Notes on Beilinson s How to glue perverse sheaves
|
|
- Kristin Pearson
- 5 years ago
- Views:
Transcription
1 Notes on Beilinson s How to glue perverse sheaves Ryan Reich June 4, 2009 In this paper I provide something o a skeleton key to A.A. Beilinson s How to glue perverse sheaves [1], which I ound hard to understand as well as lacking a discussion o some o its more interesting consequences. In order to maintain readability, I will work in the shea-theoretic context o the classical topology on complex algebraic varieties; or the necessary modiications to étale sheaves, one should consult Beilinson s paper: aside rom the shit in deinitions the only change is some Tate twists. For the D-modules case, read Sam Lichtenstein s undergraduate thesis, [2]. 1 Theoretical preliminaries The topic at hand is perverse sheaves and nearby cycles; I give a summary o the deinitions and necessary properties here. 1.0 Diagram chases Especially in 3, we will need to do some pretty serious diagram chasing. Some people make a big deal about how you can t chase diagrams in a general category because the objects aren t sets, but in act in an abelian category (or even a Quillen-exact category, i one wishes to adapt the results in 3 to the more general context in which Beilinson stated them in [1]) chasing elements is perectly possible. The important act is that an object in an abelian category A can be considered to be a shea on the canonical topology o A. This is, by deinition, the largest Grothendieck topology in which all representable unctors Hom A (, x) are sheaves, and its open covers are precisely the strict universal epimorphisms. Such a map is, in a more general category, a map : u x such that the ibered product u = u x u exists, the coequalizer x = coker(u u) exists, the natural map x x is an isomorphism, and that all o this is also true when we make any base change along a map g : y x, or the induced map x id: u x y y. In an abelian category, however, this is all equivalent merely to the statement that is a surjection. Recall the deinitions o the various constructions on sheaves: 1. Kernels o maps are taken sectionwise; i.e. or a map : F G, ker()(u) = ker((u): F(U) G(U)). Likewise, products and limits are taken sectionwise. 2. Cokernels are locally taken sectionwise: any section s coker()(u) is, on some open cover V o U, o the orm t or t G(V ). Likewise, images, coproducts, and colimits are taken locally. In an abelian category, where all o these constructions exist by assumption, these descriptions are even prescriptive: i one orms the sheaves thus described, they are representable by the objects claimed. Thereore, the ollowing common arguments in diagram chasing are valid: 1. A map : x y is surjective i and only i or every s y, there is some t x such that s = (t). This is code or: or every open set U and every s y(u), there is a surjection V U and a section t y(v ) such that s V = (t). 2. I s y, then s = 0 coker() i and only i s im(). This is code or: i s y(u) and s = 0 coker()(u), then there is some surjection V U and t x(v ) with s V = (t). 1
2 3. is zero i and only i or all s x, (x) = 0. This is actually the deinition o the zero map. Any other arguments involving elements and some concept related to exactness can also be phrased in this way. Thus, a naïve diagram-chasing argument can be converted into a rigorous one simply by replacing statements like s F with correct ones s F(U) or some open set U, and passing to surjective covers when necessary. The process being mechanical, I will just give the naive proo. 1.1 Derived category and unctors All the action takes place in the derived category; speciically, let X be an algebraic variety and denote by D(X) its derived category o bounded complexes o sheaves with constructible cohomology. By deinition, a map o complexes : K L in D(X) is an isomorphism i its associated map on cohomology sheaves H i (): H i (K ) H i (L ) is or all i. We have a notation or the index-shit: K i+1 = (K[1]) i (technically, the dierential maps also change sign, but we will never need to think about this). The derived category D(X) is a triangulated category, which means merely that in it are a class o triples o complexes and maps A B C A [1] in which two consecutive arrows compose to zero, satisying some axioms we won t need, and with the property that the associated long sequence o cohomology sheaves... H 1 (C ) H 0 (A ) H 0 (B ) H 0 (C ) H 1 (A )... is exact (note that H 0 (A [1]) = H 1 (A )); we say that the H i are cohomological. I : A B is given, there always exists a (unique up to non-unique isomorphism) triangle whose third term C = Cone() is the cone o. A unctor between two triangulated categories is triangulated i it sends triangles in one to triangles in the other. In D(X) we also have some standard constructions o shea theory. For any two complexes there is the total tensor product A B obtained by taking in degree n the direct sum o all products A i B j with i + j = n (and some dierentials that are irrelevant) and its derived unctor A L B, with H i (A L B ) = Tor i (A, B ), which is a triangulated unctor. We also have the unctor Hom(A, B ), where Hom(A, B ) i (U) = Hom(A U, B [i] U ), and its derived unctor R Hom(A, B ), with H i R Hom(A, B ) = Ext i (A, B ), which is triangulated. O course, these two have an adjunction: R Hom(A L B, C ) = R Hom(A, R Hom(B, C )). For any Zariski-open subset U X with inclusion map j, there are triangulated unctors j!, j : D(U) D(X) and j = j! : D(X) D(U); i i is the inclusion o its complement Z, then there are likewise maps i!, i : D(X) D(Z) and i = i! : D(Z) D(X). (Again, technically j! and j are derived unctors and should be denoted Lj! and Rj, but we will never have occasion to use the plain versions so we elide this extra notation.) They satisy a number o important relations, o which I will only use one here: there is a natural triangle or any complex A j! (A ) j (A ) i i j (A ) There is also a triangulated duality unctor D: D(X) D(X) op which interchanges! and, in that Dj (A ) = j! (DA ), etc., and is an involution In act, i we set D = DC, then D(A ) = R Hom(A, D ). The only thing we need to know about D is that it exchanges! and ; i.e. that Dj! F = j DF, etc. For any map : X Y o varieties, we have!, as well (also!,, and none o them are equal), with the same relationships to D, and the useul identity! R Hom(F, G) = R Hom( F,! G). 2
