On the Geometry of the Moduli Space of Real Binary Octics

advertisement
On the Geometry of the Moduli Space
of Real Binary Octics
Kenneth Chu
Department of Mathematics
University of Utah
chu@math.utah.edu
October 18, 2005
The Moduli Space of Stable Real Octics
. – p.1/22
The Moduli Space of Stable Real Octics
Intention: To produce a non-arithmetic lattice in Isom(RH5 ).
. – p.1/22
The Moduli Space of Stable Real Octics
Intention: To produce a non-arithmetic lattice in Isom(RH5 ).
MsR,cubic surfaces
. – p.1/22
The Moduli Space of Stable Real Octics
Intention: To produce a non-arithmetic lattice in Isom(RH5 ).
MsR,cubic surfaces ∼
=
5
[
i=1
ΓiR,cs \RH4
| {z }
cs
MR,
i
. – p.1/22
The Moduli Space of Stable Real Octics
Intention: To produce a non-arithmetic lattice in Isom(RH5 ).
MsR,cubic surfaces ∼
=
5
[
i=1
ΓiR,cs \RH4 ∼
= ΓR,cs \RH4 .
| {z }
cs
MR,
i
. – p.1/22
The Moduli Space of Stable Real Octics
Intention: To produce a non-arithmetic lattice in Isom(RH5 ).
MsR,cubic surfaces ∼
=
5
[
i=1
ΓiR,cs \RH4 ∼
= ΓR,cs \RH4 .
| {z }
cs
MR,
i
Hope same phenomenon will occur with MsR,binary octics , i.e.
MR
s
. – p.1/22
The Moduli Space of Stable Real Octics
Intention: To produce a non-arithmetic lattice in Isom(RH5 ).
MsR,cubic surfaces ∼
=
5
[
i=1
ΓiR,cs \RH4 ∼
= ΓR,cs \RH4 .
| {z }
cs
MR,
i
Hope same phenomenon will occur with MsR,binary octics , i.e.
?
∼
MR
s =
4
[
i=0
5
ΓR
\RH
| i {z }
MR
s,i
. – p.1/22
The Moduli Space of Stable Real Octics
Intention: To produce a non-arithmetic lattice in Isom(RH5 ).
MsR,cubic surfaces ∼
=
5
[
i=1
ΓiR,cs \RH4 ∼
= ΓR,cs \RH4 .
| {z }
cs
MR,
i
Hope same phenomenon will occur with MsR,binary octics , i.e.
MR
s
?
∼
=
4
[
i=0
5
ΓR
\RH
i
| {z }
?
∼
= ΓR \RH5 .
MR
s,i
. – p.1/22
Summary of New Results
. – p.2/22
Summary of New Results
MR
s
?
∼
=
4
S
?
5 ∼
ΓR
= ΓR \RH5 .
i \RH
i=0 | {z }
MR
s,i
. – p.2/22
Summary of New Results
?
∼
MR
s =
4
S
?
5 ∼
ΓR
= ΓR \RH5 .
i \RH
i=0 | {z }
MR
s,i
i
0
1
2
3
4
# real points
8
6
4
2
0
# cx. conj. pairs
0
1
2
3
4
. – p.2/22
Summary of New Results
?
∼
MR
s =
4
S
?
5 ∼
ΓR
= ΓR \RH5 .
i \RH
i=0 | {z }
MR
s,i
i
0
1
2
3
4
# real points
8
6
4
2
0
# cx. conj. pairs
0
1
2
3
4
R
1. ΓR
0 , . . . , Γ3 have been found explicitly.
. – p.2/22
Summary of New Results
?
∼
MR
s =
4
S
?
5 ∼
ΓR
= ΓR \RH5 .
i \RH
i=0 | {z }
MR
s,i
i
0
1
2
3
4
# real points
8
6
4
2
0
# cx. conj. pairs
0
1
2
3
4
R
1. ΓR
0 , . . . , Γ3 have been found explicitly. Each is a finite-index subgroup
of an arithmetic reflection subgroup PAi ⊂ PIsom(RH5 ).
. – p.2/22
Summary of New Results
?
∼
MR
s =
4
S
?
5 ∼
ΓR
= ΓR \RH5 .
i \RH
i=0 | {z }
MR
s,i
i
0
1
2
3
4
# real points
8
6
4
2
0
# cx. conj. pairs
0
1
2
3
4
R
1. ΓR
0 , . . . , Γ3 have been found explicitly. Each is a finite-index subgroup
of an arithmetic reflection subgroup PAi ⊂ PIsom(RH5 ). The Coxeter
diagrams of A0 , . . . , A3 have been worked out.
. – p.2/22
Summary of New Results
?
∼
MR
s =
4
S
?
5 ∼
ΓR
= ΓR \RH5 .
i \RH
i=0 | {z }
MR
s,i
i
0
1
2
3
4
# real points
8
6
4
2
0
# cx. conj. pairs
0
1
2
3
4
R
1. ΓR
0 , . . . , Γ3 have been found explicitly. Each is a finite-index subgroup
of an arithmetic reflection subgroup PAi ⊂ PIsom(RH5 ). The Coxeter
diagrams of A0 , . . . , A3 have been worked out.
2. ΓR
4 is narrowed down to one of two possibilities.
. – p.2/22
Summary of New Results
?
∼
MR
s =
4
S
?
5 ∼
ΓR
= ΓR \RH5 .
i \RH
i=0 | {z }
MR
s,i
i
0
1
2
3
4
# real points
8
6
4
2
0
# cx. conj. pairs
0
1
2
3
4
R
1. ΓR
0 , . . . , Γ3 have been found explicitly. Each is a finite-index subgroup
of an arithmetic reflection subgroup PAi ⊂ PIsom(RH5 ). The Coxeter
diagrams of A0 , . . . , A3 have been worked out.
2. ΓR
4 is narrowed down to one of two possibilities. One of these
possibilities is shown to be a finite-index subgroup of an arithmetic
subgroup in PIsom(RH5 ) and its Coxeter diagram has been worked
out.
. – p.2/22
Summary of New Results
?
∼
MR
s =
4
S
?
5 ∼
ΓR
= ΓR \RH5 .
i \RH
i=0 | {z }
MR
s,i
i
0
1
2
3
4
# real points
8
6
4
2
0
# cx. conj. pairs
0
1
2
3
4
R
1. ΓR
0 , . . . , Γ3 have been found explicitly. Each is a finite-index subgroup
of an arithmetic reflection subgroup PAi ⊂ PIsom(RH5 ). The Coxeter
diagrams of A0 , . . . , A3 have been worked out.
2. ΓR
4 is narrowed down to one of two possibilities. One of these
possibilities is shown to be a finite-index subgroup of an arithmetic
subgroup in PIsom(RH5 ) and its Coxeter diagram has been worked
out.
3. (The A-C-T construction of) MR
s cannot admit a real hyperbolic
orbifold structure:
. – p.2/22
Summary of New Results
?
∼
MR
s =
4
S
?
5 ∼
ΓR
= ΓR \RH5 .
i \RH
i=0 | {z }
MR
s,i
i
0
1
2
3
4
# real points
8
6
4
2
0
# cx. conj. pairs
0
1
2
3
4
R
1. ΓR
0 , . . . , Γ3 have been found explicitly. Each is a finite-index subgroup
of an arithmetic reflection subgroup PAi ⊂ PIsom(RH5 ). The Coxeter
diagrams of A0 , . . . , A3 have been worked out.
2. ΓR
4 is narrowed down to one of two possibilities. One of these
possibilities is shown to be a finite-index subgroup of an arithmetic
subgroup in PIsom(RH5 ) and its Coxeter diagram has been worked
out.
3. (The A-C-T construction of) MR
s cannot admit a real hyperbolic
orbifold structure: Points in the stratum ∆0,1 are not locally real
hyperbolic (modulo a finite group of isometries).
. – p.2/22
Strategy
1. Start with
. – p.3/22
Strategy
1. Start with
Ms
∼
=
PΓ\CH5
[Deligne-Mostow]
. – p.3/22
Strategy
1. Start with
`
Ms
∼
=
PΓ\CH5
[Deligne-Mostow]
k
´
CP8 − ∆≥4 /PGL(2, C)
. – p.3/22
Strategy
1. Start with
`
Ms
∼
=
PΓ\CH5
[Deligne-Mostow]
k
´
CP8 − ∆≥4 /PGL(2, C)
5
2. A-C-T observed: Periods (in CH ) of real octics lie in
S
χ∈?
RH5χ ⊂ CH5 .
. – p.3/22
Strategy
1. Start with
`
Ms
∼
=
PΓ\CH5
[Deligne-Mostow]
k
´
CP8 − ∆≥4 /PGL(2, C)
5
2. A-C-T observed: Periods (in CH ) of real octics lie in
S
χ∈?
So,
RH5χ ⊂ CH5 .
. – p.3/22
Strategy
1. Start with
`
Ms
∼
=
PΓ\CH5
[Deligne-Mostow]
k
´
CP8 − ∆≥4 /PGL(2, C)
5
2. A-C-T observed: Periods (in CH ) of real octics lie in
S
χ∈?
So,
RH5χ ⊂ CH5 .
MR
s
k
”
“
8
R
RP − ∆≥4 /PGL(2, R)
. – p.3/22
Strategy
1. Start with
`
∼
=
Ms
PΓ\CH5
[Deligne-Mostow]
k
´
CP8 − ∆≥4 /PGL(2, C)
5
2. A-C-T observed: Periods (in CH ) of real octics lie in
S
χ∈?
So,
MR
s
k
∼
= PΓ
-(
`
χ∈?
RH5χ
!,
≈
RH5χ ⊂ CH5 .
)
=: PΓ\Ks
”
“
8
R
RP − ∆≥4 /PGL(2, R)
. – p.3/22
Strategy
1. Start with
`
∼
=
Ms
PΓ\CH5
[Deligne-Mostow]
k
´
CP8 − ∆≥4 /PGL(2, C)
5
2. A-C-T observed: Periods (in CH ) of real octics lie in
S
χ∈?
So,
MR
s
∼
= PΓ
k
-(
`
χ∈?
RH5χ
!,
≈
RH5χ ⊂ CH5 .
)
=: PΓ\Ks
”
“
8
R
RP − ∆≥4 /PGL(2, R)
‘
is to undo PGL(2, C)-quotienting.
. – p.3/22
Strategy
1. Start with
`
∼
=
Ms
PΓ\CH5
[Deligne-Mostow]
k
´
CP8 − ∆≥4 /PGL(2, C)
5
2. A-C-T observed: Periods (in CH ) of real octics lie in
S
χ∈?
So,
MR
s
k
∼
= PΓ
-(
`
χ∈?
RH5χ
!,
≈
RH5χ ⊂ CH5 .
)
=: PΓ\Ks
”
“
8
R
RP − ∆≥4 /PGL(2, R)
‘
is to undo PGL(2, C)-quotienting. ≈ is to impose PGL(2, R)-quotienting.
. – p.3/22
Strategy
1. Start with
`
∼
=
Ms
PΓ\CH5
[Deligne-Mostow]
k
´
CP8 − ∆≥4 /PGL(2, C)
5
2. A-C-T observed: Periods (in CH ) of real octics lie in
S
χ∈?
So,
MR
s
∼
= PΓ
k
-(
`
χ∈?
RH5χ
!,
≈
RH5χ ⊂ CH5 .
)
=: PΓ\Ks
”
“
8
R
RP − ∆≥4 /PGL(2, R)
‘
is to undo PGL(2, C)-quotienting. ≈ is to impose PGL(2, R)-quotienting.
3. Study the geometry of PΓ\Ks
. – p.3/22
Strategy
1. Start with
`
∼
=
Ms
PΓ\CH5
[Deligne-Mostow]
k
´
CP8 − ∆≥4 /PGL(2, C)
5
2. A-C-T observed: Periods (in CH ) of real octics lie in
S
χ∈?
So,
MR
s
k
∼
= PΓ
-(
`
χ∈?
RH5χ
!,
≈
RH5χ ⊂ CH5 .
)
=: PΓ\Ks
”
“
8
R
RP − ∆≥4 /PGL(2, R)
‘
is to undo PGL(2, C)-quotienting. ≈ is to impose PGL(2, R)-quotienting.
3. Study the geometry of PΓ\Ks =: A-C-T construction of the moduli
