Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Getting a fieldMain Theorem
Applications
.
.
Zariski Geometries, Lecture 3
Masanori Itai
Dept of Math Sci, Tokai University, Japan
August 31, 2011 at Kobe university
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Table of Contents
.
. .
1 Getting a field
.
. .
2 Main Theorem
.
. .
3 Applications
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Assumption
There is an irreducible Zariski curve J with an abeliangroup strucure on it.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
.
Lemma (Lemma 4.2.1)
.
.
.
There is a one-dimensional, smooth, faithful family N ofcurves on J2 through (0, 0).
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
.
Corollary (Cor 4.2.5)
.
.
.
There are a one-dimensinal irreducible manifol U andconstructible irreducible ternary relations P,S ⊂ U2 whichdetermine a partial Z-field structure on U
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
.
Lemma (Lem 4.2.6)
.
.
.
There is a non-commutative meta abelian Z-group structuteT(U) on U × U.For some dense open U′ ⊆ U, there is a Z-embedding of thepre-group T(U′) into a connected Z-group G such that
dim G = 2
G is a solvable group with finite center.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
.
Theorem
.
.
.
There is a Z-field in M .
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (1)
Start with the Z-group G with the finite center C(G).With some arguments, we may assume that G is centerless.
Consider the commutater subgroup [G,G] of G, and1 , a ∈ G.
Claim: dim aG = 1.
This is because aG is a normal subgroup of [G,G] and0 < dim aG < dim G.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (2)
If b is another non-unit element of [G,G], we have
aG = bG or aG ∩ bG = ∅.
We have a partition of [G,G]/{1} into one-dimensionalorbits.
By dimension argumet, we see that there are only finitelymany one-dimensional orbits.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (3)
[G,G] is a constructible group of dimension 1.
[G,G] contains a connected normal subgroup [G,G]◦ ofdimension 1.
Thus [G,G]◦ = aG ∪ {1} for some non-unit a ∈ G.Fix a for the rest of the proof.
Put K+ := [G,G]◦, and write the group operation of K+
additively.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (3)
[G,G] is a constructible group of dimension 1.
[G,G] contains a connected normal subgroup [G,G]◦ ofdimension 1.
Thus [G,G]◦ = aG ∪ {1} for some non-unit a ∈ G.Fix a for the rest of the proof.
Put K+ := [G,G]◦, and write the group operation of K+
additively.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (3)
[G,G] is a constructible group of dimension 1.
[G,G] contains a connected normal subgroup [G,G]◦ ofdimension 1.
Thus [G,G]◦ = aG ∪ {1} for some non-unit a ∈ G.Fix a for the rest of the proof.
Put K+ := [G,G]◦, and write the group operation of K+
additively.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (3)
[G,G] is a constructible group of dimension 1.
[G,G] contains a connected normal subgroup [G,G]◦ ofdimension 1.
Thus [G,G]◦ = aG ∪ {1} for some non-unit a ∈ G.Fix a for the rest of the proof.
Put K+ := [G,G]◦, and write the group operation of K+
additively.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (4)
The group G acts on K+ by conjugation; for g ∈ G andx ∈ K+
gx := g−1xg
x 7→ gx is an automorphism of the Z-group K+.
Recall we have fixed non-unit a ∈ G. C(a) is a normalsubgroup of G.
Notice that g1 and g2 induces the same action if
g−11
g2 ∈ C(a)
.
Denote K× the quotient group G/C(a).
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (4)
The group G acts on K+ by conjugation; for g ∈ G andx ∈ K+
gx := g−1xg
x 7→ gx is an automorphism of the Z-group K+.
Recall we have fixed non-unit a ∈ G. C(a) is a normalsubgroup of G.
Notice that g1 and g2 induces the same action if
g−11
g2 ∈ C(a)
.
Denote K× the quotient group G/C(a).
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (4)
The group G acts on K+ by conjugation; for g ∈ G andx ∈ K+
gx := g−1xg
x 7→ gx is an automorphism of the Z-group K+.
Recall we have fixed non-unit a ∈ G. C(a) is a normalsubgroup of G.
Notice that g1 and g2 induces the same action if
g−11
g2 ∈ C(a)
.
Denote K× the quotient group G/C(a).
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (4)
The group G acts on K+ by conjugation; for g ∈ G andx ∈ K+
gx := g−1xg
x 7→ gx is an automorphism of the Z-group K+.
Recall we have fixed non-unit a ∈ G. C(a) is a normalsubgroup of G.
Notice that g1 and g2 induces the same action if
g−11
g2 ∈ C(a)
.
Denote K× the quotient group G/C(a).
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (5)
Denote K× the quotient group G/C(a).
K× acts transitively on K+ − {0}.K× is a connected one-dimensional group, hencecommutative.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (5)
Denote K× the quotient group G/C(a).
K× acts transitively on K+ − {0}.
K× is a connected one-dimensional group, hencecommutative.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (5)
Denote K× the quotient group G/C(a).
K× acts transitively on K+ − {0}.K× is a connected one-dimensional group, hencecommutative.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (6)
For g, g1, g2 ∈ K×, we have;
If ga = g1a + g2a then gx = g1x + g2x for allx ∈ K+ − {0}.If g1a = −g2a then g1x = −g2x for all x ∈ K+ − {0}.
We can identify g ∈ K× with ga ∈ K+ − {0}.
Transfer the multiplication of K× to K+.
The manifold K+ gets the addition and the multiplication,both operations are Z-closed, hence K+ is a Z-field!
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Example 4.2.9 (1)
WorkV1 ⊆ (K+)∗. Consider a family of curves of the form
g(v) = a · v + b.