3 1.2 Perverse sheaves Within D(X) there is an abelian category M(X) o perverse sheaves which has nicer properties than the category o actual sheaves. It is speciied by means o a t-structure, namely, a pair o ull subcategories p D(X) 0 and p D(X) 0, also satisying some conditions I won t use, and such that M(X) = p D(X) 0 p D(X) 0. There are truncation unctors τ 0 : D(X) p D(X) 0 and likewise or τ 0, itting into a distinguished triangle or any complex K : τ 0 K K τ >0 K (where τ >0 = τ 1 = [ 1] τ 0 [1]). This triangle is unique with respect to the property that the irst term is in p D(X) 0 and the third is in p D(X) >0. They have the obvious properties implied by the notation: τ a τ b = τ a i a b, and likewise or τ?. Furthermore, there are perverse cohomology unctors p H i : D(X) M(X), where o course p H i (A ) = p H 0 (A [i]) and p H 0 = τ 0 τ 0 = τ 0 τ 0 ; these are cohomological just like the ordinary cohomology unctors. The abelian category structure o M(X) is more or less determined by the act that i we have a map : F G o perverse sheaves (this is the notation I will be using; we will not think o perverse sheaves as complexes), then ker = p H 1 Cone() coker = p H 0 Cone(). I X is connected o dimension D, then the category M(X) is closed under a translate o the duality unctor, D[d], but not necessarily under the six unctors deined or an open/closed pair o subvarieties. However, it is true that j (F), i! (F) p D 0 and j! (F), i (F) p D 0, while j (F), i (F) M; we say these unctors are right, let, or just t-exact. Furthermore, when j is an aine morphism (the primary example being when Z is a Cartier divisor), both j! and j are t-exact, and thus their restriction to M(U) is exact with values in M(X). There is also a middle extension unctor j!, deined so that j! (F) is the image o p H 0 (j! F) in p H 0 (j F) along the natural map j! j ; it is the unique perverse shea such that i j! F p D <0 and i! j! F p D >0, but or me the most useul property is that when j is an aine, open immersion, then we have a sequence o perverse sheaves i j! F[ 1] j! F j! F j F i! j! F[1]; i.e. i j! F[ 1] = ker(j! F j F) and i! j! F[1] = coker(j! F j F) are both perverse sheaves. Perverse sheaves have good category-theoretic properties: M(X) is both artinian and noetherian, so every perverse shea has inite length. Finally, we will use the shea-theoretic act that i L is a locally constant shea on X, then F L is perverse whenever F is. Note that since L is locally ree, it is lat, and thereore F L = F L L. 1.3 Nearby cycles I we have a map : X A 1 such that Z = 1 (0), the nearby cycles unctor RΨ : D(U) D(Z) is deined. Namely, let u: G m G m be the universal cover o G m = A 1 \ {0}, let v : Ũ = U G m Gm U be its pullback, orming a diagram Z i X j U {0} A 1 G m v u Ũ G m and set RΨ (A ) = i Rj Rv v A : D(U) D(Z). 3
4 Since i and v are exact, indeed we have RΨ = R(i j v v ) is the right derived unctor o a let-exact unctor Ψ rom sheaves on U to sheaves on Z. The undamental group π 1 (G m ) acts on any j F via deck transormations o G m and thereore acts on Ψ and RΨ. There is a natural map, obtained rom (v, v )-adjunction, rom i F Ψ (j F), on whose image π 1 (G m ) acts trivially. We set, by deinition, i F Ψ (j F) Φ (F) 0 where Φ (F) is the vanishing cycles shea, and or the generator t π 1 (G m ) we get a map Var(t): Φ (F) Ψ (F) actoring 1 t on Ψ (F) through the quotient. Using some homological algebra tricks the above sequence induces a natural distinguished triangle i A RΨ (j A ) RΦ (A ) where RΦ is (morally) the right derived unctor o Φ. I beore taking the derived unctor we single out the subshea Ψ un (F) on which 1 t is locally nilpotent (in act, by the constructibility theorem o SGA7 it is actually nilpotent), we get the unctor o unipotent nearby cycles. The unit o (v, v )-adjunction i F Ψ (j F) actors through Ψ un (F) since π 1(G m ) acts trivially on i F, and the sequence i F Ψ un (j F) 1 t Ψ un (j F) induces a distinguished triangle (replacing F with the complex j A ) i j A RΨ un (A ) 1 t RΨ un (A ) (1) I L is any locally constant shea on G m with underlying vector space L, then RΨ un (A L) = RΨ un (A ) L, where t acts on the tensor product by acting on each actor (this ollows since L is trivialized on G m ). 2 The unipotent nearby cycles unctor We start o with a nice result about nearby cycles. Theorem 2.1. The unctor RΨ un [ 1] sends p D(U) 0 to p D(Z) 0 and takes M(U) to M(Z). Proo. Since j is aine and an open immersion, j and j! are t-exact, so or any A p D(U) 0, i j K = i Cone(j! A j A ) is in p D(Z) 0. I we apply the long exact sequence o perverse cohomology to equation (1), we thereore get H 3 (RΨ un A ) 1 t H 3 (RΨ un A ) H 2 (i j A ) H 2 (RΨ un A ) 1 t H 2 (RΨ un A ) H 1 (i j A ) H 1 (RΨ un A ) 1 t H 1 (RΨ un A ) H 0 (i j A ) 0 H 0 (RΨ un A ) 1 t H 0 (RΨ un A ) 0 For i 1, the map H i (RΨ un M) Hi (RΨ un M) is both given by a nilpotent operator and is injective, thereore zero. For i = 0 it is nilpotent and surjective, so also zero. It ollows that RΨ un (A ) p D(Z) <0, as promised. Now let M M(U) be a perverse shea. Then i j M p D(Z) [ 1,0] since it consists o the kernel and cokernel o the map j! M j M. This means that or i 2 in the above sequence, all the maps 1 t are injective and nilpotent, hence zero. Thus RΨ un (M) p D(Z) 1, as desired. We proceed to give a construction o RΨ un. Let La be the vector space o dimension a 0 together with the action o a matrix J a = [δ ij + δ i,j 1 ], a unipotent Jordan block o dimension n. Let L a be the locally constant shea on G m with underlying space L a and monodromy action where a generator t o π 1 (G m ) acts by J a ; we consider it a complex in cohomological degree zero. Lemma 2.2. Let A D(U); then D(A L a ) = D(A ) L a. 4
5 Proo. First we prove a related act: i we set Ľa = Hom(L a, C), then there is an isomorphism Ľa = L a. Indeed, L a has a given basis e 1,..., e a associated to J, and we ix the basis e a,..., e 1 o dual vectors or Ľa, sending J to J t, and thus inducing the desired map o local systems. In general, then, we compare: D(F) L a = R Hom(F, D ) L L a D(F L a ) = R Hom(F L L a, D ) = R Hom(F, DL a ). There is a natural map rom the ormer to the latter induced by ( L, R Hom)-adjunction: R Hom(F, D ) R Hom( L a, R Hom(F, D L a )) = R Hom(F, R Hom( L a, D L a )). Note that D =! D (one on X, the other on G m ) by deinition and D L a =! DL a. By the property o!, we have R Hom( L a,! DL a ) =! R Hom(L a, DL a ) and thereore the desired map is the one induced on R Hom(F, ) rom the one obtained by applying! to a certain map D R Hom(L a, DL a ) = D(L a L L a ). This map, in turn, is obtained by irst replacing the L with (since L a is locally ree), replacing one copy o L a with its dual, and applying D to the trace map L a Ľa C. The map D(F) L a D(F L a ) is an isomorphism locally since it would be an isomorphism i L a were replaced with a constant shea. Thus, it is an isomorphism. For each a, b 0 there is a natural exact sequence 0 L a ga,b L a+b ga+b, b L b 0; (2) that is, or any r Z, g a,r sends L a to the irst r coordinates o L a+r i r 0, and to the quotient L a ( r) given by collapsing the irst r coordinates i r 0 (that is, r 0). This sequence respects the action o J? on the terms and under the identiication o Lemma 2.2, the (a, b) sequence is dual to the (b, a) sequence; that is, Dg a,r = g a+r, r. These have various ormal identities: 1. When r 0, we have g a+r, a g a,r = 0; 2. When r 0 and a + r 0, we have g a+r, r g a,r = (1 t) r ; 3. When r and s have the same sign, and a + r 0, we have g a,r+s = g a+r,s g a,r. Let M M(U), so M L a is also perverse and has an action o the monodromy via J a acting on the second actor, which we will call simply t. The maps id g a,r between these will be called g a,r M, or just ga,r again i there is no need to mention M, and we have Dg a,r M = ga+r, r DM. Note that since the L a are locally ree, the g a,r M are all injective when r 0 and surjective when r 0. We can prove a strong uniormity lemma based on the monodromy action: Lemma 2.3. Let α a : j! (M L a ) j (M L a ). annihilates both ker α a and coker α a or all a. There exists an integer N such that (1 t) N 5
