D6FE}$ , and in particular (6.2.2) maps to (6.2.1) along the left vertical morphism. The image of $h_{\unicode[STIX]{x1D707}}$ under $H^{2}(\operatorname{pt}/T)\rightarrow H^{2}(\operatorname{pt}/\mathbb{G}_{m})=k$ is the positive integer $\langle h_{\unicode[STIX]{x1D707}},\unicode[STIX]{x1D6FE}\rangle$ , so the lemma follows.◻
6.3
We need another, more elementary lemma. Fix $\unicode[STIX]{x1D707}\in \unicode[STIX]{x1D6EC}^{\operatorname{pos}}$ and $\mathfrak{k}\in \operatorname{Kost}(\unicode[STIX]{x1D707})$ given by $\unicode[STIX]{x1D707}=\sum n_{\unicode[STIX]{x1D6FD}}\unicode[STIX]{x1D6FD}$ .
Lemma 6.3.1. If $\mathscr{L}$ is a nonconstant simple summand of $\boxtimes _{\unicode[STIX]{x1D6FD}\in R^{+}}\mathscr{P}_{n_{\unicode[STIX]{x1D6FD}}}$ , then
Proof. This follows from the fact that the local system on $X_{\operatorname{disj}}^{(n)}$ associated to a nontrivial irreducible $\unicode[STIX]{x1D6F4}_{n}$ -representation extends cleanly over the main diagonal.◻
Second proof of Theorems 1.4.1 and 5.1.1.
We proceed as in the first proof, until we have constructed the isomorphism (5.1.1) over $Z^{\unicode[STIX]{x1D707}}\setminus \unicode[STIX]{x1D6E5}(X)$ and reduced to the case that $\unicode[STIX]{x1D707}$ is a coroot. We showed that in this case $\unicode[STIX]{x1D6F9}(\,\!\mathscr{W}_{Z^{\unicode[STIX]{x1D707}}})$ contains $\unicode[STIX]{x1D6E5}_{\ast }\operatorname{IC}_{X}$ with multiplicity two but has no other subquotients supported on the main diagonal (in particular, the two sides of (5.1.1) agree in the Grothendieck group). Either $\mathfrak{sl}_{2}$ acts trivially on both copies of $\unicode[STIX]{x1D6E5}_{\ast }\operatorname{IC}_{X}$ or they have weights $1$ and $-1$ . In order to show that (5.1.1) extends to $Z^{\unicode[STIX]{x1D707}}$ we need to rule out the first case, and then show that $\operatorname{gr}\unicode[STIX]{x1D6F9}(\,\!\mathscr{W}_{Z^{\unicode[STIX]{x1D707}}})$ is semisimple.
We must rule out the possibility that $\mathfrak{sl}_{2}$ acts trivially on the subquotient $\mathscr{M}$ from Lemma 5.2.1. Since the monodromy endomorphism of $\unicode[STIX]{x1D6F9}(\,\!\mathscr{W}_{Z^{\unicode[STIX]{x1D707}}})$ is the obstruction to $\mathbb{G}_{m}$ -equivariance, this would imply that $\mathscr{M}$ is $\mathbb{G}_{m}$ -equivariant, contradicting Lemma 5.2.1.
Now we finish the proof that $\operatorname{gr}\unicode[STIX]{x1D6F9}(\,\!\mathscr{W}_{Z^{\unicode[STIX]{x1D707}}})$ is semisimple. Using the $\mathfrak{sl}_{2}$ -action, it decomposes into the direct sum of its isotypic components, indexed by the irreducible $\mathfrak{sl}_{2}$ -representations. The previous paragraph says that the two copies of $\unicode[STIX]{x1D6E5}_{\ast }\operatorname{IC}_{X}$ are subquotients of the $\operatorname{std}$ -isotypic component. Thus the other isotypic components have no subquotients supported on the main diagonal, so they are the same as the corresponding isotypic components on the right-hand side of (5.1.1) and in particular are semisimple. We will show that for any simple subquotient $\mathscr{L}\neq \unicode[STIX]{x1D6E5}_{\ast }\operatorname{IC}_{X}$ of the $\operatorname{std}$ -isotypic component we have
from which it follows that $\operatorname{IC}_{X}^{\oplus 2}$ is a direct summand of the $\operatorname{std}$ -isotypic component. Since the other summand has no subquotients supported on the main diagonal, it is semisimple by the induction hypothesis.
