Affine Toric Equivalence Relations are Effective Claudiu Raicu - - PowerPoint PPT Presentation

affine toric equivalence relations are effective
SMART_READER_LITE
LIVE PREVIEW

Affine Toric Equivalence Relations are Effective Claudiu Raicu - - PowerPoint PPT Presentation

Affine Toric Equivalence Relations are Effective Claudiu Raicu University of California, Berkeley AMS-SMM Joint Meeting, Berkeley, June 2010 Motivating Question Under what circumstances do quotients by finite equivalence relations exist?


slide-1
SLIDE 1

Affine Toric Equivalence Relations are Effective

Claudiu Raicu

University of California, Berkeley

AMS-SMM Joint Meeting, Berkeley, June 2010

slide-2
SLIDE 2

Motivating Question

Under what circumstances do quotients by finite equivalence relations exist? Outline of talk:

1

Equivalence Relations

2

The Amitsur Complex

3

A Noneffective Equivalence Relation

4

Questions

slide-3
SLIDE 3

Definition of Equivalence Relations

Given a scheme X over a base S, a scheme theoretic equivalence relation on X over S is an S-scheme R together with a morphism f : R → X ×S X

  • ver S such that for any S-scheme T, the set map

f(T) : R(T) → X(T) × X(T) is injective and its image is the graph of an equivalence relation on X(T) (here Z(T) denotes the set of S-maps from T to Z).

slide-4
SLIDE 4

Definition of Equivalence Relations

Given a scheme X over a base S, a scheme theoretic equivalence relation on X over S is an S-scheme R together with a morphism f : R → X ×S X

  • ver S such that for any S-scheme T, the set map

f(T) : R(T) → X(T) × X(T) is injective and its image is the graph of an equivalence relation on X(T) (here Z(T) denotes the set of S-maps from T to Z). R is said to be finite if the two projections R ⇒ X are finite.

slide-5
SLIDE 5

Definition of Equivalence Relations

Given a scheme X over a base S, a scheme theoretic equivalence relation on X over S is an S-scheme R together with a morphism f : R → X ×S X

  • ver S such that for any S-scheme T, the set map

f(T) : R(T) → X(T) × X(T) is injective and its image is the graph of an equivalence relation on X(T) (here Z(T) denotes the set of S-maps from T to Z). R is said to be finite if the two projections R ⇒ X are finite. A coequalizer of this two projections is called the quotient of X by the equivalence relation R.

slide-6
SLIDE 6

The Affine Case

If k is a field and X = An

k is the n-dimensional affine space over k, then

OX ≃ k[x], where x = (x1, · · · , xn). An equivalence relation R ⊂ X ×k X corresponds to an ideal I(x, y) ⊂ k[x, y]

slide-7
SLIDE 7

The Affine Case

If k is a field and X = An

k is the n-dimensional affine space over k, then

OX ≃ k[x], where x = (x1, · · · , xn). An equivalence relation R ⊂ X ×k X corresponds to an ideal I(x, y) ⊂ k[x, y] satisfying:

1

(reflexivity) I(x, y) ⊂ (x1 − y1, · · · , xn − yn)

slide-8
SLIDE 8

The Affine Case

If k is a field and X = An

k is the n-dimensional affine space over k, then

OX ≃ k[x], where x = (x1, · · · , xn). An equivalence relation R ⊂ X ×k X corresponds to an ideal I(x, y) ⊂ k[x, y] satisfying:

1

(reflexivity) I(x, y) ⊂ (x1 − y1, · · · , xn − yn)

2

(symmetry) I(x, y) = I(y, x)

slide-9
SLIDE 9

The Affine Case

If k is a field and X = An

k is the n-dimensional affine space over k, then

OX ≃ k[x], where x = (x1, · · · , xn). An equivalence relation R ⊂ X ×k X corresponds to an ideal I(x, y) ⊂ k[x, y] satisfying:

1

(reflexivity) I(x, y) ⊂ (x1 − y1, · · · , xn − yn)

2

(symmetry) I(x, y) = I(y, x)