6 Proo. We consider again the triangle (1); i we could ind such an N annihilating Ψ un (M L a ), then it would annihilate i j (M L a ) as well, and hence its perverse cohomologies ker α a and coker α a. We know that Ψ un (M L a ) = Ψ un (M) La and that 1 t is nilpotent on Ψ un (M), say (1 t)n = 0, so it suices to show that this power kills the tensor product as well. Indeed, the short exact sequences (2) with (a, b) = (1, a 1): 0 L 1 L a L a 1 0 induce short exact sequences o perverse sheaves respecting the monodromy action: 0 Ψ un Thus the claim is proved by induction. (M) Ψ un (M) L a Ψ un (M) L a 1 0. Corollary 2.4. There exists a map β a : j! (M L a ) j (M L a ) such that α a β a = β a α a = (1 t) 2N. Proo. Consider the diagram j! (M L a ) j (M L α! a ) β coim α a = im α a (1 t) N β! α j! (M L a ) j (M L a ) (1 t) N where α α! = α a, and set β a = β! β. Note that β! and β exist and β! α! = α β = (1 t) N because (1 t) N annihilates ker α a and coker α a, respectively. We claim that α! β! = β α = (1 t) N as well. Indeed, consider the latter and apply α to both sides: α β α = (1 t) N α = α (1 t) N, since α a, and hence α and α!, respect the monodromy action. Since α is injective, β α = (1 t) N, as desired. Likewise or α! β!, applying it to α! and using surjectivity. Thereore, α a β a = α α! β! β = α (1 t) N β = (1 t) N α β = (1 t) 2N. Likwise or β a α a. Let r Z with a + r 0, and set α a,r = j (g a,r ) α a = α a+r j! (g a,r ), so α a,r : j! (M L a ) j (M L a+r ). Setting β a,r = j! (g a+r, r ) β a+r = β a g (g a+r, r ), we can generalize Corollary 2.4: Corollary 2.5. We have: 1. When r 0, α a,r β a,r = (1 t) 2N+r ; 2. When r 0, β a,r α a,r = (1 t) 2N+r ; 3. For any r, β a+2r, r α a,r = (1 t) 2(N+r) j! (g a,2r ) and α a+r,r β a+r, r = (1 t) 2(N+r) j (g a,2r ). Proposition 2.6. For a 0, the natural maps j! (g a,1 ) and j (g a+r,1 ) respectively induce isomorphisms ker(α a,r ) ker(α a+1,r ) and coker(α a,r ) coker(α a+1,r ). When r 0, the ormer are isomorphisms or all a, and when r 0, the latter are. These statements also hold with α replaced by β and r negated. Proo. The latter statement is obvious: when r 0, g a,r and hence j (g a,r ) is injective, so α a,r has the same kernel as α a ; likewise or r 0 and j! (g a,r ) being surjective. 6
7 Suppose r 0; then, applying Corollary 2.5, we have β a,r α a,r = (1 t) 2N+r, so (1 t) 2N+r annihilates ker(α a,r ). Likewise, when r 0, α a,r β a,r = (1 t) 2N+r, so (1 t) 2N+r annihilates coker(α a,r ). Using the natural maps j! (g a,1 ) and j (g a+r,1 ) we get a commutative square j! (M L a ) α a,r j (M L a+r ) j! (g a,1 ) j! (M L a+1 ) j (g a+r,1 ) α a+1,r j (M L a+r+1 ) showing that j! (g a,1 ) induces a map on kernels and j (g a+r,1 ) on cokernels. Since the j! (g a,1 ) are injective, we get a long sequence o inclusions o kernels: ker α a 1,r ker α a,r ker α a+1,r. Each kernel is annihilated by (1 t) 2N+r, whose kernel is (or a 2N +r) the perverse shea j! (M L 2N+r ). Since perverse sheaves have inite length, this chain must have a maximum, whence the claim or kernels. For cokernels, they are all quotients j (M L 2N+r ), and stabilized since their kernels are a similar increasing subsequence. To obtain the results or β, just swap it with α in all the arguments and negate r; we have only used Corollary 2.5, which is symmetric under this transormation. Departing slightly rom Beilinson s notation, we denote these stable kernels ker α,r or r 0 and coker α,r or r 0; when r = 0 we drop it. Lemma 2.7. Suppose that in Proposition 2.6 the stable cokernels coker(α a,r ), coker(β a,r ) obtain or a L. Then in this range the ollowing natural maps are isomorphisms: im(α a+1,r )/ im(α a,r ) j (M L a+1+r )/j (M L a+r ) im(β a+1,r )/ im(β a,r ) j! (M L a+1 )/j! (M L a ). In particular, (α a,r ) 1 (im(1 t) n ) im(1 t) n and the same or β, or n a L. Proo. This is o course merely a general algebraic act: suppose we have objects A 1 A 2 in an abelian category and subobjects B i A i with B 1 B 2. Then the ollowing three conditions are equivalent: 1. The inclusion B 2 A 2 induces an inclusion B 2 /B 1 A 2 /A 1 ; 2. The inclusion A 1 A 2 induces an inclusion A 1 /B 1 A 2 /B 2 ; 3. The natural map B 1 B 2 A1 A 2 is an isomorphism. Furthermore, B 2 /B 1 A 2 /A 1 i and only i A 1 /B 1 A 2 /B 2. For the proo, assume by symmetry that the irst map is injective (resp. surjective), and chase the ollowing diagram: B 1 A 1 A 1 /B B 2 A 2 A 2 /B 2 0 B 1 /B 1 A 2 /A
8 Proposition 2.8. For r 0, there is an isomorphism coker α,r = ker α, r. Proo. Starting with the map β a o Corollary 2.4, and or a M, let γ a,r = (1 t) a 2N β a+r, so γ a : j (M L a+r ) j! (M L a+r ). By deinition o α a,r, we have β a α a,r = (1 t) 2N j! (g a,r ), so γ a,r α a,r = j! (g a,r ) (1 t) a = 0 since (1 t) a annihilates L a. Likewise, α a+r, r β a = (1 t) 2N j (g a+r, r ), so α a+r, r γ a,r = 0 or the same reason. So, in act γ a,r : coker α a,r ker α a+r, r. The kernel o γ a,r is by deinition (β a+r ) 1 im(1 t) 2N+r mod im α a,r, but since α a,r β a,r = (1 t) 2N+r, to show that γ a,r is injective it suices to show that (β a+r ) 1 im(1 t) 2N+r im(1 t) 2N+r, which ollows rom Lemma 2.7 when a 0. For surjectivity, using the act that (1 t) 2N r annihilates ker(α a+r, r ) (so ker(α a+r, r ) im(1 t) a 2N ), a quick diagram chase shows that it suices that (α a+r, r ) 1 im(1 t) 2N im(β a+r ), or since β a+r α a+r = (1 t) 2N, that (α a,r ) 1 im(1 t) 2N im(1 t) 2N, which is again true or a 0 by Lemma 2.7. Because they are equal, we will give a single name to the stable kernel and cokernel: Π r (M) = ker(α, r ) = coker(α,r ). The deinitions o α a and g a,r easily imply that DΠ r (M) = Π r (DM). Theorem 2.9. There is an isomorphism RΨ un (M)[ 1] = Π 0 (M), so that RΨun commutes with D. Proo. This is really just a more speciic version o the argument in Lemma 2.3. Consider again the triangle (1) with A = M L a. Since we know that RΨ un (M)[ 1] is a perverse shea and that i j (M L a ) p D [ 1,0] (X), it suices to show that ker(1 t) = RΨ un (M)[ 1]. Since RΨ un (M L a )[ 1] = RΨ un (M)[ 1] L a = a i=1 RΨ un (M)[ 1] (a), where the i th coordinate o t is t (i) + t (i+1). Thus, i c i is the i th coordinate (unction) o ker(1 t), we have c i+1 = (1 t 1 (i) )c i, or by induction, c i = (1 t 1 ) i 1 c 1 and (1 t 1 ) a c 1 = 0. Since 1 t is nilpotent on RΨ un (M), or a suiciently large, this ormula or c i in act deines ker(1 t) to be the isomorphic image o RΨ un (M)[ 1], as desired. Note that by rearranging some deinitions (and doing the completely analogous proo or the cokernel), we could have given Theorem 2.9 right ater Lemma 2.2 and proven the r = 0 cases o Proposition 2.6 and Proposition 2.8 as a side eect. However, we need the r = 1 case or the gluing constructions. Note also the corollary: Corollary Suppose (1 t) N annihilates RΨ un (M). Then we have (and conversely, o course) RΨ un (M)[ 1] = i j! (M L N )[ 1] = i! j! (M L N )[1]. Nearby cycles and the perverse t-structure One thing Beilinson does not at all address in his paper is the interaction o RΨ un with the entire perverse t-structure. However, this an easy consequence o two general lemmas. To this end, let T be a triangulated unctor on a t-category C and let the core o C be the abelian category A. We will assume that the objects o C are bounded above, meaning that C = b Z C b, and that C is nondegenerate, meaning that b Z C b = 0. These assumptions hold or the perverse t-structure. Lemma Suppose T is right t-exact and that T A A; then T is t-exact. 8