We need g(1) = 1. Hence, put b = 1− a. Thus,
ga(v) = a · v + 1− a
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Example 4.2.9 (2)
For ga(v) = a · v + 1− a, g(v) = b · v + 1− b, we have
ga(gb(v)) = a(b · v + 1− b) + 1− a = ab · v + 1− ab
On the other hand,
gab(v) = ab · v + 1− ab
Thus, we have
ga ◦ gb(v) = ga(gb(v)) = gab(v)
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Example 4.2.9 (3)
We use the multiplication to introduce
ga(v) ⊕ gb(v) = (a · v + 1− a) · (b · v + 1− b)= ab · v2 + (a + b− 2ab) · v + ab + 1− a− b= f (v)
This curve has a derivative at 1; f ′(1) = a + b.Thus f (v) is tangent to ga+b = (a + b) · v + 1− a− b at (1, 1).Therefore,
ga ⊕ gbTga+b
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Main Theorem
.
Theorem (Thm 4.4.1)
.
.
.
Let M be a Zariski structure satisfying (EU) andC be apre-smooth Zariski curve inM . Assume thatC is non-linear(i.e., ample). Then there is a non-constant continous map
f : C→ P1(K)
for some algebraically closed fieldK.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Chow Theorem and Purity Theorem
.
Theorem (Thm 4.3.9 generalized Chow Theorem)
.
.
.
Any closed subset of Pn is an algebraic subvariety of Pn.
Let K be the acf interpreted in an ample Zariski geometry M .
.
Theorem (Thm 4.3.10, purity Theorem)
.
.
.
Any relation R ⊆ Km definable in the natural language of Mis constructible i.e., a boolean conbination of algebraic sets.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Simple Zariski groups are algebraic
.
Theorem (Thm 4.4.6)
.
.
.
G is a simple Z-group satisfying (EU)
Some one-dim irred Z-subsetC ⊆ G is pre-smooth.
ThenG is Z-isomorphic to an algbraic groupG(K) for somealgebraically closed fieldK.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Z-meromorphic functions
Given a Zariski set N and an algebraically closed field K.
.
Definition (Def 4.4.2)
.
.
.
A continous function g : N → K with the domain containingan open subset of N is called Z-meromorphic.
.
Proposition
.
.
.
The set of all Z-meromorphic functions from N to K forms afield KZ(N).
.
Definition (p. 101)
.
.
.
KZ(N) is called the set of all Z-meromorphic functions fromN to K.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof of Theorem 4.4.6. (1)
Let m = rk( G).
(General property) For any G, simple group of finiteMorley rank, Th(G) is almost strongly-minimal, henceℵ1-categorical.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (2)
Let C be a strongly minimal definable subetset of G.There is a definable relation F ⊆ G× Cm giving afinite-to-finte correspondence between a subset R ⊆ Gand a subset D ⊆ Cm.
For G in the theorem, there exists a non-constantZ-meromorphic function G→ K.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (2)
Let C be a strongly minimal definable subetset of G.There is a definable relation F ⊆ G× Cm giving afinite-to-finte correspondence between a subset R ⊆ Gand a subset D ⊆ Cm.
For G in the theorem, there exists a non-constantZ-meromorphic function G→ K.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (3)
Consider the field KZ(G) of Z-meromorphic functionsfrom G to K.
Each g ∈ G acts on KZ(G) by f (x) 7→ f (g · c) giving arepresentation of G as a group of automorphisms ofKZ(G).
By Rosenlicht’s Theorem and the purity theorem, wesee that there is a G-invariant finite-dimensionalK-subspace V of KZ(G).
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (3)
Consider the field KZ(G) of Z-meromorphic functionsfrom G to K.
Each g ∈ G acts on KZ(G) by f (x) 7→ f (g · c) giving arepresentation of G as a group of automorphisms ofKZ(G).
By Rosenlicht’s Theorem and the purity theorem, wesee that there is a G-invariant finite-dimensionalK-subspace V of KZ(G).
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (3)
Consider the field KZ(G) of Z-meromorphic functionsfrom G to K.
Each g ∈ G acts on KZ(G) by f (x) 7→ f (g · c) giving arepresentation of G as a group of automorphisms ofKZ(G).
By Rosenlicht’s Theorem and the purity theorem, wesee that there is a G-invariant finite-dimensionalK-subspace V of KZ(G).
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (4)
G can be represented as a definable subgroup G(K) ofGL(V).
By the purity theorem again, the definable subgroupG(K) is algebraic.
The representation is isomorphism since G is simple.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (4)
G can be represented as a definable subgroup G(K) ofGL(V).
By the purity theorem again, the definable subgroupG(K) is algebraic.
The representation is isomorphism since G is simple.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Outline of proof (4)
G can be represented as a definable subgroup G(K) ofGL(V).
By the purity theorem again, the definable subgroupG(K) is algebraic.
The representation is isomorphism since G is simple.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
Geometric Mordel-Lang conjecture
Hrushovski proved Geometric Mordel-Lang conjecture usingZariski geometry argument.
Masanori Itai Zariski Geometries, Lecture 3
Getting a fieldMain Theorem
Applications
A recent application by Zilber, 2010
Zilber reproved a theorem of Bogomorph, Korotaev andTschinkel with the help of Zariski geometry.
.
Theorem
.
.
.
For any algebraically closed fields K1 and K2 there is a fieldisomorphism β : K1 → K2 inducing an isomorphism of pairsand a bijective isogeny ψ such that
α = ψ ◦ β : (J1; C1,+) → (J2 : C2,+)
Masanori Itai Zariski Geometries, Lecture 3