3

(transitivity) I(x, z) ⊂ I(x, y) + I(y, z)

slide-10
SLIDE 10

The Affine Case

If k is a field and X = An

k is the n-dimensional affine space over k, then

OX ≃ k[x], where x = (x1, · · · , xn). An equivalence relation R ⊂ X ×k X corresponds to an ideal I(x, y) ⊂ k[x, y] satisfying:

1

(reflexivity) I(x, y) ⊂ (x1 − y1, · · · , xn − yn)

2

(symmetry) I(x, y) = I(y, x)

3

(transitivity) I(x, z) ⊂ I(x, y) + I(y, z) R is finite if and only if I satisfies

4

(finiteness) k[x, y]/I(x, y) is finite over k[x]

slide-11
SLIDE 11

Effective Equivalence Relations

Definition

An equivalence relation R on X is said to be effective if there exists a morphism X → Y such that R ≃ X ×Y X. In the affine case effectivity corresponds to the ideal I(x, y) of the equivalence relation being generated by differences f(x) − f(y).

slide-12
SLIDE 12

Effective Equivalence Relations

Definition

An equivalence relation R on X is said to be effective if there exists a morphism X → Y such that R ≃ X ×Y X. In the affine case effectivity corresponds to the ideal I(x, y) of the equivalence relation being generated by differences f(x) − f(y).

Question (Koll´ ar)

Is every finite equivalence relation effective?

slide-13
SLIDE 13

Effective Equivalence Relations

Definition

An equivalence relation R on X is said to be effective if there exists a morphism X → Y such that R ≃ X ×Y X. In the affine case effectivity corresponds to the ideal I(x, y) of the equivalence relation being generated by differences f(x) − f(y).

Question (Koll´ ar)

Is every finite equivalence relation effective? Answer: No.

slide-14
SLIDE 14

Effective Equivalence Relations

Definition

An equivalence relation R on X is said to be effective if there exists a morphism X → Y such that R ≃ X ×Y X. In the affine case effectivity corresponds to the ideal I(x, y) of the equivalence relation being generated by differences f(x) − f(y).

Question (Koll´ ar)

Is every finite equivalence relation effective? Answer: No. Example: to come.

slide-15
SLIDE 15

Effective Equivalence Relations

Definition

An equivalence relation R on X is said to be effective if there exists a morphism X → Y such that R ≃ X ×Y X. In the affine case effectivity corresponds to the ideal I(x, y) of the equivalence relation being generated by differences f(x) − f(y).

Question (Koll´ ar)

Is every finite equivalence relation effective? Answer: No. Example: to come. Also, Hironaka’s.

slide-16
SLIDE 16

Effective Equivalence Relations

Definition

An equivalence relation R on X is said to be effective if there exists a morphism X → Y such that R ≃ X ×Y X. In the affine case effectivity corresponds to the ideal I(x, y) of the equivalence relation being generated by differences f(x) − f(y).

Question (Koll´ ar)

Is every finite equivalence relation effective? Answer: No. Example: to come. Also, Hironaka’s. “Theorem” If X, Y and f : X → Y are “nice”, and if it happens that the effective equivalence relation R = X ×Y X defined by f is finite, then the quotient X/R exists.

slide-17
SLIDE 17

Toric Equivalence Relations

If X is a (not necessarily normal) toric variety, an equivalence relation R on X is said to be toric if it is invariant under the diagonal action of the torus.

slide-18
SLIDE 18

Toric Equivalence Relations

If X is a (not necessarily normal) toric variety, an equivalence relation R on X is said to be toric if it is invariant under the diagonal action of the torus. In the affine case, this suffices to insure effectivity:

Theorem (–, 2009)

Let k be a field, X/k an affine toric variety, and R a toric equivalence relation on X. Then there exists an affine toric variety Y together with a toric map X → Y such that R ≃ X ×Y X.

slide-19
SLIDE 19

Toric Equivalence Relations

If X is a (not necessarily normal) toric variety, an equivalence relation R on X is said to be toric if it is invariant under the diagonal action of the torus. In the affine case, this suffices to insure effectivity:

Theorem (–, 2009)

Let k be a field, X/k an affine toric variety, and R a toric equivalence relation on X. Then there exists an affine toric variety Y together with a toric map X → Y such that R ≃ X ×Y X. Remarks: The theorem holds without any finiteness assumptions.