9 Proo. We will show that T commutes with all truncations. Suppose we have an object x C b, so that there is a distinguished triangle τ <b x x τ b x where by deinition, τ b x = H b (x)[ b] A[ b]. By hypothesis on T, we have T (x) C b, T (τ <b x) C <b, and T (H b x[ b]) A[ b] C b. Since T is triangulated, there is a triangle T (τ <b x) T (x) T (H b x[ b]) and thereore, by uniqueness o the truncation triangle, it must be that T (τ <b x) = τ <b T (x). This is under the hypothesis that x C b ; since then τ <b x C b 1 and since τ <b 1 τ <b = τ <b 1, we can apply truncationsby-one repeatedly and conclude that or all n, we have τ n T (x) = T (τ n x). Now suppose we have any x, and or any n orm the distinguished triangle τ <n x x τ n to which we apply T. Since T (τ <n x) = τ <n T (x), the cone o the resulting triangle τ <n T (x) T (x) T (τ n x) must be isomorphic to τ n T (x), by uniqueness o cones and the truncation triangle or T (x). τ n T (x) = T (τ n x). Since then T commutes with all trunctions, it is a ortiori t-exact. Thus, Lemma Suppose that T is t-exact and that T vanishes on A; then T = 0. Proo. T commutes with cohomology, so or any x we have H n (T x) = T H n (x) = 0. Thus, by nondegeneracy, we must have T (x) = 0. In the irst lemma, take T = i RΨ un [ 1]; by Theorem 2.1, it satisies the hypothesis o Lemma 2.11, and thereore is t-exact. By deinition, the Verdier duality unctor D is (contravariantly) t-exact or the perverse t-structure, so taking T = D i RΨ un [ 1] i RΨ un [ 1] D in Lemma 2.12 we get T = 0 on all o D(X) by Theorem 2.9. That is: Theorem The unipotent nearby cycles RΨ un 3 Vanishing cycles and gluing commute with Verdier duality. Since Π 0 = RΨ un [ 1], we will reer to it simply as Ψ (M), and Π 1 as Ξ, what Beilinson calls the maximal extension unctor. Proposition 3.1. There are two natural exact sequences exchanged by duality and M DM: where α + α = α and β β + = 1 t. 0 j! (M) α Ξ (M) β Ψ (M) 0 0 Ψ (M) β+ Ξ (M) α+ j (M) 0, Proo. Since α a,r = α a+r j! (g a,r ) we have im(α a,r ) im(α a+r ) and thus we obtain a natural surjection coker(α,r ) coker(α ). The kernel o this map is im(α a,r )/ im(α a ) or a 0; this receives a natural surjection, via α a+r, rom coker(1 t) r inside j! (M L a+r ), which is injective by Lemma 2.7. In particular, or r = 1 we get the irst short exact sequence. Likewise, since α a+r, r = j (g a+r, r ) α a+r, we have ker(α ) ker(α, r ), where the quotient is ker(α a+r, r )/ ker(α a+r ) or a 0; it maps, via α a+r, injectively to ker(g a+r, r ) = im(1 t) a, and surjectively by Lemma 2.7. For r = 1 we get the second short exact sequence. 9
10 The duality claim ollows rom the act that D(α a,r M ) = αa+r, r DM and that D is contravariant and exact. Since j Ψ = 0 and j (α a ) = id, we conclude that j (α + α ) = id as well; by adjunction, α + α = α. Observe that the constructions above take place in j!, (M L a+1 ). Recall the isomorphism γ a,r = (1 t) a 2N β a+r o Proposition 2.8; what we must really prove is that β (γ a,1 ) 1 β + = (1 t)(γ a+1,0 ) 1. First we show that β + γ a+1,0 β = (1 t)γ a,1, which is immediately obvious rom the deinition o γ a,r (note that the symbolic ormula given above is not the isomorphism o coker(α a,r ) with ker(α a+r, r ), but that this isomorphism is canonically induced by it. This induction is precisely the contribution o the β ± ). Thereore, multiplying on the let by β (γ a,1 ) 1, we get: β (γ a,1 ) 1 β + γ a+1,0 β = β (1 t) = (1 t) β. Since β is surjective, we may remove it rom the right, and we are done ater multiplying by (γ a+1,0 ) 1. The remainder o the paper is simply what Beilinson calls linear algebra. Let M = j F or a perverse shea F M(X). From the maps in these two exact sequences we can orm a complex: j! j (F) (α,γ ) Ξ j (F) F (α+, γ+) j j (F), (3) where γ : j! j (F) F and γ + : M j j (F) are deined by the let- and right-adjunctions (j!, j ) and (j, j ) and the property that j (γ ) = j (γ + ) = id. Proposition 3.2. The complex (3) is in act a complex; let Φ (F) be its cohomology shea. Then Φ is an exact unctor whose values are supported on Z and there are maps u, v such that v u = 1 t as in the ollowing diagram: Ψ j (F) u Φ (F) v Ψ j (F). Proo. That (3) is a complex amounts to showing that γ + γ = α = α + α, which is true by deinition o the γ ± and adjunction. To show that it is exact, suppose we have 0 F 1 F 2 F 3 0, so that we get a short exact sequence o complexes 0 C (F 1 ) C (F 2 ) C (F 3 ) 0, where by C (F) we have denoted the complex (3) padded with zeroes on both sides. Note that since α is injective and α + surjective, C (F) ails to be exact only at the middle term. Thereore we have a long exact sequence o cohomology sheaves: (0 = H 1 C (F 3 )) Φ (F 1 ) Φ (F 2 ) Φ (F 3 ) (0 = H 1 (C (F 1 )))... which shows that Φ is unctorial and an exact unctor. I we apply j to (3), it becomes simply j F (id,id) j F j F (id, id) j F which is actually exact, so j Φ (F) = 0; i.e. Φ (F) is supported on Z. Finally, to deine u and v, let pr: Ξ j (F) F Ξ j (F), and set u = (β +, 0) in coordinates, and v = β pr. Then v u = β β + = 1 t by Proposition 3.1. Deine a vanishing cycles gluing data or to be a quadruple (F U, F Z, u, v) modeled on Proposition 3.2; or any F M(X), we have F (F) = (j F, Φ (F), u, v) as there. Let M (U, Z) be the category o such data, and F : M(X) M (U, Y ) is a unctor. Conversely, given a vanishing cycles data, we can orm the complex Ψ (F U ) (β+,u) (β, v) Ξ (F U ) F Z Ψ (F U ) (4) since v u = 1 t = β β +, and let G (F U, F Y, u, v) be its cohomology shea. We could prove Theorem 3.6 directly, but some o the ormalism in Beilinson s paper is too elegant to cut. Thereore, we make two urther deinitions (which are a little dierent rom the terminology he uses): 10
11 Deinition 3.3. Let a diad be a complex o perverse sheaves on X o the orm ( ) D (L) = (a L,b L ) = F L A B (R) = (a R,b R ) F R in which a L is injective and a R is surjective (so it is exact on the ends). Let the category o diads be denoted M 2 (X). Let a triad be a short exact sequence o the orm S = (c,d (0 1 F,d2 ) A B 1 B 2 (c+,d1 +,d2 + ) ) F + 0 in which both (c, d i ): F A B i are injections and (c +, d i +): A B i F + are surjections. Let the category o triads be denoted M 3 (X); it has a natural relection unctor r : M 3 (X) M 3 (X) which invokes the natural symmetry 1 2, and is an involution. We can deine a map T : M 2 (X) M 3 (X) by setting ( ) T (D) = ker(r) (ι A,ι B,h) A B H(D ) (π A,π B, k) coker(l), where: 1. ι = (ι A, ι B ): ker(r) A B is the natural inclusion, 2. π = (π A, π B ): A B coker(l) is the natural projection, 3. h: ker(r) H(D ) is the natural projection onto H(D ) = ker(r)/f L, 4. k : H(D ) coker(l) is the natural inclusion o H(D ) = ker(coker(l) F R ). These our maps are related: we have πι = kh. Deine the inverse T 1 by the ormula T 1 (S) = (ker(d 2 ) (c,d1 ) ) A B 1 coker(c, d 1 ). Lemma 3.4. The unctors T, T 1 are mutually inverse equivalences o M 2 (X) with M 3 (X). Proo. First we have to show that T (D ) M 3 (X) at all. The conditions that (ι A, ι B ) = ι be injective and (π A, π B ) = π be surjective are true by deinition. To show that (ι A, h) is injective we do a diagram chase: i x ker(r) and ι A (x) = h(x) = 0, then x im(l): x = (a L (y), b L (y)) with y F L, and a L (y) = ι A (x) = 0. But a L is injective, so y = 0 and hence x = 0. Likewise, we show that (π A, k) is surjective: suppose that x coker(l), and pick a lit (y 1, y 2 ) A B, with π A (y 1 ) + π B (y 2 ) = x. Since a R is surjective, there is some z A such that a R (z) = b R (y 2 ), and thereore (R)( z, y 2 ) = 0: ( z, y 2 ) ker(r). By deinition, kh( z, y 2 ) = π( z, y 2 ) = π B (y 2 ) π A (z), so x = π A (y 1 + z) kh(z, y 2 ), as desired. Now we have to check exactness in the middle. First o all, T (D ) is a complex, since π ι = k h. We prove exactness: suppose (x, y) ker(π, k), so π(x) = k(y) coker(l), with x A B and y H(D ). Pick z ker(r) such that y = h(z), so π(x) = k(y) = πι(z); thus, x ι(z) ker(π) = im(l) ker(r) and thus x ker(r), or more precisely, there is w ker(r) such that x = ι(w). Also, k(y) = π(x) = πι(w) = kh(w), so since k is injective, y = h(w); thus, (x, y) = (ι, h)(w), as desired. Likewise, we must check that T 1 (S) M 2 (X). That it is a complex is obvious, so we must show that c is injective on ker(d 2 ) and that A alone surjects onto coker(c, d 1 ). The ormer ollows rom the act that (c, d 2 ) is assumed to be injective. For the latter, let (x 1, x 2 ) A B 1 ; since (c +, d 2 +) is surjective, there is some (y 1, y 3 ) A B 2 such that c + (y 1 ) + d 2 +(y 3 ) = d 1 +(x 2 ). That is, (y 1, x 2, y 3 ) ker(c +, d 1 +, d 2 +), and hence in im(c, d 1, d 2 ). In particular, (y 1, x 2 ) im(c, d 1 ), so (x 1, x 2 ) and (x 1 + y 1, 0) have the same image in coker(c, d 1 ), as desired. It is easy to check that T 1 T = id: its F L is ker(h) = im(l) = F L ; its A and B are indeed A and B, and its F R is coker(ι) = F R ; one checks quickly that the maps are right as well. Conversely, T T 1 = id: 11
12 its F is ker(a B 1 coker(c, d 1 )) = F since (c, d 1 ) is an injection; its A and B 1 are obviously the original A and B 1 ; we will deal at greater length with B 2 and F +. For B 2, we must show that F / ker(d 2 ) = B 2, or in other words, that d 2 is surjective. This is another diagram chase: i x B 2, then since (c +, d 1 +) is surjective there is some (y 1, y 2 ) A B 1 such that c + (y 1 ) + d 1 +(y 2 ) = d + 2 (x). Thereore (y 1, y 2, x) ker(c +, d 1 +, d 2 +) = im(c, d 1, d 2 ), so in particular x = d 2 +(z), or x = d 2 +( z), or some z F, as desired. For F +, we must show that the smaller sequence 0 ker(d 2 ) (c,d1 ) A B 1 (c+,d1 + ) F + 0 is still exact. The hypotheses on S already show that it is exact at the ends, so we deal only with the middle. First o all, it is a complex, since c + c + d 1 +d 1 = d 2 +d 2 = 0 on ker(d 2 ). Now we do still another diagram chase: i x ker(c +, d 1 +), then (x, 0) ker(c +, d 1 +, d 2 +), so there is some y F such that (c (y), d 1 (y)) = x and d 2 (y) = 0; i.e. y ker(d 2 ), as desired. Corollary 3.5. The relection unctor on a diad T 1 rt (D ) is the complex r(d ) = ( ) ker(a R ) (a L,b L ) A H(D ) (a R,b R ) coker(a L ), where a L is the natural inclusion and a R the natural projection, b L = h (a L, 0), and b R actorizes k through coker(a L ) coker(l). Proo. The d 2 o rt (D ) is ι B, whose kernel (in ker(r)) we show is ker(a R ). Indeed, the map (id, 0) exhibits ker(a R ) ker(ι B ) ker(r); conversely, i (x 1, x 2 ) ker(ι B ) ker(r), then x 2 = 0 and a R (x 1 ) = 0, so x 1 ker(a R ). Likewise, the d 1 o rt (D ) is h, so we must show that coker(ι A, h) = coker(a L ). The map (id, 0) rom A to A H(D ) induces a map A coker(ι A, h); to show that it descends to coker(a L ), we must show that or x F L, (a L (x), 0) = (ι A (y), h(y)) or some y ker(r). In particular, y im(l) since h(y) = 0, so we can take y = (L)(x) = (a L (x), b L (x)). Thus we have a map coker(a L ) coker(ι A, h), which this argument also shows is injective. It is also surjective: i (x 1, x 2 ) A H(D ), then there is some y ker(r) with h(y) = x 2, so (x 1 ι(y), 0) has the same image in coker(ι A, h). The injectivity argument just given shows that coker(a L ) coker(l), so the description o b R makes sense. Indeed, k does actor through coker(a L ), as shown by the surjectivity argument above, and in act the computation there shows that it coincides with the map H(D ) coker(a L ). Similar considerations prove the identities o the other maps. Theorem 3.6. The gluing category M (U, Z) is abelian, and F : M(X) M (U, Z) and G : M (U, Z) M(X) are mutually inverse exact unctors, so M (U, Z) is equivalent to M(X). Proo. That M (U, Z) is abelian amounts to proving that taking coordinatewise kernels and cokernels works. That is, i we have (F U, F Z, u, v) and (F U, F Z, u, v ) with maps a U : F U F U, a Z : F Z F Z and such that the ollowing diagram commutes: Ψ (F U ) u v F Z Ψ (F U ) Ψ (a U ) Ψ (F U ) u F Z a Z Ψ (a U ) v Ψ (F U ) then (ker a U, ker a Z, ũ, ṽ) is a kernel or (a U, a V ), where ũ and ṽ are induced maps; likewise or the cokernel; and we must show that (a U, a V ) is an isomorphism i and only i the kernel and cokernel vanish. The maps ũ and ṽ are constructed rom the natural sequence o kernels (or cokernels) in the above diagram, and the exactness o Ψ, and once they exist it is obvious rom the deinition o morphisms in M (U, Z) that the 12
13 desired gluing data is a kernel (resp. cokernel). Since M(U) and M(Z) are abelian and kernels and cokernels are taken coordinatewise, the last claim ollows. That F is exact ollows rom the orm taken by kernels and cokernels, and the exactness o Φ ; that G is unctorial and exact is exactly the same proo as that Φ is. All that remains is to prove that they are mutually inverse. We do this by interpreting both M(X) and M (U, Z) as diads in the orm given, respectively, by diagrams (3) and (4). Then Corollary 3.5 shows that these two types are interchanged by the relection unctor, with the correspondence explicitly realizing the two unctors F and G previously given. Reerences [1] A. A. Beilinson, How to glue perverse sheaves, K-theory, arithmetic and geometry (Moscow, 1984), Lecture Notes in Math., vol. 1289, Springer, Berlin, 1987, pp , available at 20glue%20perverse%20sheaves.pd. [2] Sam Lichtenstein, Vanishing cycles or algebraic D-modules, Bachelor s thesis, Harvard University, 2009, harvard.edu/~gaitsgde/grad_2009/lichtenstein(2009).pd. 13
How to glue perverse sheaves
How to glue perverse sheaves A.A. Beilinson The aim o this note [0] is to give a short, sel-contained account o the vanishing cycle constructions o perverse sheaves; e.g., or the needs o [1]. It diers
More informationDescent on the étale site Wouter Zomervrucht, October 14, 2014
Descent on the étale site Wouter Zomervrucht, October 14, 2014 We treat two eatures o the étale site: descent o morphisms and descent o quasi-coherent sheaves. All will also be true on the larger pp and
More informationMath 248B. Base change morphisms
Math 248B. Base change morphisms 1. Motivation A basic operation with shea cohomology is pullback. For a continuous map o topological spaces : X X and an abelian shea F on X with (topological) pullback
More informationSEPARATED AND PROPER MORPHISMS
SEPARATED AND PROPER MORPHISMS BRIAN OSSERMAN The notions o separatedness and properness are the algebraic geometry analogues o the Hausdor condition and compactness in topology. For varieties over the
More information1 Categories, Functors, and Natural Transformations. Discrete categories. A category is discrete when every arrow is an identity.
MacLane: Categories or Working Mathematician 1 Categories, Functors, and Natural Transormations 1.1 Axioms or Categories 1.2 Categories Discrete categories. A category is discrete when every arrow is an
More informationSEPARATED AND PROPER MORPHISMS
SEPARATED AND PROPER MORPHISMS BRIAN OSSERMAN Last quarter, we introduced the closed diagonal condition or a prevariety to be a prevariety, and the universally closed condition or a variety to be complete.