space of stable real binary octics.
. – p.3/22
Deligne-Mostow Construction of Ms
. – p.4/22
Deligne-Mostow Construction of Ms
Let p(x) ∈ P0 .
. – p.4/22
Deligne-Mostow Construction of Ms
Let p(x) ∈ P0 . Define Xp to be the completion of
4
(x, y) ∈ C
y − p(x) = 0
2
. – p.4/22
Deligne-Mostow Construction of Ms
Let p(x) ∈ P0 . Define Xp to be the completion of
4
(x, y) ∈ C
y − p(x) = 0
2
Xp is a Riemann surface
. – p.4/22
Deligne-Mostow Construction of Ms
Let p(x) ∈ P0 . Define Xp to be the completion of
4
(x, y) ∈ C
y − p(x) = 0
2
π
Xp is a Riemann surface and the map Xp → C : (x, y) 7→ x is a quadruple
cyclic covering of CP1 branched over the 8 distinct roots of p(x).
. – p.4/22
Deligne-Mostow Construction of Ms
Let p(x) ∈ P0 . Define Xp to be the completion of
4
(x, y) ∈ C
y − p(x) = 0
2
π
Xp is a Riemann surface and the map Xp → C : (x, y) 7→ x is a quadruple
cyclic covering of CP1 branched over the 8 distinct roots of p(x). Each
ramification point has ramification index 4. Riemann-Hurwitz ⇒ genus(Xp )
= H 1,0 (Xp ) = 9 and rankZ H 1 (Xp , Z) = dimC H 1 (Xp , C) = 2 · 9 = 18.
. – p.4/22
Deligne-Mostow Construction of Ms
Let p(x) ∈ P0 . Define Xp to be the completion of
4
√
σ
(x, y) ∈ C
y − p(x) = 0 =: Xp −→ Xp : (x, y) 7→ (x, −1y).
2
π
Xp is a Riemann surface and the map Xp → C : (x, y) 7→ x is a quadruple
cyclic covering of CP1 branched over the 8 distinct roots of p(x). Each
ramification point has ramification index 4. Riemann-Hurwitz ⇒ genus(Xp )
= H 1,0 (Xp ) = 9 and rankZ H 1 (Xp , Z) = dimC H 1 (Xp , C) = 2 · 9 = 18.
. – p.4/22
Deligne-Mostow Construction of Ms
Let p(x) ∈ P0 . Define Xp to be the completion of
4
√
σ
(x, y) ∈ C
y − p(x) = 0 =: Xp −→ Xp : (x, y) 7→ (x, −1y).
2
π
Xp is a Riemann surface and the map Xp → C : (x, y) 7→ x is a quadruple
cyclic covering of CP1 branched over the 8 distinct roots of p(x). Each
ramification point has ramification index 4. Riemann-Hurwitz ⇒ genus(Xp )
= H 1,0 (Xp ) = 9 and rankZ H 1 (Xp , Z) = dimC H 1 (Xp , C) = 2 · 9 = 18.
Let Λ(Xp ) := Hσ12 =−1 (Xp , Z).
. – p.4/22
Deligne-Mostow Construction of Ms
Let p(x) ∈ P0 . Define Xp to be the completion of
4
√
σ
(x, y) ∈ C
y − p(x) = 0 =: Xp −→ Xp : (x, y) 7→ (x, −1y).
2
π
Xp is a Riemann surface and the map Xp → C : (x, y) 7→ x is a quadruple
cyclic covering of CP1 branched over the 8 distinct roots of p(x). Each
ramification point has ramification index 4. Riemann-Hurwitz ⇒ genus(Xp )
= H 1,0 (Xp ) = 9 and rankZ H 1 (Xp , Z) = dimC H 1 (Xp , C) = 2 · 9 = 18.
Let Λ(Xp ) :=
σ 2 + 1 = 0.
Hσ12 =−1 (Xp , Z).
σ∗
Then, Λ(Xp ) −→ Λ(Xp ) satisfies
. – p.4/22
Deligne-Mostow Construction of Ms
Let p(x) ∈ P0 . Define Xp to be the completion of
4
√
σ
(x, y) ∈ C
y − p(x) = 0 =: Xp −→ Xp : (x, y) 7→ (x, −1y).
2
π
Xp is a Riemann surface and the map Xp → C : (x, y) 7→ x is a quadruple
cyclic covering of CP1 branched over the 8 distinct roots of p(x). Each
ramification point has ramification index 4. Riemann-Hurwitz ⇒ genus(Xp )
= H 1,0 (Xp ) = 9 and rankZ H 1 (Xp , Z) = dimC H 1 (Xp , C) = 2 · 9 = 18.
Hσ12 =−1 (Xp , Z).
σ∗
Let Λ(Xp ) :=
Then, Λ(Xp ) −→ Λ(Xp ) satisfies
√
2
σ + 1 = 0. Hence Λ(Xp ) becomes a Z[ −1]-module
. – p.4/22
Deligne-Mostow Construction of Ms
Let p(x) ∈ P0 . Define Xp to be the completion of
4
√
σ
(x, y) ∈ C
y − p(x) = 0 =: Xp −→ Xp : (x, y) 7→ (x, −1y).
2
π
Xp is a Riemann surface and the map Xp → C : (x, y) 7→ x is a quadruple
cyclic covering of CP1 branched over the 8 distinct roots of p(x). Each
ramification point has ramification index 4. Riemann-Hurwitz ⇒ genus(Xp )
= H 1,0 (Xp ) = 9 and rankZ H 1 (Xp , Z) = dimC H 1 (Xp , C) = 2 · 9 = 18.
Hσ12 =−1 (Xp , Z).
σ∗
Let Λ(Xp ) :=
Then, Λ(Xp ) −→ Λ(Xp ) satisfies
√
2
σ + 1 = 0. Hence Λ(Xp ) becomes a Z[ −1]-module via
√
− −1 · φ := σ ∗ (φ).
. – p.4/22
Deligne-Mostow Construction of Ms
Let p(x) ∈ P0 . Define Xp to be the completion of
4
√
σ
(x, y) ∈ C
y − p(x) = 0 =: Xp −→ Xp : (x, y) 7→ (x, −1y).
2
π
Xp is a Riemann surface and the map Xp → C : (x, y) 7→ x is a quadruple
cyclic covering of CP1 branched over the 8 distinct roots of p(x). Each
ramification point has ramification index 4. Riemann-Hurwitz ⇒ genus(Xp )
= H 1,0 (Xp ) = 9 and rankZ H 1 (Xp , Z) = dimC H 1 (Xp , C) = 2 · 9 = 18.
Hσ12 =−1 (Xp , Z).
σ∗
Let Λ(Xp ) :=
Then, Λ(Xp ) −→ Λ(Xp ) satisfies
√
2
σ + 1 = 0. Hence Λ(Xp ) becomes a Z[ −1]-module via
√
− −1 · φ := σ ∗ (φ).
FACT:
√
∼
Λ(Xp ) = Z[ −1]6 .
. – p.4/22
Hermitian Form on Λ(Xp )
. – p.5/22
Hermitian Form on Λ(Xp )
Consider the embedding
Λ(Xp )
k
Hσ12 =−1 (Xp , Z)
. – p.5/22
Hermitian Form on Λ(Xp )
Consider the embedding
Λ(Xp )
֒→
Hσ12 =−1 (Xp , Z) ⊗Z C
k
Hσ12 =−1 (Xp , Z)
. – p.5/22
Hermitian Form on Λ(Xp )
Consider the embedding
Λ(Xp )
֒→
Hσ12 =−1 (Xp , Z) ⊗Z C
∼
=
Hσ12 =−1 (Xp , C)
k
Hσ12 =−1 (Xp , Z)
. – p.5/22
Hermitian Form on Λ(Xp )
Consider the embedding
Λ(Xp )
k
Hσ12 =−1 (Xp , Z)
֒→
Hσ12 =−1 (Xp , Z) ⊗Z C
∼
=
Hσ12 =−1 (Xp , C)
k
1 √
1
√
(X
,
C)
⊕
H
(Xp , C)
Hσ=−
p
σ= −1
−1
. – p.5/22
Hermitian Form on Λ(Xp )
Consider the embedding
Λ(Xp )
k
Hσ12 =−1 (Xp , Z)
֒→
Hσ12 =−1 (Xp , Z) ⊗Z C
∼
=
Hσ12 =−1 (Xp , C)
k
1 √
1
√
(X
,
C)
⊕
H
(Xp , C)
Hσ=−
p
σ= −1
−1
↓
1
√
Hσ=−
(Xp , C)
−1
. – p.5/22
Hermitian Form on Λ(Xp )
Consider the embedding
Λ(Xp )
Hσ12 =−1 (Xp , Z) ⊗Z C
֒→
∼
=
Hσ12 =−1 (Xp , C)
k
k
1 √
1
√
(X
,
C)
⊕
H
(Xp , C)
Hσ=−
p
σ= −1
−1
Hσ12 =−1 (Xp , Z)
↓
1
√
Hσ=−
(Xp , C)
−1
Computations show
0
B
B
B
B
B
@
1
√
Hσ=−
(Xp , C)
−1
|
{z
k
}
0,1
1,0
√
√
(X)
⊕
H
(X)
Hσ=−
σ=− −1
−1
(+)
(−−−−−)
√
h
, (α, β) 7−→ 2 −1
′
Z
Xp
1
C
C
C
α∧β C ∼
= C1,5
C
A
. – p.5/22
Hermitian Form on Λ(Xp ) (cont’d)
. – p.6/22
Hermitian Form on Λ(Xp ) (cont’d)
Λ(Xp ) =
Hσ12 =−1 (Xp , Z)
֒→
1
√
Hσ=−
(Xp , C) ,
−1
′
h
∼
= C1,5
. – p.6/22
Hermitian Form on Λ(Xp ) (cont’d)
Λ(Xp ) =
Hσ12 =−1 (Xp , Z)
֒→
1
√
Hσ=−
(Xp , C) ,
−1
′
h
∼
= C1,5
The pullback Hermitian form on Λ(Xp ) is given by:
√
h(ξ, η) = − Ω(ξ, σ(η)) − −1 Ω(ξ, η),
where Ω(ξ, η) := h ξ ∪ η , Xp i.
. – p.6/22
Hermitian Form on Λ(Xp ) (cont’d)
Λ(Xp ) =
Hσ12 =−1 (Xp , Z)
֒→
1
√
Hσ=−
(Xp , C) ,
−1
′
h
∼
= C1,5
The pullback Hermitian form on Λ(Xp ) is given by:
√
h(ξ, η) = − Ω(ξ, σ(η)) − −1 Ω(ξ, η),
where Ω(ξ, η) := h ξ ∪ η , Xp i.
√
Computations show (Λ(Xp ) , h) is abstractly isometric to Λ := Z[ −1]6 ,
equipped with
2
4
−2
√
1 − −1
1+
√
−1
−2
3
2
5⊕4
−2
√
1 − −1
1+
√
−1
−2
3
2
5⊕4
1−
0
√
1+
−1
√
0
−1
3
5
. – p.6/22
2
Explanation of 4
−2
√
1 − −1
1+
√
−1
−2
3
5
. – p.7/22
2
Explanation of 4
−2
√
1 − −1
1+
√
−1
−2
3
5
The vanishing cohomology corresponding to a nodal octic has
√
Z[ −1]-rank one
. – p.7/22
2
Explanation of 4
−2
√
1 − −1
1+
√
−1
−2
3
5
The vanishing cohomology corresponding to a nodal octic has
√
Z[ −1]-rank one and is generated by a vector of norm -2.
. – p.7/22
2
Explanation of 4
−2
√
1 − −1
1+
√
−1
−2
3
5
The vanishing cohomology corresponding to a nodal octic has
√
Z[ −1]-rank one and is generated by a vector of norm -2.
The vanishing cohomology corresponding to a cuspidal octic is an
orthogonal summand of Λ(Xp ) ∼
=Λ
. – p.7/22
2
Explanation of 4
−2
√
1 − −1
1+
√
−1
−2
3
5
The vanishing cohomology corresponding to a nodal octic has
√
Z[ −1]-rank one and is generated by a vector of norm -2.
The vanishing cohomology corresponding to a cuspidal octic is an
√
∼
orthogonal summand of Λ(Xp ) = Λ of Z[ −1]-rank two
. – p.7/22
2
Explanation of 4
−2
√
1 − −1
1+
√
−1
−2
3
5
The vanishing cohomology corresponding to a nodal octic has
√
Z[ −1]-rank one and is generated by a vector of norm -2.
The vanishing cohomology corresponding to a cuspidal octic is an
√
∼
orthogonal summand of Λ(Xp ) = Λ of Z[ −1]-rank two with inherited
√
Z[ −1]-Hermitian form:


√
 Z[ −1]2 , 
−2
√
1 − −1
1+
√
−1
−2


. – p.7/22
2
Explanation of 4
−2
√
1 − −1
1+
√
−1
−2
3
5
The vanishing cohomology corresponding to a nodal octic has
√
Z[ −1]-rank one and is generated by a vector of norm -2.
The vanishing cohomology corresponding to a cuspidal octic is an
√
∼
orthogonal summand of Λ(Xp ) = Λ of Z[ −1]-rank two with inherited
√
Z[ −1]-Hermitian form:


√
 Z[ −1]2 , 
−2
√
1 − −1
1+
√
−1
−2


OUTLINE OF PROOF
Local “pictorial” computations of the intersection form of the vanishing 1-homology of
H1,σ2 =−1 (Xp , Z) over a coalescing two-point or three-point configurations,
. – p.7/22
2
Explanation of 4
−2
√
1 − −1
1+
√
−1
−2
3
5
The vanishing cohomology corresponding to a nodal octic has
√
Z[ −1]-rank one and is generated by a vector of norm -2.
The vanishing cohomology corresponding to a cuspidal octic is an
√
∼
orthogonal summand of Λ(Xp ) = Λ of Z[ −1]-rank two with inherited
√
Z[ −1]-Hermitian form:


√
 Z[ −1]2 , 
−2
√
1 − −1
1+
√
−1
−2


OUTLINE OF PROOF
Local “pictorial” computations of the intersection form of the vanishing 1-homology of
H1,σ2 =−1 (Xp , Z) over a coalescing two-point or three-point configurations, i.e. the
intersection form of the vanishing 1-homology of
y 4 = x2 − ǫ2
and
y 4 = x3 − ǫ3 ,
as ǫ
→ 0, ǫ ≥ 0.
. – p.7/22
“Fiberwise” Summary
p ∈ P0
. – p.8/22
“Fiberwise” Summary
p ∈ P0
8
>
>
<
>
>
:
quadruple cyclic covering Xp
of CP1 branched at roots of p
with cyclic action Xp
σ
→ Xp of order 4
. – p.8/22
“Fiberwise” Summary
p ∈ P0
8
>
>
<
>
>
:
quadruple cyclic covering Xp
of CP1 branched at roots of p
with cyclic action Xp
σ
→ Xp of order 4
“
”
√
1
Z[ −1]-lattice Λ(Xp ) := Hσ2 =−1 (X, Z) , h
. – p.8/22
“Fiberwise” Summary
p ∈ P0
8
>
>
<
>
>
:
quadruple cyclic covering Xp
of CP1 branched at roots of p
with cyclic action Xp
σ
→ Xp of order 4
“
”
√
1
Z[ −1]-lattice Λ(Xp ) := Hσ2 =−1 (X, Z) , h ∼
=Λ
. – p.8/22
“Fiberwise” Summary
p ∈ P0
8
>
>
<
>
>
:
quadruple cyclic covering Xp
of CP1 branched at roots of p
with cyclic action Xp
σ
→ Xp of order 4
“
”
√
1
Z[ −1]-lattice Λ(Xp ) := Hσ2 =−1 (X, Z) , h ∼
=Λ
Recall
Λ ⊗Z[√−1] C ∼
= C1,5 = C1+,5− ,
. – p.8/22
“Fiberwise” Summary
8
>
>
<
p ∈ P0
>
>
:
quadruple cyclic covering Xp
of CP1 branched at roots of p
with cyclic action Xp
σ
→ Xp of order 4
“
”
√
1
Z[ −1]-lattice Λ(Xp ) := Hσ2 =−1 (X, Z) , h ∼
=Λ
Recall
2
4
Λ ⊗Z[√−1] C ∼
= C1,5 = C1+,5− ,
−2
√
1 − −1
1+
√
−1
−2
3
2
5⊕4
√
and Λ := Z[ −1]6 equipped with
−2
√
1 − −1
1+
√
−1
−2
3
2
5⊕4
1−
0
√
1+
−1
√
0
−1
3
5
. – p.8/22
“Fiberwise” Summary
8
>
>
<
p ∈ P0
>
>
:
quadruple cyclic covering Xp
of CP1 branched at roots of p
with cyclic action Xp
σ
→ Xp of order 4
“
”
√
1
Z[ −1]-lattice Λ(Xp ) := Hσ2 =−1 (X, Z) , h ∼
=Λ
Recall
2
4
Λ ⊗Z[√−1] C ∼
= C1,5 = C1+,5− ,
−2
√
1 − −1
1+
√
−1
−2
3
2
5⊕4
√
and Λ := Z[ −1]6 equipped with
−2
√
1 − −1
1+
√
−1
−2
3
2
5⊕4
1−
0
√
1+
−1
√
0
−1
3
5
1,0
√
(Xp ) is a positive 1-dimensional subspace of
Recall also Hσ=−
−1
“
1
√
Hσ=−
(Xp , C) ,
−1
h
”
∼
= C1,5 = C1+,5− .
. – p.8/22
“Fiberwise” Summary
8
>
>
<
p ∈ P0
>
>
:
quadruple cyclic covering Xp
of CP1 branched at roots of p
with cyclic action Xp
σ
→ Xp of order 4
“
”
√
1
Z[ −1]-lattice Λ(Xp ) := Hσ2 =−1 (X, Z) , h ∼
=Λ
Recall
2
4
Λ ⊗Z[√−1] C ∼
= C1,5 = C1+,5− ,
−2
√
1 − −1
1+
√
−1
−2
3
2
5⊕4
√
and Λ := Z[ −1]6 equipped with
−2
√
1 − −1
1+
√
−1
−2
3
2
5⊕4
1−
0
√
1+
−1
√
0
−1
3
5
1,0
√
(Xp ) is a positive 1-dimensional subspace of
Recall also Hσ=−
−1
“
Hence,
1
√
Hσ=−
(Xp , C) ,
−1
1,0
√
Hσ=−
(Xp )
−1
h
∈ CH
”
“
∼
= C1,5 = C1+,5− .
1
√
Hσ=−
(Xp , C)
−1
”
∼
= CH5
. – p.8/22
Construction of F0 = Domain(Period Map) −→ P0
. – p.9/22
Construction of F0 = Domain(Period Map) −→ P0
A framed smooth form over p ∈ P0 is a “projective equivalence class” of
∼
an (abstract) isometry of Λ(Xp ) −→ Λ,
. – p.9/22
Construction of F0 = Domain(Period Map) −→ P0
A framed smooth form over p ∈ P0 is a “projective equivalence class” of
∼
an (abstract) isometry of Λ(Xp ) −→ Λ, where two such isometries are
√
“projectively equivalent” if one is a Z[ −1]-unit scalar multiple of the other.
. – p.9/22
Construction of F0 = Domain(Period Map) −→ P0
A framed smooth form over p ∈ P0 is a “projective equivalence class” of
∼
an (abstract) isometry of Λ(Xp ) −→ Λ, where two such isometries are
√
“projectively equivalent” if one is a Z[ −1]-unit scalar multiple of the other.
F0 is the space of all framed smooth forms, and we get a natural
projection map F0 → P0 , which is in fact an unbranched covering.
. – p.9/22
Construction of F0 = Domain(Period Map) −→ P0
A framed smooth form over p ∈ P0 is a “projective equivalence class” of
∼
an (abstract) isometry of Λ(Xp ) −→ Λ, where two such isometries are
√
“projectively equivalent” if one is a Z[ −1]-unit scalar multiple of the other.
F0 is the space of all framed smooth forms, and we get a natural
projection map F0 → P0 , which is in fact an unbranched covering.
GL(2, C)-action
cohomology.
on P0 naturally extends to F0 via induced action on
. – p.9/22
Construction of F0 = Domain(Period Map) −→ P0
A framed smooth form over p ∈ P0 is a “projective equivalence class” of
∼
an (abstract) isometry of Λ(Xp ) −→ Λ, where two such isometries are
√
“projectively equivalent” if one is a Z[ −1]-unit scalar multiple of the other.
F0 is the space of all framed smooth forms, and we get a natural
projection map F0 → P0 , which is in fact an unbranched covering.
GL(2, C)-action
on P0 naturally extends to F0 via induced action on
√
cohomology. Let G := GL(2, C)/h±1, ± −1i.
. – p.9/22
Construction of F0 = Domain(Period Map) −→ P0
A framed smooth form over p ∈ P0 is a “projective equivalence class” of
∼
an (abstract) isometry of Λ(Xp ) −→ Λ, where two such isometries are
√
“projectively equivalent” if one is a Z[ −1]-unit scalar multiple of the other.
F0 is the space of all framed smooth forms, and we get a natural
projection map F0 → P0 , which is in fact an unbranched covering.
GL(2, C)-action
on P0 naturally extends to F0 via induced action on
√
cohomology. Let G := GL(2, C)/h±1, ± −1i. Then G acts freely on F0 .
. – p.9/22
Construction of F0 = Domain(Period Map) −→ P0
A framed smooth form over p ∈ P0 is a “projective equivalence class” of
∼
an (abstract) isometry of Λ(Xp ) −→ Λ, where two such isometries are
√
“projectively equivalent” if one is a Z[ −1]-unit scalar multiple of the other.
F0 is the space of all framed smooth forms, and we get a natural
projection map F0 → P0 , which is in fact an unbranched covering.
GL(2, C)-action
on P0 naturally extends to F0 via induced action on
√
cohomology. Let G := GL(2, C)/h±1, ± −1i. Then G acts freely on F0 .
FACTS
1. Let PΓ be the Deck transformation group of the covering F0 → P0 . Then we know
PΓ ⊆ PIsom(Λ),
. – p.9/22
Construction of F0 = Domain(Period Map) −→ P0
A framed smooth form over p ∈ P0 is a “projective equivalence class” of
∼
an (abstract) isometry of Λ(Xp ) −→ Λ, where two such isometries are
√
“projectively equivalent” if one is a Z[ −1]-unit scalar multiple of the other.
F0 is the space of all framed smooth forms, and we get a natural
projection map F0 → P0 , which is in fact an unbranched covering.
GL(2, C)-action
on P0 naturally extends to F0 via induced action on
√
cohomology. Let G := GL(2, C)/h±1, ± −1i. Then G acts freely on F0 .
FACTS
1. Let PΓ be the Deck transformation group of the covering F0 → P0 . Then we know
PΓ ⊆ PIsom(Λ), since for each p ∈ P0 , Λ(Xp ) is abstractly isometric to Λ and each
Deck transformation must preserve Λ(Xp ).
. – p.9/22
Construction of F0 = Domain(Period Map) −→ P0
A framed smooth form over p ∈ P0 is a “projective equivalence class” of
∼
an (abstract) isometry of Λ(Xp ) −→ Λ, where two such isometries are
√
“projectively equivalent” if one is a Z[ −1]-unit scalar multiple of the other.
F0 is the space of all framed smooth forms, and we get a natural
projection map F0 → P0 , which is in fact an unbranched covering.
GL(2, C)-action
on P0 naturally extends to F0 via induced action on
√
cohomology. Let G := GL(2, C)/h±1, ± −1i. Then G acts freely on F0 .
FACTS
1. Let PΓ be the Deck transformation group of the covering F0 → P0 . Then we know
PΓ ⊆ PIsom(Λ), since for each p ∈ P0 , Λ(Xp ) is abstractly isometric to Λ and each
Deck transformation must preserve Λ(Xp ). Clearly, PΓ\F0 ∼
= P0 .
. – p.9/22
Construction of F0 = Domain(Period Map) −→ P0
A framed smooth form over p ∈ P0 is a “projective equivalence class” of
∼
an (abstract) isometry of Λ(Xp ) −→ Λ, where two such isometries are
√
“projectively equivalent” if one is a Z[ −1]-unit scalar multiple of the other.
F0 is the space of all framed smooth forms, and we get a natural
projection map F0 → P0 , which is in fact an unbranched covering.
GL(2, C)-action
on P0 naturally extends to F0 via induced action on
√
cohomology. Let G := GL(2, C)/h±1, ± −1i. Then G acts freely on F0 .
FACTS
1. Let PΓ be the Deck transformation group of the covering F0 → P0 . Then we know
PΓ ⊆ PIsom(Λ), since for each p ∈ P0 , Λ(Xp ) is abstractly isometric to Λ and each
Deck transformation must preserve Λ(Xp ). Clearly, PΓ\F0 ∼
= P0 .
2. F0 is also the covering corresponding to the kernel of the representation