slide-20
SLIDE 20

Toric Equivalence Relations

If X is a (not necessarily normal) toric variety, an equivalence relation R on X is said to be toric if it is invariant under the diagonal action of the torus. In the affine case, this suffices to insure effectivity:

Theorem (–, 2009)

Let k be a field, X/k an affine toric variety, and R a toric equivalence relation on X. Then there exists an affine toric variety Y together with a toric map X → Y such that R ≃ X ×Y X. Remarks: The theorem holds without any finiteness assumptions. If R is finite, the quotient exists and is also an affine toric variety.

slide-21
SLIDE 21

Toric Equivalence Relations

If X is a (not necessarily normal) toric variety, an equivalence relation R on X is said to be toric if it is invariant under the diagonal action of the torus. In the affine case, this suffices to insure effectivity:

Theorem (–, 2009)

Let k be a field, X/k an affine toric variety, and R a toric equivalence relation on X. Then there exists an affine toric variety Y together with a toric map X → Y such that R ≃ X ×Y X. Remarks: The theorem holds without any finiteness assumptions. If R is finite, the quotient exists and is also an affine toric variety. The theorem is false in the nonaffine case: an equivalence relation on X = P2 identifying the points of a (torus-invariant) line L can’t be effective; if it were, then the map X → Y defining it would have to contract L and therefore be constant.

slide-22
SLIDE 22

Definition of the Amitsur Complex

Given a commutative ring A and an A-algebra B, we consider the Amitsur complex C(A, B) : B → B ⊗A B → · · · → B⊗Am → · · · with differentials given by the formula d(b1 ⊗ b2 ⊗ · · · ⊗ bm) =

m+1

  • i=1

(−1)ib1 ⊗ · · · ⊗ bi−1 ⊗ 1 ⊗ bi ⊗ · · · ⊗ bm.

slide-23
SLIDE 23

Definition of the Amitsur Complex

Given a commutative ring A and an A-algebra B, we consider the Amitsur complex C(A, B) : B → B ⊗A B → · · · → B⊗Am → · · · with differentials given by the formula d(b1 ⊗ b2 ⊗ · · · ⊗ bm) =

m+1

  • i=1

(−1)ib1 ⊗ · · · ⊗ bi−1 ⊗ 1 ⊗ bi ⊗ · · · ⊗ bm. It is well known that if B is a faithfully flat or augmented A-algebra, then C(A, B) is exact. In these cases, the kernel of the first differential is A.

slide-24
SLIDE 24

Exactness of the Amitsur Complex

It turns out that exactness holds also when A, B are monoid rings and the map A → B is defined on the monoid level:

Theorem (–, 2009)

Let k be any commutative ring, let τ and σ be commutative monoids, and let ϕ : τ → σ be a map of monoids. If A = k[τ], B = k[σ], and B is considered as an A-algebra via the map A → B induced by ϕ, then the Amitsur complex C(A, B) is exact.

slide-25
SLIDE 25

Exactness of the Amitsur Complex

It turns out that exactness holds also when A, B are monoid rings and the map A → B is defined on the monoid level:

Theorem (–, 2009)

Let k be any commutative ring, let τ and σ be commutative monoids, and let ϕ : τ → σ be a map of monoids. If A = k[τ], B = k[σ], and B is considered as an A-algebra via the map A → B induced by ϕ, then the Amitsur complex C(A, B) is exact. As opposed to the faithfully flat and augmented cases, the kernel of the first differential d : B → B ⊗A B, b → b ⊗ 1 − 1 ⊗ b is usually larger than A.

slide-26
SLIDE 26

A 1–Dimensional Zig–zag

If we consider A = k[t3, t5] ⊂ B = k[t] then t7 ∈ B is not an element of A, but it goes to zero under the first differential in the Amitsur complex.