More informationVALUATIVE CRITERIA BRIAN OSSERMAN
VALUATIVE CRITERIA BRIAN OSSERMAN Intuitively, one can think o separatedness as (a relative version o) uniqueness o limits, and properness as (a relative version o) existence o (unique) limits. It is not
More informationMath 216A. A gluing construction of Proj(S)
Math 216A. A gluing construction o Proj(S) 1. Some basic deinitions Let S = n 0 S n be an N-graded ring (we ollows French terminology here, even though outside o France it is commonly accepted that N does
More informationVALUATIVE CRITERIA FOR SEPARATED AND PROPER MORPHISMS
VALUATIVE CRITERIA FOR SEPARATED AND PROPER MORPHISMS BRIAN OSSERMAN Recall that or prevarieties, we had criteria or being a variety or or being complete in terms o existence and uniqueness o limits, where
More information2 Coherent D-Modules. 2.1 Good filtrations
2 Coherent D-Modules As described in the introduction, any system o linear partial dierential equations can be considered as a coherent D-module. In this chapter we ocus our attention on coherent D-modules
More informationMath 754 Chapter III: Fiber bundles. Classifying spaces. Applications
Math 754 Chapter III: Fiber bundles. Classiying spaces. Applications Laurențiu Maxim Department o Mathematics University o Wisconsin maxim@math.wisc.edu April 18, 2018 Contents 1 Fiber bundles 2 2 Principle
More informationUniversity of Cape Town
The copyright o this thesis rests with the. No quotation rom it or inormation derived rom it is to be published without ull acknowledgement o the source. The thesis is to be used or private study or non-commercial
More informationPERVERSE SHEAVES. Contents
PERVERSE SHEAVES SIDDHARTH VENKATESH Abstract. These are notes for a talk given in the MIT Graduate Seminar on D-modules and Perverse Sheaves in Fall 2015. In this talk, I define perverse sheaves on a
More information8 Perverse Sheaves. 8.1 Theory of perverse sheaves
8 Perverse Sheaves In this chapter we will give a self-contained account of the theory of perverse sheaves and intersection cohomology groups assuming the basic notions concerning constructible sheaves
More informationLIMITS AND COLIMITS. m : M X. in a category G of structured sets of some sort call them gadgets the image subset
5 LIMITS ND COLIMITS In this chapter we irst briely discuss some topics namely subobjects and pullbacks relating to the deinitions that we already have. This is partly in order to see how these are used,
More informationHSP SUBCATEGORIES OF EILENBERG-MOORE ALGEBRAS
HSP SUBCATEGORIES OF EILENBERG-MOORE ALGEBRAS MICHAEL BARR Abstract. Given a triple T on a complete category C and a actorization system E /M on the category o algebras, we show there is a 1-1 correspondence
More informationGENERALIZED ABSTRACT NONSENSE: CATEGORY THEORY AND ADJUNCTIONS
GENERALIZED ABSTRACT NONSENSE: CATEGORY THEORY AND ADJUNCTIONS CHRIS HENDERSON Abstract. This paper will move through the basics o category theory, eventually deining natural transormations and adjunctions
More informationTHE SNAIL LEMMA ENRICO M. VITALE
THE SNIL LEMM ENRICO M. VITLE STRCT. The classical snake lemma produces a six terms exact sequence starting rom a commutative square with one o the edge being a regular epimorphism. We establish a new
More informationCategories and Natural Transformations
Categories and Natural Transormations Ethan Jerzak 17 August 2007 1 Introduction The motivation or studying Category Theory is to ormalise the underlying similarities between a broad range o mathematical
More informationIn the index (pages ), reduce all page numbers by 2.
Errata or Nilpotence and periodicity in stable homotopy theory (Annals O Mathematics Study No. 28, Princeton University Press, 992) by Douglas C. Ravenel, July 2, 997, edition. Most o these were ound by
More informationVariations on a Casselman-Osborne theme
Variations on a Casselman-Osborne theme Dragan Miličić Introduction This paper is inspired by two classical results in homological algebra o modules over an enveloping algebra lemmas o Casselman-Osborne
More informationA NOTE ON SHEAVES WITHOUT SELF-EXTENSIONS ON THE PROJECTIVE n-space.
A NOTE ON SHEAVES WITHOUT SELF-EXTENSIONS ON THE PROJECTIVE n-space. DIETER HAPPEL AND DAN ZACHARIA Abstract. Let P n be the projective n space over the complex numbers. In this note we show that an indecomposable
More informationDERIVED CATEGORIES OF STACKS. Contents 1. Introduction 1 2. Conventions, notation, and abuse of language The lisse-étale and the flat-fppf sites
DERIVED CATEGORIES OF STACKS Contents 1. Introduction 1 2. Conventions, notation, and abuse of language 1 3. The lisse-étale and the flat-fppf sites 1 4. Derived categories of quasi-coherent modules 5
More informationA TALE OF TWO FUNCTORS. Marc Culler. 1. Hom and Tensor
A TALE OF TWO FUNCTORS Marc Culler 1. Hom and Tensor It was the best of times, it was the worst of times, it was the age of covariance, it was the age of contravariance, it was the epoch of homology, it
More informationCHOW S LEMMA. Matthew Emerton
CHOW LEMMA Matthew Emerton The aim o this note is to prove the ollowing orm o Chow s Lemma: uppose that : is a separated inite type morphism o Noetherian schemes. Then (or some suiciently large n) there
More informationMATH 101B: ALGEBRA II PART A: HOMOLOGICAL ALGEBRA 23
MATH 101B: ALGEBRA II PART A: HOMOLOGICAL ALGEBRA 23 6.4. Homotopy uniqueness of projective resolutions. Here I proved that the projective resolution of any R-module (or any object of an abelian category
More informationMADE-TO-ORDER WEAK FACTORIZATION SYSTEMS
MADE-TO-ORDER WEAK FACTORIZATION SYSTEMS EMILY RIEHL The aim o this note is to briely summarize techniques or building weak actorization systems whose right class is characterized by a particular liting
More informationTCC Homological Algebra: Assignment #3 (Solutions)
TCC Homological Algebra: Assignment #3 (Solutions) David Loeffler, d.a.loeffler@warwick.ac.uk 30th November 2016 This is the third of 4 problem sheets. Solutions should be submitted to me (via any appropriate
More informationTHE GORENSTEIN DEFECT CATEGORY
THE GORENSTEIN DEFECT CATEGORY PETTER ANDREAS BERGH, DAVID A. JORGENSEN & STEFFEN OPPERMANN Dedicated to Ranar-Ola Buchweitz on the occasion o his sixtieth birthday Abstract. We consider the homotopy cateory
More informationTHE HOMOTOPY THEORY OF EQUIVALENCE RELATIONS
THE HOMOTOPY THEORY OF EQUIVALENCE RELATIONS FINNUR LÁRUSSON Abstract. We give a detailed exposition o the homotopy theory o equivalence relations, perhaps the simplest nontrivial example o a model structure.
More informationHomotopy, Quasi-Isomorphism, and Coinvariants
LECTURE 10 Homotopy, Quasi-Isomorphism, an Coinvariants Please note that proos o many o the claims in this lecture are let to Problem Set 5. Recall that a sequence o abelian groups with ierential is a
More informationThe Clifford algebra and the Chevalley map - a computational approach (detailed version 1 ) Darij Grinberg Version 0.6 (3 June 2016). Not proofread!
The Cliord algebra and the Chevalley map - a computational approach detailed version 1 Darij Grinberg Version 0.6 3 June 2016. Not prooread! 1. Introduction: the Cliord algebra The theory o the Cliord
More informationAlgebraic Geometry Spring 2009
MIT OpenCourseWare http://ocw.mit.edu 18.726 Algebraic Geometry Spring 2009 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms. 18.726: Algebraic Geometry
More informationEXT, TOR AND THE UCT
EXT, TOR AND THE UCT CHRIS KOTTKE Contents 1. Left/right exact functors 1 2. Projective resolutions 2 3. Two useful lemmas 3 4. Ext 6 5. Ext as a covariant derived functor 8 6. Universal Coefficient Theorem
More informationAFFINE PUSHFORWARD AND SMOOTH PULLBACK FOR PERVERSE SHEAVES
AFFINE PUSHFORWARD AND SMOOTH PULLBACK FOR PERVERSE SHEAVES YEHAO ZHOU Conventions In this lecture note, a variety means a separated algebraic variety over complex numbers, and sheaves are C-linear. 1.