We will show that $H^{i}(\unicode[STIX]{x1D6E5}^{!}\mathscr{L})=0$ for $i\leqslant 1$ , which implies that $\operatorname{Ext}^{1}(\unicode[STIX]{x1D6E5}_{\ast }\operatorname{IC}_{X},\mathscr{L})=0$ . The other vanishing follows by applying Verdier duality: since $\operatorname{gr}\unicode[STIX]{x1D6F9}(\,\!\mathscr{W}_{Z^{\unicode[STIX]{x1D707}}})$ is Verdier self-dual, so is its $\operatorname{std}$ -isotypic component. Observe that $\mathscr{L}$ has the form $\text{}^{\prime }\mathfrak{j}_{=\unicode[STIX]{x1D708},!\ast }^{\unicode[STIX]{x1D707}}\text{}^{\prime }\mathfrak{m}_{\unicode[STIX]{x1D708}}^{\unicode[STIX]{x1D707},\unicode[STIX]{x1D6E5}}\mathscr{F}$ for some $0<\unicode[STIX]{x1D708}\leqslant \unicode[STIX]{x1D707}$ and a simple summand $\mathscr{F}$ of $\mathscr{P}^{\unicode[STIX]{x1D708}}$ . The case $\unicode[STIX]{x1D708}=0$ is excluded because then $\mathscr{L}=\operatorname{IC}_{Z^{\unicode[STIX]{x1D707}}}$ has weight $0$ .
First suppose that $\unicode[STIX]{x1D708}<\unicode[STIX]{x1D707}$ . Then we have
Since $\unicode[STIX]{x1D6E5}^{!}\operatorname{IC}_{Z^{\unicode[STIX]{x1D707}-\unicode[STIX]{x1D708}}}$ is concentrated in cohomological degrees greater than or equal to $1$ (see [Reference Braverman and GaitsgoryBG08]) and $\unicode[STIX]{x1D6E5}^{!}\mathscr{F}$ is concentrated in degrees greater than or equal to $0$ , and both complexes have lisse cohomology sheaves, their $!$ -tensor product is concentrated in degrees greater than or equal to $2$ as desired.
Finally, we address the case $\unicode[STIX]{x1D708}=\unicode[STIX]{x1D707}$ , where $\mathscr{F}=\mathscr{L}$ . By Lemma 6.3.1, we can assume $\mathscr{L}$ is a summand of $\operatorname{add}_{\ast }\operatorname{IC}_{\prod X^{(n_{\unicode[STIX]{x1D6FD}})}}$ for some Kostant partition $\unicode[STIX]{x1D707}=\sum n_{\unicode[STIX]{x1D6FD}}\unicode[STIX]{x1D6FD}$ . If $\sum n_{\unicode[STIX]{x1D6FD}}\geqslant 3$ then the claim follows by base change. By assumption $\mathscr{L}\neq \unicode[STIX]{x1D6E5}_{\ast }\operatorname{IC}_{X}$ , so $\sum n_{\unicode[STIX]{x1D6FD}}>1$ . This leaves only the case $\sum n_{\unicode[STIX]{x1D6FD}}=2$ , and since $\unicode[STIX]{x1D707}$ is a coroot the only possibility is that $\unicode[STIX]{x1D707}=\unicode[STIX]{x1D6FD}_{1}+\unicode[STIX]{x1D6FD}_{2}$ is a sum of two distinct coroots. As shown in Lemma 5.1.4, in this case the $\operatorname{std}$ -isotypic component of $\operatorname{gr}\unicode[STIX]{x1D6F9}(\mathscr{W}_{Z^{\unicode[STIX]{x1D707}}})$ is just $\operatorname{IC}_{X}^{\oplus 2}$ .◻
Acknowledgements
I thank my doctoral advisor Dennis Gaitsgory for suggesting that I study this particular nearby cycles sheaf, and for many invaluable conversations full of technical assistance and motivation. The ways in which he has helped and guided me through this project are too numerous to list, but I am especially grateful to him for explaining the statement of Conjecture 1.5.2 and its relevance to the semisimplicity assertion of Theorem 1.4.1. I thank Sam Raskin for many illuminating discussions about this material, which included helping me to formulate Lemma 4.3.1 and explaining Definition 3.1.1. I am grateful to Simon Schieder for explaining to me the definition of the Picard–Lefschetz oscillators for an arbitrary reductive group. I also thank the referee for comments which substantially improved the paper.