slide-27
SLIDE 27

A 1–Dimensional Zig–zag

If we consider A = k[t3, t5] ⊂ B = k[t] then t7 ∈ B is not an element of A, but it goes to zero under the first differential in the Amitsur complex. t7 ⊗ 1

slide-28
SLIDE 28

A 1–Dimensional Zig–zag

If we consider A = k[t3, t5] ⊂ B = k[t] then t7 ∈ B is not an element of A, but it goes to zero under the first differential in the Amitsur complex. t7 ⊗ 1 = t2 ⊗ t5

slide-29
SLIDE 29

A 1–Dimensional Zig–zag

If we consider A = k[t3, t5] ⊂ B = k[t] then t7 ∈ B is not an element of A, but it goes to zero under the first differential in the Amitsur complex. t7 ⊗ 1 = t2 ⊗ t5 = t2 · t3 ⊗ t2

slide-30
SLIDE 30

A 1–Dimensional Zig–zag

If we consider A = k[t3, t5] ⊂ B = k[t] then t7 ∈ B is not an element of A, but it goes to zero under the first differential in the Amitsur complex. t7 ⊗ 1 = t2 ⊗ t5 = t2 · t3 ⊗ t2 = t5 ⊗ t2

slide-31
SLIDE 31

A 1–Dimensional Zig–zag

If we consider A = k[t3, t5] ⊂ B = k[t] then t7 ∈ B is not an element of A, but it goes to zero under the first differential in the Amitsur complex. t7 ⊗ 1 = t2 ⊗ t5 = t2 · t3 ⊗ t2 = t5 ⊗ t2 = 1 ⊗ t7

slide-32
SLIDE 32

A 1–Dimensional Zig–zag

If we consider A = k[t3, t5] ⊂ B = k[t] then t7 ∈ B is not an element of A, but it goes to zero under the first differential in the Amitsur complex. t7 ⊗ 1 = t2 ⊗ t5 = t2 · t3 ⊗ t2 = t5 ⊗ t2 = 1 ⊗ t7

slide-33
SLIDE 33

A 2–Dimensional Zig–zag

slide-34
SLIDE 34

A Noneffective Affine Equivalence Relation

If k is any ring, A = k[f1, · · · , fm] ⊂ B = k[x], and f(x, y) is a 1–cocycle in the Amitsur complex C(A, B), i.e. f(y, z) − f(x, z) + f(x, y) = 0 ∈ k[x, y, z]/(fi(x) − fi(y), fi(x) − fi(z)), then the ideal I(x, y) = (f(x, y), fi(x) − fi(y) : i = 1, · · · , m) ⊂ k[x, y] defines an equivalence relation on Spec(B). When the fi’s are homogeneous, noneffectivity of this equivalence relation amounts to f not being a coboundary.

slide-35
SLIDE 35

A Noneffective Affine Equivalence Relation

If k is any ring, A = k[f1, · · · , fm] ⊂ B = k[x], and f(x, y) is a 1–cocycle in the Amitsur complex C(A, B), i.e. f(y, z) − f(x, z) + f(x, y) = 0 ∈ k[x, y, z]/(fi(x) − fi(y), fi(x) − fi(z)), then the ideal I(x, y) = (f(x, y), fi(x) − fi(y) : i = 1, · · · , m) ⊂ k[x, y] defines an equivalence relation on Spec(B). When the fi’s are homogeneous, noneffectivity of this equivalence relation amounts to f not being a coboundary.

Example

f1(x) = x2

1, f2(x) = x1x2 − x2 2, f3(x) = x3 2,

f(x, y) = (x1y2 − x2y1)y3

2.

slide-36
SLIDE 36

Questions

Do quotients by finite equivalence relations exist in characteristic 0?

slide-37
SLIDE 37

Questions

Do quotients by finite equivalence relations exist in characteristic 0? Given a finite equivalence relation on an affine variety, is there a method of producing invariant sections?

slide-38
SLIDE 38

Questions

Do quotients by finite equivalence relations exist in characteristic 0? Given a finite equivalence relation on an affine variety, is there a method of producing invariant sections? Are finite toric equivalence relations effective?

slide-39
SLIDE 39

Questions

Do quotients by finite equivalence relations exist in characteristic 0? Given a finite equivalence relation on an affine variety, is there a method of producing invariant sections? Are finite toric equivalence relations effective? Is there a geometric way of explaining the noneffective equivalence relations coming from the nonvanishing of the first cohomology of the Amitsur complex?