More informationPERVERSE SHEAVES: PART I
PERVERSE SHEAVES: PART I Let X be an algebraic variety (not necessarily smooth). Let D b (X) be the bounded derived category of Mod(C X ), the category of left C X -Modules, which is in turn a full subcategory
More informationCATEGORIES. 1.1 Introduction
1 CATEGORIES 1.1 Introduction What is category theory? As a irst approximation, one could say that category theory is the mathematical study o (abstract) algebras o unctions. Just as group theory is the
More informationCLASS NOTES MATH 527 (SPRING 2011) WEEK 6
CLASS NOTES MATH 527 (SPRING 2011) WEEK 6 BERTRAND GUILLOU 1. Mon, Feb. 21 Note that since we have C() = X A C (A) and the inclusion A C (A) at time 0 is a coibration, it ollows that the pushout map i
More informationMATH 101B: ALGEBRA II PART A: HOMOLOGICAL ALGEBRA
MATH 101B: ALGEBRA II PART A: HOMOLOGICAL ALGEBRA These are notes for our first unit on the algebraic side of homological algebra. While this is the last topic (Chap XX) in the book, it makes sense to
More informationRepresentation Theory of Hopf Algebroids. Atsushi Yamaguchi
Representation Theory o H Algebroids Atsushi Yamaguchi Contents o this slide 1. Internal categories and H algebroids (7p) 2. Fibered category o modules (6p) 3. Representations o H algebroids (7p) 4. Restrictions
More informationTRIANGULATED CATEGORIES, SUMMER SEMESTER 2012
TRIANGULATED CATEGORIES, SUMMER SEMESTER 2012 P. SOSNA Contents 1. Triangulated categories and functors 2 2. A first example: The homotopy category 8 3. Localization and the derived category 12 4. Derived
More informationCOHEN-MACAULAY RINGS SELECTED EXERCISES. 1. Problem 1.1.9
COHEN-MACAULAY RINGS SELECTED EXERCISES KELLER VANDEBOGERT 1. Problem 1.1.9 Proceed by induction, and suppose x R is a U and N-regular element for the base case. Suppose now that xm = 0 for some m M. We
More informationDe Rham Cohomology. Smooth singular cochains. (Hatcher, 2.1)
II. De Rham Cohomology There is an obvious similarity between the condition d o q 1 d q = 0 for the differentials in a singular chain complex and the condition d[q] o d[q 1] = 0 which is satisfied by the
More informationEILENBERG-ZILBER VIA ACYCLIC MODELS, AND PRODUCTS IN HOMOLOGY AND COHOMOLOGY
EILENBERG-ZILBER VIA ACYCLIC MODELS, AND PRODUCTS IN HOMOLOGY AND COHOMOLOGY CHRIS KOTTKE 1. The Eilenberg-Zilber Theorem 1.1. Tensor products of chain complexes. Let C and D be chain complexes. We define
More informationDerived Algebraic Geometry I: Stable -Categories
Derived Algebraic Geometry I: Stable -Categories October 8, 2009 Contents 1 Introduction 2 2 Stable -Categories 3 3 The Homotopy Category of a Stable -Category 6 4 Properties of Stable -Categories 12 5
More informationAN INTRODUCTION TO PERVERSE SHEAVES. Antoine Chambert-Loir
AN INTRODUCTION TO PERVERSE SHEAVES Antoine Chambert-Loir Antoine Chambert-Loir Université Paris-Diderot. E-mail : Antoine.Chambert-Loir@math.univ-paris-diderot.fr Version of February 15, 2018, 10h48 The
More informationAlgebraic Geometry Spring 2009
MIT OpenCourseWare http://ocw.mit.edu 18.726 Algebraic Geometry Spring 2009 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms. 18.726: Algebraic Geometry
More information2. ETA EVALUATIONS USING WEBER FUNCTIONS. Introduction
. ETA EVALUATIONS USING WEBER FUNCTIONS Introduction So ar we have seen some o the methods or providing eta evaluations that appear in the literature and we have seen some o the interesting properties
More informationGENERAL ABSTRACT NONSENSE
GENERAL ABSTRACT NONSENSE MARCELLO DELGADO Abstract. In this paper, we seek to understand limits, a uniying notion that brings together the ideas o pullbacks, products, and equalizers. To do this, we will
More informationIntroduction to modules
Chapter 3 Introduction to modules 3.1 Modules, submodules and homomorphisms The problem of classifying all rings is much too general to ever hope for an answer. But one of the most important tools available
More informationMath 210B. Artin Rees and completions
Math 210B. Artin Rees and completions 1. Definitions and an example Let A be a ring, I an ideal, and M an A-module. In class we defined the I-adic completion of M to be M = lim M/I n M. We will soon show
More informationGood tilting modules and recollements of derived module categories, II.
Good tilting modules and recollements of derived module categories, II. Hongxing Chen and Changchang Xi Abstract Homological tilting modules of finite projective dimension are investigated. They generalize
More informationREPRESENTATION THEORY WEEK 9
REPRESENTATION THEORY WEEK 9 1. Jordan-Hölder theorem and indecomposable modules Let M be a module satisfying ascending and descending chain conditions (ACC and DCC). In other words every increasing sequence
More informationGrothendieck construction for bicategories
Grothendieck construction or bicategories Igor Baković Rudjer Bošković Institute Abstract In this article, we give the generalization o the Grothendieck construction or pseudo unctors given in [5], which
More informationThe Universal Coefficient Theorem
The Universal Coefficient Theorem Renzo s math 571 The Universal Coefficient Theorem relates homology and cohomology. It describes the k-th cohomology group with coefficients in a(n abelian) group G in
More information1 Replete topoi. X = Shv proét (X) X is locally weakly contractible (next lecture) X is replete. D(X ) is left complete. K D(X ) we have R lim
Reference: [BS] Bhatt, Scholze, The pro-étale topology for schemes In this lecture we consider replete topoi This is a nice class of topoi that include the pro-étale topos, and whose derived categories
More information(C) The rationals and the reals as linearly ordered sets. Contents. 1 The characterizing results
(C) The rationals and the reals as linearly ordered sets We know that both Q and R are something special. When we think about about either o these we usually view it as a ield, or at least some kind o
More informationEquivalence Relations
Equivalence Relations Definition 1. Let X be a non-empty set. A subset E X X is called an equivalence relation on X if it satisfies the following three properties: 1. Reflexive: For all x X, (x, x) E.
More informationDUALITY AND SMALL FUNCTORS
DUALITY AND SMALL FUNCTORS GEORG BIEDERMANN AND BORIS CHORNY Abstract. The homotopy theory o small unctors is a useul tool or studying various questions in homotopy theory. In this paper, we develop the
More informationFOUNDATIONS OF ALGEBRAIC GEOMETRY CLASS 37
FOUNDATIONS OF ALGEBRAIC GEOMETRY CLASS 37 RAVI VAKIL CONTENTS 1. Motivation and game plan 1 2. The affine case: three definitions 2 Welcome back to the third quarter! The theme for this quarter, insofar
More informationFILTERED RINGS AND MODULES. GRADINGS AND COMPLETIONS.
FILTERED RINGS AND MODULES. GRADINGS AND COMPLETIONS. Let A be a ring, for simplicity assumed commutative. A filtering, or filtration, of an A module M means a descending sequence of submodules M = M 0
More informationA NOTE ON HENSEL S LEMMA IN SEVERAL VARIABLES
PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 125, Number 11, November 1997, Pages 3185 3189 S 0002-9939(97)04112-9 A NOTE ON HENSEL S LEMMA IN SEVERAL VARIABLES BENJI FISHER (Communicated by
More informationSpan, Cospan, and Other Double Categories
ariv:1201.3789v1 [math.ct] 18 Jan 2012 Span, Cospan, and Other Double Categories Susan Nieield July 19, 2018 Abstract Given a double category D such that D 0 has pushouts, we characterize oplax/lax adjunctions
More informationThe V -filtration and vanishing and nearby cycles
The V -filtration and vanishing and nearby cycles Gus Lonergan Disclaimer: We will work with right D-modules. Any D-module we consider will be at least coherent. We will work locally, choosing etale coordinates
More informationNotes on the definitions of group cohomology and homology.
Notes on the definitions of group cohomology and homology. Kevin Buzzard February 9, 2012 VERY sloppy notes on homology and cohomology. Needs work in several places. Last updated 3/12/07. 1 Derived functors.
More informationCategories and functors
Lecture 1 Categories and functors Definition 1.1 A category A consists of a collection ob(a) (whose elements are called the objects of A) for each A, B ob(a), a collection A(A, B) (whose elements are called
More informationTangent Categories. David M. Roberts, Urs Schreiber and Todd Trimble. September 5, 2007
Tangent Categories David M Roberts, Urs Schreiber and Todd Trimble September 5, 2007 Abstract For any n-category C we consider the sub-n-category T C C 2 o squares in C with pinned let boundary This resolves
More informationThe Hecke category (part I factorizable structure)
The Hecke category (part I factorizable structure) Ryan Reich 16 February 2010 In this lecture and the next, we will describe the Hecke category, namely, the thing which acts on D-modules on Bun G and
More informationPerverse Sheaves. Bhargav Bhatt. Fall The goal of this class is to introduce perverse sheaves, and how to work with it; plus some applications.