ρ
π1 (P0 , p0 ) −→ PIsom(Λ(Xp0 )),
. – p.9/22
Construction of F0 = Domain(Period Map) −→ P0
A framed smooth form over p ∈ P0 is a “projective equivalence class” of
∼
an (abstract) isometry of Λ(Xp ) −→ Λ, where two such isometries are
√
“projectively equivalent” if one is a Z[ −1]-unit scalar multiple of the other.
F0 is the space of all framed smooth forms, and we get a natural
projection map F0 → P0 , which is in fact an unbranched covering.
GL(2, C)-action
on P0 naturally extends to F0 via induced action on
√
cohomology. Let G := GL(2, C)/h±1, ± −1i. Then G acts freely on F0 .
FACTS
1. Let PΓ be the Deck transformation group of the covering F0 → P0 . Then we know
PΓ ⊆ PIsom(Λ), since for each p ∈ P0 , Λ(Xp ) is abstractly isometric to Λ and each
Deck transformation must preserve Λ(Xp ). Clearly, PΓ\F0 ∼
= P0 .
2. F0 is also the covering corresponding to the kernel of the representation
ρ
π1 (P0 , p0 ) −→ PIsom(Λ(Xp0 )), where p0 ∈ P0 is some fixed smooth octic. PΓ is thus
also the monodromy group of the representation ρ.
. – p.9/22
Construction of F0 = Domain(Period Map) −→ P0
A framed smooth form over p ∈ P0 is a “projective equivalence class” of
∼
an (abstract) isometry of Λ(Xp ) −→ Λ, where two such isometries are
√
“projectively equivalent” if one is a Z[ −1]-unit scalar multiple of the other.
F0 is the space of all framed smooth forms, and we get a natural
projection map F0 → P0 , which is in fact an unbranched covering.
GL(2, C)-action
on P0 naturally extends to F0 via induced action on
√
cohomology. Let G := GL(2, C)/h±1, ± −1i. Then G acts freely on F0 .
FACTS
1. Let PΓ be the Deck transformation group of the covering F0 → P0 . Then we know
PΓ ⊆ PIsom(Λ), since for each p ∈ P0 , Λ(Xp ) is abstractly isometric to Λ and each
Deck transformation must preserve Λ(Xp ). Clearly, PΓ\F0 ∼
= P0 .
2. F0 is also the covering corresponding to the kernel of the representation
ρ
π1 (P0 , p0 ) −→ PIsom(Λ(Xp0 )), where p0 ∈ P0 is some fixed smooth octic. PΓ is thus
also the monodromy group of the representation ρ.
3. In fact, PΓ = PIsom(Λ).
. – p.9/22
The Period Map F0 → CH5
5
F0 −→ CH = CH Λ
h
i
f
Λ(Xp ) → Λ 7−→
⊗Z[√−1]
C
. – p.10/22
The Period Map F0 → CH5
5
⊗Z[√−1]
F0 −→ CH = CH Λ
h
i
f
1,0
√ (Xp ))
Λ(Xp ) → Λ 7−→ f (Hσ=−
−1
C
. – p.10/22
The Period Map F0 → CH5
5
⊗Z[√−1]
F0 −→ CH = CH Λ
h
i
f
1,0
√ (Xp ))
Λ(Xp ) → Λ 7−→ f (Hσ=−
−1
C
FACTS
. – p.10/22
The Period Map F0 → CH5
5
⊗Z[√−1]
F0 −→ CH = CH Λ
h
i
f
1,0
√ (Xp ))
Λ(Xp ) → Λ 7−→ f (Hσ=−
−1
C
FACTS
1. The period map is equivariant with respect to the actions of PΓ on F0
(via Deck transformations ↔ change of basis of projectivized frames)
. – p.10/22
The Period Map F0 → CH5
5
⊗Z[√−1]
F0 −→ CH = CH Λ
h
i
f
1,0
√ (Xp ))
Λ(Xp ) → Λ 7−→ f (Hσ=−
−1
C
FACTS
1. The period map is equivariant with respect to the actions of PΓ on F0
(via Deck transformations ↔ change of basis of projectivized frames) and on CH5
(via isometries).
. – p.10/22
The Period Map F0 → CH5
5
⊗Z[√−1]
F0 −→ CH = CH Λ
h
i
f
1,0
√ (Xp ))
Λ(Xp ) → Λ 7−→ f (Hσ=−
−1
C
FACTS
1. The period map is equivariant with respect to the actions of PΓ on F0
(via Deck transformations ↔ change of basis of projectivized frames) and on CH5
(via isometries). It is also (G y F0 )-invariant.
. – p.10/22
The Period Map F0 → CH5
5
⊗Z[√−1]
F0 −→ CH = CH Λ
h
i
f
1,0
√ (Xp ))
Λ(Xp ) → Λ 7−→ f (Hσ=−
−1
C
FACTS
1. The period map is equivariant with respect to the actions of PΓ on F0
(via Deck transformations ↔ change of basis of projectivized frames) and on CH5
(via isometries). It is also (G y F0 )-invariant.
S
2. Let H := r∈R CH(r ⊥ ) ⊂ CH5 , where R is the set of all vectors in
Λ of norm -2.
. – p.10/22
The Period Map F0 → CH5
5
⊗Z[√−1]
F0 −→ CH = CH Λ
h
i
f
1,0
√ (Xp ))
Λ(Xp ) → Λ 7−→ f (Hσ=−
−1
C
FACTS
1. The period map is equivariant with respect to the actions of PΓ on F0
(via Deck transformations ↔ change of basis of projectivized frames) and on CH5
(via isometries). It is also (G y F0 )-invariant.
S
2. Let H := r∈R CH(r ⊥ ) ⊂ CH5 , where R is the set of all vectors in
Λ of norm -2. FACT: the period map maps F0 onto CH5 − H and
. – p.10/22
The Period Map F0 → CH5
5
⊗Z[√−1]
F0 −→ CH = CH Λ
h
i
f
1,0
√ (Xp ))
Λ(Xp ) → Λ 7−→ f (Hσ=−
−1
C
FACTS
1. The period map is equivariant with respect to the actions of PΓ on F0
(via Deck transformations ↔ change of basis of projectivized frames) and on CH5
(via isometries). It is also (G y F0 )-invariant.
S
2. Let H := r∈R CH(r ⊥ ) ⊂ CH5 , where R is the set of all vectors in
Λ of norm -2. FACT: the period map maps F0 onto CH5 − H and it
maps F0 /G biholomorphically onto CH5 − H.
. – p.10/22
The Period Map F0 → CH5
5
⊗Z[√−1]
F0 −→ CH = CH Λ
h
i
f
1,0
√ (Xp ))
Λ(Xp ) → Λ 7−→ f (Hσ=−
−1
C
FACTS
1. The period map is equivariant with respect to the actions of PΓ on F0
(via Deck transformations ↔ change of basis of projectivized frames) and on CH5
(via isometries). It is also (G y F0 )-invariant.
S
2. Let H := r∈R CH(r ⊥ ) ⊂ CH5 , where R is the set of all vectors in
Λ of norm -2. FACT: the period map maps F0 onto CH5 − H and it
maps F0 /G biholomorphically onto CH5 − H.
3. We now have
M0 ↔ P0 /G
. – p.10/22
The Period Map F0 → CH5
5
⊗Z[√−1]
F0 −→ CH = CH Λ
h
i
f
1,0
√ (Xp ))
Λ(Xp ) → Λ 7−→ f (Hσ=−
−1
C
FACTS
1. The period map is equivariant with respect to the actions of PΓ on F0
(via Deck transformations ↔ change of basis of projectivized frames) and on CH5
(via isometries). It is also (G y F0 )-invariant.
S
2. Let H := r∈R CH(r ⊥ ) ⊂ CH5 , where R is the set of all vectors in
Λ of norm -2. FACT: the period map maps F0 onto CH5 − H and it
maps F0 /G biholomorphically onto CH5 − H.
3. We now have
M0 ↔ P0 /G ∼
= (PΓ\F0 ) /G
. – p.10/22
The Period Map F0 → CH5
5
⊗Z[√−1]
F0 −→ CH = CH Λ
h
i
f
1,0
√ (Xp ))
Λ(Xp ) → Λ 7−→ f (Hσ=−
−1
C
FACTS
1. The period map is equivariant with respect to the actions of PΓ on F0
(via Deck transformations ↔ change of basis of projectivized frames) and on CH5
(via isometries). It is also (G y F0 )-invariant.
S
2. Let H := r∈R CH(r ⊥ ) ⊂ CH5 , where R is the set of all vectors in
Λ of norm -2. FACT: the period map maps F0 onto CH5 − H and it
maps F0 /G biholomorphically onto CH5 − H.
3. We now have
M0 ↔ P0 /G ∼
= (PΓ\F0 ) /G ↔ PΓ\ (F0 /G)
. – p.10/22
The Period Map F0 → CH5
5
⊗Z[√−1]
F0 −→ CH = CH Λ
h
i
f
1,0
√ (Xp ))
Λ(Xp ) → Λ 7−→ f (Hσ=−
−1
C
FACTS
1. The period map is equivariant with respect to the actions of PΓ on F0
(via Deck transformations ↔ change of basis of projectivized frames) and on CH5
(via isometries). It is also (G y F0 )-invariant.
S
2. Let H := r∈R CH(r ⊥ ) ⊂ CH5 , where R is the set of all vectors in
Λ of norm -2. FACT: the period map maps F0 onto CH5 − H and it
maps F0 /G biholomorphically onto CH5 − H.
3. We now have
5
∼
∼
M0 ↔ P0 /G = (PΓ\F0 ) /G ↔ PΓ\ (F0 /G) = PΓ\ CH − H .
. – p.10/22
(Relevant Properties of the) Fox Completion Fs → Ps
. – p.11/22
(Relevant Properties of the) Fox Completion Fs → Ps
The Fox completion Fs ⊃ F0 “fills up” the gaps in F0 in such a way that:
. – p.11/22
(Relevant Properties of the) Fox Completion Fs → Ps
The Fox completion Fs ⊃ F0 “fills up” the gaps in F0 in such a way that:
1. the points of Fs − F0 “lie above” Ps − P0 , the stable but non-smooth
octics.
. – p.11/22
(Relevant Properties of the) Fox Completion Fs → Ps
The Fox completion Fs ⊃ F0 “fills up” the gaps in F0 in such a way that:
1. the points of Fs − F0 “lie above” Ps − P0 , the stable but non-smooth
octics. Elements of Fs are called framed stable forms,
. – p.11/22
(Relevant Properties of the) Fox Completion Fs → Ps
The Fox completion Fs ⊃ F0 “fills up” the gaps in F0 in such a way that:
1. the points of Fs − F0 “lie above” Ps − P0 , the stable but non-smooth
octics. Elements of Fs are called framed stable forms,
2. the actions G y F0 and PΓ y F0 extend to Fs ,
. – p.11/22
(Relevant Properties of the) Fox Completion Fs → Ps
The Fox completion Fs ⊃ F0 “fills up” the gaps in F0 in such a way that:
1. the points of Fs − F0 “lie above” Ps − P0 , the stable but non-smooth
octics. Elements of Fs are called framed stable forms,
2. the actions G y F0 and PΓ y F0 extend to Fs ,
3. the period map F0 → CH5 extends to Fs “ CH5 ,