Perverse Sheaves Bhargav Bhatt Fall 2015 1 September 8, 2015 The goal o this class is to introduce perverse sheaves, and how to work with it; plus some applications. Background For more background, see
More informationDeformation theory of representable morphisms of algebraic stacks
Deformation theory of representable morphisms of algebraic stacks Martin C. Olsson School of Mathematics, Institute for Advanced Study, 1 Einstein Drive, Princeton, NJ 08540, molsson@math.ias.edu Received:
More informationFOUNDATIONS OF ALGEBRAIC GEOMETRY CLASS 25
FOUNDATIONS OF ALGEBRAIC GEOMETRY CLASS 25 RAVI VAKIL CONTENTS 1. Quasicoherent sheaves 1 2. Quasicoherent sheaves form an abelian category 5 We began by recalling the distinguished affine base. Definition.
More informationDerived categories, perverse sheaves and intermediate extension functor
Derived categories, perverse sheaves and intermediate extension functor Riccardo Grandi July 26, 2013 Contents 1 Derived categories 1 2 The category of sheaves 5 3 t-structures 7 4 Perverse sheaves 8 1
More informationUMS 7/2/14. Nawaz John Sultani. July 12, Abstract
UMS 7/2/14 Nawaz John Sultani July 12, 2014 Notes or July, 2 2014 UMS lecture Abstract 1 Quick Review o Universals Deinition 1.1. I S : D C is a unctor and c an object o C, a universal arrow rom c to S
More informationLecture 9 - Faithfully Flat Descent
Lecture 9 - Faithfully Flat Descent October 15, 2014 1 Descent of morphisms In this lecture we study the concept of faithfully flat descent, which is the notion that to obtain an object on a scheme X,
More informationThe basics of frame theory
First version released on 30 June 2006 This version released on 30 June 2006 The basics o rame theory Harold Simmons The University o Manchester hsimmons@ manchester.ac.uk This is the irst part o a series
More informationExercises on chapter 0
Exercises on chapter 0 1. A partially ordered set (poset) is a set X together with a relation such that (a) x x for all x X; (b) x y and y x implies that x = y for all x, y X; (c) x y and y z implies that
More informationChern classes à la Grothendieck
Chern classes à la Grothendieck Theo Raedschelders October 16, 2014 Abstract In this note we introduce Chern classes based on Grothendieck s 1958 paper [4]. His approach is completely formal and he deduces
More informationare additive in each variable. Explicitly, the condition on composition means that given a diagram
1. Abelian categories Most of homological algebra can be carried out in the setting of abelian categories, a class of categories which includes on the one hand all categories of modules and on the other
More informationAlgebraic Geometry Spring 2009
MIT OpenCourseWare http://ocw.mit.edu 18.726 Algebraic Geometry Spring 2009 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms. 18.726: Algebraic Geometry
More informationarxiv:math/ v1 [math.at] 6 Oct 2004
arxiv:math/0410162v1 [math.at] 6 Oct 2004 EQUIVARIANT UNIVERSAL COEFFICIENT AND KÜNNETH SPECTRAL SEQUENCES L. GAUNCE LEWIS, JR. AND MICHAEL A. MANDELL Abstract. We construct hyper-homology spectral sequences
More informationFORMAL GLUEING OF MODULE CATEGORIES
FORMAL GLUEING OF MODULE CATEGORIES BHARGAV BHATT Fix a noetherian scheme X, and a closed subscheme Z with complement U. Our goal is to explain a result of Artin that describes how coherent sheaves on
More information1. THE CONSTRUCTIBLE DERIVED CATEGORY
1. THE ONSTRUTIBLE DERIVED ATEGORY DONU ARAPURA Given a family of varieties, we want to be able to describe the cohomology in a suitably flexible way. We describe with the basic homological framework.
More informationD-Modules and Mixed Hodge Modules
D-Modules and Mixed Hodge Modules Notes by Takumi Murayama Fall 2016 and Winter 2017 Contents 1 September 19 (Harold Blum) 4 1.1 Deinitions [HTT08, 1.1]...................................... 4 1.2 The
More informationwhere Σ is a finite discrete Gal(K sep /K)-set unramified along U and F s is a finite Gal(k(s) sep /k(s))-subset
Classification of quasi-finite étale separated schemes As we saw in lecture, Zariski s Main Theorem provides a very visual picture of quasi-finite étale separated schemes X over a henselian local ring
More informationA Primer on Homological Algebra
A Primer on Homological Algebra Henry Y Chan July 12, 213 1 Modules For people who have taken the algebra sequence, you can pretty much skip the first section Before telling you what a module is, you probably
More informationPART II.1. IND-COHERENT SHEAVES ON SCHEMES
PART II.1. IND-COHERENT SHEAVES ON SCHEMES Contents Introduction 1 1. Ind-coherent sheaves on a scheme 2 1.1. Definition of the category 2 1.2. t-structure 3 2. The direct image functor 4 2.1. Direct image
More informationHOMOLOGY AND COHOMOLOGY. 1. Introduction
HOMOLOGY AND COHOMOLOGY ELLEARD FELIX WEBSTER HEFFERN 1. Introduction We have been introduced to the idea of homology, which derives from a chain complex of singular or simplicial chain groups together
More informationThe Diamond Category of a Locally Discrete Ordered Set.
The Diamond Category of a Locally Discrete Ordered Set Claus Michael Ringel Let k be a field Let I be a ordered set (what we call an ordered set is sometimes also said to be a totally ordered set or a
More informationCategory Theory. Course by Dr. Arthur Hughes, Typset by Cathal Ormond
Category Theory Course by Dr. Arthur Hughes, 2010 Typset by Cathal Ormond Contents 1 Types, Composition and Identities 3 1.1 Programs..................................... 3 1.2 Functional Laws.................................
More informationALGEBRAIC GEOMETRY COURSE NOTES, LECTURE 9: SCHEMES AND THEIR MODULES.
ALGEBRAIC GEOMETRY COURSE NOTES, LECTURE 9: SCHEMES AND THEIR MODULES. ANDREW SALCH 1. Affine schemes. About notation: I am in the habit of writing f (U) instead of f 1 (U) for the preimage of a subset
More informationAn introduction to derived and triangulated categories. Jon Woolf
An introduction to derived and triangulated categories Jon Woolf PSSL, Glasgow, 6 7th May 2006 Abelian categories and complexes Derived categories and functors arise because 1. we want to work with complexes
More informationFourier Mukai transforms II Orlov s criterion
Fourier Mukai transforms II Orlov s criterion Gregor Bruns 07.01.2015 1 Orlov s criterion In this note we re going to rely heavily on the projection formula, discussed earlier in Rostislav s talk) and
More informationUNIVERSAL DERIVED EQUIVALENCES OF POSETS
UNIVERSAL DERIVED EQUIVALENCES OF POSETS SEFI LADKANI Abstract. By using only combinatorial data on two posets X and Y, we construct a set of so-called formulas. A formula produces simultaneously, for
More informationJournal of Pure and Applied Algebra
Journal o Pure and Applied Algebra () 9 9 Contents lists available at ScienceDirect Journal o Pure and Applied Algebra journal homepage: www.elsevier.com/locate/jpaa When the heart o a aithul torsion pair
More informationTheta divisors and the Frobenius morphism
Theta divisors and the Frobenius morphism David A. Madore Abstract We introduce theta divisors for vector bundles and relate them to the ordinariness of curves in characteristic p > 0. We prove, following
More informationFinite Dimensional Hilbert Spaces are Complete for Dagger Compact Closed Categories (Extended Abstract)
Electronic Notes in Theoretical Computer Science 270 (1) (2011) 113 119 www.elsevier.com/locate/entcs Finite Dimensional Hilbert Spaces are Complete or Dagger Compact Closed Categories (Extended bstract)
More informationCALABI-YAU ALGEBRAS AND PERVERSE MORITA EQUIVALENCES
CALABI-YAU ALGEBRAS AND PERERSE MORITA EQUIALENCES JOSEPH CHUANG AND RAPHAËL ROUQUIER Preliminary Draft Contents 1. Notations 2 2. Tilting 2 2.1. t-structures and filtered categories 2 2.1.1. t-structures
More information3. The Sheaf of Regular Functions
24 Andreas Gathmann 3. The Sheaf of Regular Functions After having defined affine varieties, our next goal must be to say what kind of maps between them we want to consider as morphisms, i. e. as nice
More information1 Relative degree and local normal forms
THE ZERO DYNAMICS OF A NONLINEAR SYSTEM 1 Relative degree and local normal orms The purpose o this Section is to show how single-input single-output nonlinear systems can be locally given, by means o a
More informationProperties of Triangular Matrix and Gorenstein Differential Graded Algebras
Properties of Triangular Matrix and Gorenstein Differential Graded Algebras Daniel Maycock Thesis submitted for the degree of Doctor of Philosophy chool of Mathematics & tatistics Newcastle University
More information