. – p.11/22
(Relevant Properties of the) Fox Completion Fs → Ps
The Fox completion Fs ⊃ F0 “fills up” the gaps in F0 in such a way that:
1. the points of Fs − F0 “lie above” Ps − P0 , the stable but non-smooth
octics. Elements of Fs are called framed stable forms,
2. the actions G y F0 and PΓ y F0 extend to Fs ,
3. the period map F0 → CH5 extends to Fs “ CH5 , holomorphically,
. – p.11/22
(Relevant Properties of the) Fox Completion Fs → Ps
The Fox completion Fs ⊃ F0 “fills up” the gaps in F0 in such a way that:
1. the points of Fs − F0 “lie above” Ps − P0 , the stable but non-smooth
octics. Elements of Fs are called framed stable forms,
2. the actions G y F0 and PΓ y F0 extend to Fs ,
3. the period map F0 → CH5 extends to Fs “ CH5 , holomorphically,
equivariantly,
. – p.11/22
(Relevant Properties of the) Fox Completion Fs → Ps
The Fox completion Fs ⊃ F0 “fills up” the gaps in F0 in such a way that:
1. the points of Fs − F0 “lie above” Ps − P0 , the stable but non-smooth
octics. Elements of Fs are called framed stable forms,
2. the actions G y F0 and PΓ y F0 extend to Fs ,
3. the period map F0 → CH5 extends to Fs “ CH5 , holomorphically,
equivariantly, and surjectively,
. – p.11/22
(Relevant Properties of the) Fox Completion Fs → Ps
The Fox completion Fs ⊃ F0 “fills up” the gaps in F0 in such a way that:
1. the points of Fs − F0 “lie above” Ps − P0 , the stable but non-smooth
octics. Elements of Fs are called framed stable forms,
2. the actions G y F0 and PΓ y F0 extend to Fs ,
3. the period map F0 → CH5 extends to Fs “ CH5 , holomorphically,
equivariantly, and surjectively,
4. PΓ\Fs ∼
= Ps and Fs /G ∼
= CH5 .
. – p.11/22
(Relevant Properties of the) Fox Completion Fs → Ps
The Fox completion Fs ⊃ F0 “fills up” the gaps in F0 in such a way that:
1. the points of Fs − F0 “lie above” Ps − P0 , the stable but non-smooth
octics. Elements of Fs are called framed stable forms,
2. the actions G y F0 and PΓ y F0 extend to Fs ,
3. the period map F0 → CH5 extends to Fs “ CH5 , holomorphically,
equivariantly, and surjectively,
4. PΓ\Fs ∼
= Ps and Fs /G ∼
= CH5 .
The Deligne-Mostow Construction of Ms
Ms ↔ Ps /G = (PΓ\Fs ) /G ↔ PΓ\ (Fs /G) ∼
= PΓ\CH5
. – p.11/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 }
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
»
–
If Λ(Xp ) −→ Λ is a framed smooth form over p ∈ P0R ,
f
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
»
–
If Λ(Xp ) −→ Λ is a framed smooth form over p ∈ P0R , then
f
χp,f := f ◦ κp ◦ f −1 : Λ → Λ is an IAAI of Λ.
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
»
–
If Λ(Xp ) −→ Λ is a framed smooth form over p ∈ P0R , then
f
χp,f := f ◦ κp ◦ f −1 : Λ → Λ is an IAAI of Λ.
If [f1 ] and [f2 ] are framed smooth forms over the same p ∈ P0R ,
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
»
–
If Λ(Xp ) −→ Λ is a framed smooth form over p ∈ P0R , then
f
χp,f := f ◦ κp ◦ f −1 : Λ → Λ is an IAAI of Λ.
If [f1 ] and [f2 ] are framed smooth forms over the same p ∈ P0R , then χp,f1 and χp,f2
belong to the same Isom(Λ)-conjugacy class of IAAI’s of Λ.
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
»
–
If Λ(Xp ) −→ Λ is a framed smooth form over p ∈ P0R , then
f
χp,f := f ◦ κp ◦ f −1 : Λ → Λ is an IAAI of Λ.
If [f1 ] and [f2 ] are framed smooth forms over the same p ∈ P0R , then χp,f1 and χp,f2
belong to the same Isom(Λ)-conjugacy class of IAAI’s of Λ. Thus [χp ] := [χp,f ] is a
well-defined Isom(Λ)-conjugacy class, depending only on p, not on [f ].
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
»
–
If Λ(Xp ) −→ Λ is a framed smooth form over p ∈ P0R , then
f
χp,f := f ◦ κp ◦ f −1 : Λ → Λ is an IAAI of Λ.
If [f1 ] and [f2 ] are framed smooth forms over the same p ∈ P0R , then χp,f1 and χp,f2
belong to the same Isom(Λ)-conjugacy class of IAAI’s of Λ. Thus [χp ] := [χp,f ] is a
well-defined Isom(Λ)-conjugacy class, depending only on p, not on [f ].
Let (p1 , [f1 ]), (p2 , [f2 ]) be ordered pairs with p1 , p2 ∈ P0R and [f1 ], [f2 ] being framed smooth
forms over p1 , p2 respectively.
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
»
–
If Λ(Xp ) −→ Λ is a framed smooth form over p ∈ P0R , then
f
χp,f := f ◦ κp ◦ f −1 : Λ → Λ is an IAAI of Λ.
If [f1 ] and [f2 ] are framed smooth forms over the same p ∈ P0R , then χp,f1 and χp,f2
belong to the same Isom(Λ)-conjugacy class of IAAI’s of Λ. Thus [χp ] := [χp,f ] is a
well-defined Isom(Λ)-conjugacy class, depending only on p, not on [f ].
Let (p1 , [f1 ]), (p2 , [f2 ]) be ordered pairs with p1 , p2 ∈ P0R and [f1 ], [f2 ] being framed smooth
forms over p1 , p2 respectively. If p1 , p2 are of the same topological type,
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
»
–
If Λ(Xp ) −→ Λ is a framed smooth form over p ∈ P0R , then
f
χp,f := f ◦ κp ◦ f −1 : Λ → Λ is an IAAI of Λ.
If [f1 ] and [f2 ] are framed smooth forms over the same p ∈ P0R , then χp,f1 and χp,f2
belong to the same Isom(Λ)-conjugacy class of IAAI’s of Λ. Thus [χp ] := [χp,f ] is a
well-defined Isom(Λ)-conjugacy class, depending only on p, not on [f ].
Let (p1 , [f1 ]), (p2 , [f2 ]) be ordered pairs with p1 , p2 ∈ P0R and [f1 ], [f2 ] being framed smooth
forms over p1 , p2 respectively. If p1 , p2 are of the same topological type, i.e. one can be
deformed to the other via smooth real octics,
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
»
–
If Λ(Xp ) −→ Λ is a framed smooth form over p ∈ P0R , then
f
χp,f := f ◦ κp ◦ f −1 : Λ → Λ is an IAAI of Λ.
If [f1 ] and [f2 ] are framed smooth forms over the same p ∈ P0R , then χp,f1 and χp,f2
belong to the same Isom(Λ)-conjugacy class of IAAI’s of Λ. Thus [χp ] := [χp,f ] is a
well-defined Isom(Λ)-conjugacy class, depending only on p, not on [f ].
Let (p1 , [f1 ]), (p2 , [f2 ]) be ordered pairs with p1 , p2 ∈ P0R and [f1 ], [f2 ] being framed smooth
forms over p1 , p2 respectively. If p1 , p2 are of the same topological type, i.e. one can be
deformed to the other via smooth real octics, then we can deform (p1 , [f1 ]) to some
(p2 , [f2′ ]).
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
»
–
If Λ(Xp ) −→ Λ is a framed smooth form over p ∈ P0R , then
f
χp,f := f ◦ κp ◦ f −1 : Λ → Λ is an IAAI of Λ.
If [f1 ] and [f2 ] are framed smooth forms over the same p ∈ P0R , then χp,f1 and χp,f2
belong to the same Isom(Λ)-conjugacy class of IAAI’s of Λ. Thus [χp ] := [χp,f ] is a
well-defined Isom(Λ)-conjugacy class, depending only on p, not on [f ].
Let (p1 , [f1 ]), (p2 , [f2 ]) be ordered pairs with p1 , p2 ∈ P0R and [f1 ], [f2 ] being framed smooth
forms over p1 , p2 respectively. If p1 , p2 are of the same topological type, i.e. one can be
deformed to the other via smooth real octics, then we can deform (p1 , [f1 ]) to some
(p2 , [f2′ ]). Noting that IAAI(Λ) is a lattice in IAAI(Λ ⊗ C) = IAAI(C1,5 ),
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
»
–
If Λ(Xp ) −→ Λ is a framed smooth form over p ∈ P0R , then
f
χp,f := f ◦ κp ◦ f −1 : Λ → Λ is an IAAI of Λ.
If [f1 ] and [f2 ] are framed smooth forms over the same p ∈ P0R , then χp,f1 and χp,f2
belong to the same Isom(Λ)-conjugacy class of IAAI’s of Λ. Thus [χp ] := [χp,f ] is a
well-defined Isom(Λ)-conjugacy class, depending only on p, not on [f ].
Let (p1 , [f1 ]), (p2 , [f2 ]) be ordered pairs with p1 , p2 ∈ P0R and [f1 ], [f2 ] being framed smooth
forms over p1 , p2 respectively. If p1 , p2 are of the same topological type, i.e. one can be
deformed to the other via smooth real octics, then we can deform (p1 , [f1 ]) to some
(p2 , [f2′ ]). Noting that IAAI(Λ) is a lattice in IAAI(Λ ⊗ C) = IAAI(C1,5 ), we see
p1 , p2 ∈ P0R of same topological type =⇒ [χp1 ] := [χp1 ,f1 ] = [χp2 ,f ′ ] = [χp2 ,f2 ] =: [χp2 ].
2
. – p.12/22
p ∈ P0R
Involutive Anti-isometry κp of Λ(Xp )
κ
If p ∈ P0R , then complex conjugation CP1 → CP1 : x 7→ x induces an antiholomorphic
involution κp on Xp := { (x, y) ∈ C2 | y 4 − p(x) = 0 } via (x, y) 7→ ( x , y ).
κp in turn induces an involutive (antilinear) anti-isometry (IAAI) κp on Λ(Xp ).
»
–
If Λ(Xp ) −→ Λ is a framed smooth form over p ∈ P0R , then
f
χp,f := f ◦ κp ◦ f −1 : Λ → Λ is an IAAI of Λ.
If [f1 ] and [f2 ] are framed smooth forms over the same p ∈ P0R , then χp,f1 and χp,f2
belong to the same Isom(Λ)-conjugacy class of IAAI’s of Λ. Thus [χp ] := [χp,f ] is a
well-defined Isom(Λ)-conjugacy class, depending only on p, not on [f ].
Let (p1 , [f1 ]), (p2 , [f2 ]) be ordered pairs with p1 , p2 ∈ P0R and [f1 ], [f2 ] being framed smooth
forms over p1 , p2 respectively. If p1 , p2 are of the same topological type, i.e. one can be
deformed to the other via smooth real octics, then we can deform (p1 , [f1 ]) to some
(p2 , [f2′ ]). Noting that IAAI(Λ) is a lattice in IAAI(Λ ⊗ C) = IAAI(C1,5 ), we see
p1 , p2 ∈ P0R of same topological type =⇒ [χp1 ] := [χp1 ,f1 ] = [χp2 ,f ′ ] = [χp2 ,f2 ] =: [χp2 ].
2
FACT: Converse holds.
. – p.12/22
“Real” Octics Have “Real” Periods
. – p.13/22
“Real” Octics Have “Real” Periods
κp
Xp → Xp also induces an anti-linear involution on H 1 (Xp , C)
. – p.13/22
“Real” Octics Have “Real” Periods
κp
Xp → Xp also induces an anti-linear involution on H 1 (Xp , C) via
H 1 (Xp , C)
φ
−→
7−→
H 1 (Xp , C)
(κp )∗ (φ)
. – p.13/22
“Real” Octics Have “Real” Periods
κp
Xp → Xp also induces an anti-linear involution on H 1 (Xp , C) via
H 1 (Xp , C)
φ
−→
7−→
H 1 (Xp , C)
(κp )∗ (φ)
This involution preserves both Hodge decomposition
. – p.13/22
“Real” Octics Have “Real” Periods
κp
Xp → Xp also induces an anti-linear involution on H 1 (Xp , C) via
H 1 (Xp , C)
φ
−→
7−→
H 1 (Xp , C)
(κp )∗ (φ)
This involution preserves both Hodge decomposition and the σ-eigenspaces of H 1 (Xp , C).
. – p.13/22
“Real” Octics Have “Real” Periods
κp
Xp → Xp also induces an anti-linear involution on H 1 (Xp , C) via
H 1 (Xp , C)
φ
−→
7−→
H 1 (Xp , C)
(κp )∗ (φ)
This involution preserves both Hodge decomposition and the σ-eigenspaces of H 1 (Xp , C).
It turns out that κp restricts to an IAAI on
Λ(Xp )
. – p.13/22
“Real” Octics Have “Real” Periods
κp
Xp → Xp also induces an anti-linear involution on H 1 (Xp , C) via
H 1 (Xp , C)
φ
−→
7−→
H 1 (Xp , C)
(κp )∗ (φ)
This involution preserves both Hodge decomposition and the σ-eigenspaces of H 1 (Xp , C).
It turns out that κp restricts to an IAAI on
Λ(Xp ) ⊗Z[√−1] C
. – p.13/22
“Real” Octics Have “Real” Periods
κp
Xp → Xp also induces an anti-linear involution on H 1 (Xp , C) via
H 1 (Xp , C)
φ
−→
7−→
H 1 (Xp , C)
(κp )∗ (φ)
This involution preserves both Hodge decomposition and the σ-eigenspaces of H 1 (Xp , C).
It turns out that κp restricts to an IAAI on
1
√
(Xp , C)
Λ(Xp ) ⊗Z[√−1] C ∼
= Hσ=−
−1
|
{z
}
C1,5 =C1+,5−
. – p.13/22
“Real” Octics Have “Real” Periods
κp
Xp → Xp also induces an anti-linear involution on H 1 (Xp , C) via
H 1 (Xp , C)
φ
−→
7−→
H 1 (Xp , C)
(κp )∗ (φ)
This involution preserves both Hodge decomposition and the σ-eigenspaces of H 1 (Xp , C).
It turns out that κp restricts to an IAAI on
0,1
1,0
1
√
√
√
⊕
H
(X
,
C)
(Xp , C),
=
H
(X
,
C)
Λ(Xp ) ⊗Z[√−1] C ∼
= Hσ=−
p
p
−1
σ=− −1
σ=− −1
|
{z
} |
{z
} |
{z
}
C1,5 =C1+,5−
(+)
(−−−−−)
. – p.13/22
“Real” Octics Have “Real” Periods
κp
Xp → Xp also induces an anti-linear involution on H 1 (Xp , C) via
H 1 (Xp , C)
φ
−→
7−→
H 1 (Xp , C)
(κp )∗ (φ)
This involution preserves both Hodge decomposition and the σ-eigenspaces of H 1 (Xp , C).
It turns out that κp restricts to an IAAI on
0,1
1,0
1
√
√
√
⊕
H
(X
,
C)
(Xp , C),
=
H
(X
,
C)
Λ(Xp ) ⊗Z[√−1] C ∼
= Hσ=−
p
p
−1
σ=− −1
σ=− −1
|
{z
} |
{z
} |
{z
}
C1,5 =C1+,5−
(+)
(−−−−−)
thereby preserving each summands.
. – p.13/22
“Real” Octics Have “Real” Periods
κp
Xp → Xp also induces an anti-linear involution on H 1 (Xp , C) via
H 1 (Xp , C)
φ
−→
7−→
H 1 (Xp , C)
(κp )∗ (φ)
This involution preserves both Hodge decomposition and the σ-eigenspaces of H 1 (Xp , C).
It turns out that κp restricts to an IAAI on
0,1
1,0
1
√
√
√
⊕
H
(X
,
C)
(Xp , C),
=
H
(X
,
C)
Λ(Xp ) ⊗Z[√−1] C ∼
= Hσ=−
p
p
−1
σ=− −1
σ=− −1
|
{z
} |
{z
} |
{z
}
C1,5 =C1+,5−
(+)
(−−−−−)
thereby preserving each summands.
1,0
√
Thus, Hσ=−
(Xp , C) ∈ CH (Λ(Xp ) ⊗ C) is fixed by [κp ].
−1
. – p.13/22
“Real” Octics Have “Real” Periods
κp
Xp → Xp also induces an anti-linear involution on H 1 (Xp , C) via
H 1 (Xp , C)
φ
H 1 (Xp , C)
−→
(κp )∗ (φ)
7−→
This involution preserves both Hodge decomposition and the σ-eigenspaces of H 1 (Xp , C).
It turns out that κp restricts to an IAAI on
0,1
1,0
1
√
√
√
⊕
H
(X
,
C)
(Xp , C),
=
H
(X
,
C)
Λ(Xp ) ⊗Z[√−1] C ∼
= Hσ=−
p
p
−1
σ=− −1
σ=− −1
|
{z
} |
{z
} |
{z
}
C1,5 =C1+,5−
(+)
(−−−−−)
thereby preserving each summands.
1,0
√
Thus, Hσ=−
(Xp , C) ∈ CH (Λ(Xp ) ⊗ C) is fixed by [κp ].
−1
f
Hence, for a given framed smooth form [Λ(Xp ) → Λ] over p ∈ P0R , its period
1,0
5
√
f (Hσ=−
(X
,
C))
∈
CH
= CH (Λ ⊗ C) is fixed by the projective class
p
−1
[χp ] = [f ◦ κp ◦ f −1 ] ∈ PIAAI(Λ).
. – p.13/22
“Real” Octics Have “Real” Periods
κp
Xp → Xp also induces an anti-linear involution on H 1 (Xp , C) via
H 1 (Xp , C)
φ
H 1 (Xp , C)
−→
(κp )∗ (φ)
7−→
This involution preserves both Hodge decomposition and the σ-eigenspaces of H 1 (Xp , C).
It turns out that κp restricts to an IAAI on
0,1
1,0
1
√
√
√
⊕
H
(X
,
C)
(Xp , C),
=
H
(X
,
C)
Λ(Xp ) ⊗Z[√−1] C ∼
= Hσ=−
p
p
−1
σ=− −1
σ=− −1
|
{z
} |
{z
} |
{z
}
C1,5 =C1+,5−
(+)
(−−−−−)
thereby preserving each summands.
1,0
√
Thus, Hσ=−
(Xp , C) ∈ CH (Λ(Xp ) ⊗ C) is fixed by [κp ].
−1
f
Hence, for a given framed smooth form [Λ(Xp ) → Λ] over p ∈ P0R , its period
1,0
5
√
f (Hσ=−
(X
,
C))
∈
CH
= CH (Λ ⊗ C) is fixed by the projective class
p
−1
[χp ] = [f ◦ κp ◦ f −1 ] ∈ PIAAI(Λ).
We call an element x ∈ CH5 a real period if x ∈ Fix([χp ]) for some χp ∈ IAAI(Λ) arising as
described above.
. – p.13/22
Real Periods Lie on Copies of RH5 ⊂ CH5
. – p.14/22
Real Periods Lie on Copies of RH5 ⊂ CH5
1. For each χ ∈ IAAI(Λ), the metric on Λ restricts to a metric on the
Z-module Fix(χ) ∼
= Z6 of signature (1+, 5−).
. – p.14/22
Real Periods Lie on Copies of RH5 ⊂ CH5
1. For each χ ∈ IAAI(Λ), the metric on Λ restricts to a metric on the
Z-module Fix(χ) ∼
= Z6 of signature (1+, 5−). Thus
Fix(χ) ⊗Z
R ∼
= R1,5 = R1+,5− ,
. – p.14/22
Real Periods Lie on Copies of RH5 ⊂ CH5
1. For each χ ∈ IAAI(Λ), the metric on Λ restricts to a metric on the
Z-module Fix(χ) ∼
= Z6 of signature (1+, 5−). Thus
Fix(χ) ⊗Z
R ∼
= R1,5 = R1+,5− ,
and
RH (Fix(χ) ⊗Z R)
∼
= RH5
. – p.14/22
Real Periods Lie on Copies of RH5 ⊂ CH5
1. For each χ ∈ IAAI(Λ), the metric on Λ restricts to a metric on the
Z-module Fix(χ) ∼
= Z6 of signature (1+, 5−). Thus
Fix(χ) ⊗Z
R ∼
= R1,5 = R1+,5− ,
and
RH (Fix(χ) ⊗Z R)
∼
= RH5
CH Λ ⊗Z[√−1] C
∼
= CH5
∩
∩
. – p.14/22
Real Periods Lie on Copies of RH5 ⊂ CH5
1. For each χ ∈ IAAI(Λ), the metric on Λ restricts to a metric on the
Z-module Fix(χ) ∼
= Z6 of signature (1+, 5−). Thus
Fix(χ) ⊗Z
R ∼
= R1,5 = R1+,5− ,
and
RH (Fix(χ) ⊗Z R)
∼
= RH5
CH Λ ⊗Z[√−1] C
∼
= CH5
∩
∩
2. Hence, the periods of real octics lie on copies of real hyperbolic
space RH5 within CH5 .
. – p.14/22
The Allcock-Carlson-Toledo Construction of MRs
. – p.15/22
The Allcock-Carlson-Toledo Construction of MRs
R
:=
P(P
MR
s ) / PGL(2, R)
s
. – p.15/22
The Allcock-Carlson-Toledo Construction of MRs
R
R
)
/
PGL
(2,
R)
↔
P
:=
P(P
MR
s / GL(2, R)
s
s
. – p.15/22
The Allcock-Carlson-Toledo Construction of MRs
R
R
)
/
PGL
(2,
R)
↔
P
:=
P(P
MR
s / GL(2, R)
s
s
↔ PsR / ( GL(2, R)/h±1i )
. – p.15/22
The Allcock-Carlson-Toledo Construction of MRs
R
R
)
/
PGL
(2,
R)
↔
P
:=
P(P
MR
s / GL(2, R)
s
s
↔ PsR / ( GL(2, R)/h±1i ) =: PsR /GR
. – p.15/22
The Allcock-Carlson-Toledo Construction of MRs
R
R
)
/
PGL
(2,
R)
↔
P
:=
P(P
MR
s / GL(2, R)
s
s
↔ PsR / ( GL(2, R)/h±1i ) =: PsR /GR
o .
n
GR
↔
PΓ \ preimage of PsR under Fs → Ps
. – p.15/22
The Allcock-Carlson-Toledo Construction of MRs
R
R
)
/
PGL
(2,
R)
↔
P
:=
P(P
MR
s / GL(2, R)
s
s
↔ PsR / ( GL(2, R)/h±1i ) =: PsR /GR
o .
n
GR
↔
PΓ \ preimage of PsR under Fs → Ps
n
o.
=:
PΓ \ FsR
GR
. – p.15/22
The Allcock-Carlson-Toledo Construction of MRs
R
R
)
/
PGL
(2,
R)
↔
P
:=
P(P
MR
s / GL(2, R)
s
s
↔ PsR / ( GL(2, R)/h±1i ) =: PsR /GR
o .
n
GR
↔
PΓ \ preimage of PsR under Fs → Ps
n
o.
=:
PΓ \ FsR
GR
o
/n
↔ PΓ
FsR /GR
. – p.15/22
The Allcock-Carlson-Toledo Construction of MRs
R
R
)
/
PGL
(2,
R)
↔
P
:=
P(P
MR
s / GL(2, R)
s
s
↔ PsR / ( GL(2, R)/h±1i ) =: PsR /GR
o .
n
GR
↔
PΓ \ preimage of PsR under Fs → Ps
n
o.
=:
PΓ \ FsR
GR
o
/n
↔ PΓ
FsR /GR

,
- 


a
∼

RH5[χ] 
≈
= PΓ


R
[χ]∈PIAAI (Λ)
. – p.15/22
The Allcock-Carlson-Toledo Construction of MRs
R
R
)
/
PGL
(2,
R)
↔
P
:=
P(P
MR
s / GL(2, R)
s
s
↔ PsR / ( GL(2, R)/h±1i ) =: PsR /GR
o .
n
GR
↔
PΓ \ preimage of PsR under Fs → Ps
n
o.
=:
PΓ \ FsR
GR
o
/n
↔ PΓ
FsR /GR

,
- 


a
∼

RH5[χ] 
≈
= PΓ


R
[χ]∈PIAAI (Λ)
=: PΓ \ Ks
. – p.15/22
The Allcock-Carlson-Toledo Construction of MRs
R
R
)
/
PGL
(2,
R)
↔
P
:=
P(P
MR
s / GL(2, R)
s
s
↔ PsR / ( GL(2, R)/h±1i ) =: PsR /GR
o .
n
GR
↔
PΓ \ preimage of PsR under Fs → Ps
n
o.
=:
PΓ \ FsR
GR
o
/n
↔ PΓ
FsR /GR

,
- 


a
∼

RH5[χ] 
≈
= PΓ


R
[χ]∈PIAAI (Λ)
=: PΓ \ Ks =:
R
A-C-T construction of Ms
. – p.15/22
Uniformizations of MRs,i (i = 0, . . . , 4)
. – p.16/22
Uniformizations of MRs,i (i = 0, . . . , 4)
Recall: MR
s ↔ PΓ\Ks = PΓ
-(
‘
[χ]∈PIAAIR (Λ)
RH5[χ]
!,
≈
)
.
. – p.16/22
Uniformizations of MRs,i (i = 0, . . . , 4)
Recall: MR
s ↔ PΓ\Ks = PΓ
FACTS
-(
‘
[χ]∈PIAAIR (Λ)
RH5[χ]
!,
≈
)
.
1. There are either 6 (or 7) PΓ = PIsom(Λ)-conjugacy classes of IAAI’s of Λ.
. – p.16/22
Uniformizations of MRs,i (i = 0, . . . , 4)
Recall: MR
s ↔ PΓ\Ks = PΓ
FACTS
-(
‘
[χ]∈PIAAIR (Λ)
RH5[χ]
!,
≈
)
.
1. There are either 6 (or 7) PΓ = PIsom(Λ)-conjugacy classes of IAAI’s of Λ. Five of
them correspond to κCP1 ,
. – p.16/22
Uniformizations of MRs,i (i = 0, . . . , 4)
Recall: MR
s ↔ PΓ\Ks = PΓ
FACTS
-(
‘
[χ]∈PIAAIR (Λ)
RH5[χ]
!,
≈
)
.
1. There are either 6 (or 7) PΓ = PIsom(Λ)-conjugacy classes of IAAI’s of Λ. Five of
them correspond to κCP1 , and the remaining one to the antipodal map on CP1 .
. – p.16/22
Uniformizations of MRs,i (i = 0, . . . , 4)
Recall: MR
s ↔ PΓ\Ks = PΓ
FACTS
-(
‘
[χ]∈PIAAIR (Λ)
RH5[χ]
!,
≈
)
.
1. There are either 6 (or 7) PΓ = PIsom(Λ)-conjugacy classes of IAAI’s of Λ. Five of
them correspond to κCP1 , and the remaining one to the antipodal map on CP1 .
2. PΓ obviously acts transitively on the collection of the copies RH5χ , where all the χ
belong to one PΓ-conjugacy class;
. – p.16/22
Uniformizations of MRs,i (i = 0, . . . , 4)
Recall: MR
s ↔ PΓ\Ks = PΓ
FACTS
-(
‘
[χ]∈PIAAIR (Λ)
RH5[χ]
!,
≈
)
.
1. There are either 6 (or 7) PΓ = PIsom(Λ)-conjugacy classes of IAAI’s of Λ. Five of
them correspond to κCP1 , and the remaining one to the antipodal map on CP1 .
2. PΓ obviously acts transitively on the collection of the copies RH5χ , where all the χ
belong to one PΓ-conjugacy class; equivalently, the corresponding octics have the
same topological type.
. – p.16/22
Uniformizations of MRs,i (i = 0, . . . , 4)
Recall: MR
s ↔ PΓ\Ks = PΓ
-(
‘
[χ]∈PIAAIR (Λ)
FACTS
RH5[χ]
!,
≈
)
.
1. There are either 6 (or 7) PΓ = PIsom(Λ)-conjugacy classes of IAAI’s of Λ. Five of
them correspond to κCP1 , and the remaining one to the antipodal map on CP1 .
2. PΓ obviously acts transitively on the collection of the copies RH5χ , where all the χ
belong to one PΓ-conjugacy class; equivalently, the corresponding octics have the
same topological type.
It should now be clear that
∼
MR
s,i =
StabPΓ (FixΛ (χi ))
| {z }
-
RH5χi , i = 0, . . . , 4,
Z6
where RH5χi := RH (FixΛ (χi ) ⊗Z R) ∼
= RH5 ,
. – p.16/22
Uniformizations of MRs,i (i = 0, . . . , 4)
Recall: MR
s ↔ PΓ\Ks = PΓ
-(
‘
[χ]∈PIAAIR (Λ)
FACTS
RH5[χ]
!,
≈
)
.
1. There are either 6 (or 7) PΓ = PIsom(Λ)-conjugacy classes of IAAI’s of Λ. Five of
them correspond to κCP1 , and the remaining one to the antipodal map on CP1 .
2. PΓ obviously acts transitively on the collection of the copies RH5χ , where all the χ
belong to one PΓ-conjugacy class; equivalently, the corresponding octics have the
same topological type.
It should now be clear that
∼
MR
s,i =
StabPΓ (FixΛ (χi ))
| {z }
-
RH5χi , i = 0, . . . , 4,
Z6
where RH5χi := RH (FixΛ (χi ) ⊗Z R) ∼
= RH5 , and StabPΓ (FixΛ (χi )) can be described by the
following abstract isomorphism:
91
˛
08
˛
<
˛ A extends to some =
R
A
Γi := StabPΓ (FixΛ (χi )) ∼
= P @ A ∈ Isom (FixΛ (χi )) ˛˛
:
˛ element of Isom(Λ) ;
. – p.16/22
Fix(χ0 ), . . . , Fix(χ4 )
. – p.17/22
Fix(χ0 ), . . . , Fix(χ4 )
Fix(χ0 )
Fix(χ1 )
Fix(χ3 )
∼
=
∼
=
∼
=
diag(1, −1, −1, −1, −1, −1)
diag(1, −1, −1, −1, −1, −2)
diag(1, −1, −1, −2, −2, −2)
. – p.17/22
Fix(χ0 ), . . . , Fix(χ4 )
Fix(χ0 )
Fix(χ1 )
∼
=
∼
=
Fix(χ3 )
∼
=
Fix(χ2 )
∼
=
∼
=
diag(1, −1, −1, −1, −1, −1)
diag(1, −1, −1, −1, −1, −2)
diag(1, −1, −1, −2, −2, −2)
diag(1, −1, −1, −1, −2, −2)
diag(1, −1, −1, −1, −2, −2)
. – p.17/22
Fix(χ0 ), . . . , Fix(χ4 )
Fix(χ0 )
Fix(χ1 )
2
Fix(χ4 ) ∼
=4
|
0
1
3
∼
=
∼
=
Fix(χ3 )
∼
=
Fix(χ2 )
∼
=
∼
=
2
6
6
1
5⊕6
6
6
0
4
−2
0
1
1
{z
diag(1, −1, −1, −1, −1, −1)
diag(1, −1, −1, −1, −1, −2)
diag(1, −1, −1, −2, −2, −2)
diag(1, −1, −1, −1, −2, −2)
diag(1, −1, −1, −1, −2, −2)
0
1
−2
−1
1
0
−1
lattice, det=−4
−2
2
−2
6
6 −2
1
6
7
6
6 1
1 7
7
7 or 6
6
0 7
6 0
5
6
6 0
−2
4
}
−2
|
3
−2
1
0
0
3
0
0
3
−2
0
0
−2
1
−6
0
0
−4
0
0
1
0
2 −1
{z
0
det=−4
−2
3
7
−4 7
7
7
2 7
7
7
1 7
7
0 7
5
−4
}
. – p.17/22
Fix(χ0 ), . . . , Fix(χ4 )
Fix(χ0 )
Fix(χ1 )
2
Fix(χ4 ) ∼
=4
0
1
|
3
∼
=
∼
=
Fix(χ3 )
∼
=
Fix(χ2 )
∼
=
∼
=
2
6
6
1
5⊕6
6
6
0
4
−2
0
1
1
{z
diag(1, −1, −1, −1, −1, −2)
diag(1, −1, −1, −2, −2, −2)
diag(1, −1, −1, −1, −2, −2)
diag(1, −1, −1, −1, −2, −2)
0
1
−2
−1
1
0
−1
lattice, det=−4
NON-COMMENSURABILITY
diag(1, −1, −1, −1, −1, −1)
−2
2
−2
6
6 −2
1
6
7
6
6 1
1 7
7
7 or 6
6
0 7
6 0
5
6
6 0
−2
4
}
−2
|
3
−2
1
0
0
3
0
0
3
−2
0
0
−2
1
−6
0
0
−4
0
0
1
0
2 −1
{z
0
det=−4
−2
3
7
−4 7
7
7
2 7
7
7
1 7
7
0 7
5
−4
}
Isom(Fix(χ0 ))
fi Isom(Fix(χ1 ))
Isom(Fix(χ0 ))
fi Isom(Fix(χ3 ))
Isom(Fix(χ1 ))
fi Isom(Fix(χ2 ))
Isom(Fix(χ1 ))
fi Isom(Fix(χ4 ))
Isom(Fix(χ2 ))
fi Isom(Fix(χ3 ))
Isom(Fix(χ3 ))
fi Isom(Fix(χ4 ))
. – p.17/22
MRs ↔ PΓ\Ks Is Not Real Hyperbolic
. – p.18/22
MRs ↔ PΓ\Ks Is Not Real Hyperbolic
Recall again:
- 


MR
s ↔ PΓ\Ks = PΓ

a
[χ]∈PIAAIR (Λ)
,
RH5[χ] 
≈



. – p.18/22
MRs ↔ PΓ\Ks Is Not Real Hyperbolic
Recall again:
- 


MR
s ↔ PΓ\Ks = PΓ

a
[χ]∈PIAAIR (Λ)
,
RH5[χ] 
≈



We check the local quotient structure of PΓ\Ks stratum by stratum.
. – p.18/22
MRs ↔ PΓ\Ks Is Not Real Hyperbolic
Recall again:
- 


MR
s ↔ PΓ\Ks = PΓ

a
[χ]∈PIAAIR (Λ)
,
RH5[χ] 
≈



We check the local quotient structure of PΓ\Ks stratum by stratum.
NEGATIVE RESULT
Points in PΓ\Ks corresponding to the stratum ∆0,1
R (of real octics having one
(real) triple point and no other singularities) can not admit a local real hyperbolic
orbifold structure.
. – p.18/22
MRs ↔ PΓ\Ks Is Not Real Hyperbolic
Recall again:
- 


MR
s ↔ PΓ\Ks = PΓ

a
[χ]∈PIAAIR (Λ)
,
RH5[χ] 
≈



We check the local quotient structure of PΓ\Ks stratum by stratum.
NEGATIVE RESULT
Points in PΓ\Ks corresponding to the stratum ∆0,1
R (of real octics having one
(real) triple point and no other singularities) can not admit a local real hyperbolic
orbifold structure. Hence PΓ\Ks itself cannot be a real hyperbolic orbifold.
. – p.18/22
Why Points in ∆R0,1 Are Not Hyerpbolic
. – p.19/22
Why Points in ∆R0,1 Are Not Hyerpbolic
A point in ∆0,1
R can be locally described by p0,0 (x), where
pa0 ,a1 (x) = (x3 + a1 x + a0 ) · r(x), a0 , a1 ∈ R.
. – p.19/22
Why Points in ∆R0,1 Are Not Hyerpbolic
A point in ∆0,1
R can be locally described by p0,0 (x), where
pa0 ,a1 (x) = (x3 + a1 x + a0 ) · r(x), a0 , a1 ∈ R.
We thus examine the vanishing (σ 2 = −1)-homology of
y 4 = x3 + a1 x + a0 , as a0 , a1 → 0,
. – p.19/22
Why Points in ∆R0,1 Are Not Hyerpbolic
A point in ∆0,1
R can be locally described by p0,0 (x), where
pa0 ,a1 (x) = (x3 + a1 x + a0 ) · r(x), a0 , a1 ∈ R.
We thus examine the vanishing (σ 2 = −1)-homology of
y 4 = x3 + a1 x + a0 , as a0 , a1 → 0,
preserved by the action induced by x 7→ x.
. – p.19/22
Why Points in ∆R0,1 Are Not Hyerpbolic
A point in ∆0,1
R can be locally described by p0,0 (x), where
pa0 ,a1 (x) = (x3 + a1 x + a0 ) · r(x), a0 , a1 ∈ R.
We thus examine the vanishing (σ 2 = −1)-homology of
y 4 = x3 + a1 x + a0 , as a0 , a1 → 0,
preserved by the action induced by x 7→ x.
2
Computations show that Λ0 = 4
IAAI’s,
say χ1 and χ2 .
−2
√
1 − −1
1+
√
−1
−2
3
5 has two conjugacy classes of
. – p.19/22
Why Points in ∆R0,1 Are Not Hyerpbolic
A point in ∆0,1
R can be locally described by p0,0 (x), where
pa0 ,a1 (x) = (x3 + a1 x + a0 ) · r(x), a0 , a1 ∈ R.
We thus examine the vanishing (σ 2 = −1)-homology of
y 4 = x3 + a1 x + a0 , as a0 , a1 → 0,
preserved by the action induced by x 7→ x.
2
Computations show that Λ0 = 4
−2
√
1 − −1
1+
√
−1
−2
3
5 has two conjugacy classes of
say χ1 and χ2 . We expect this since a real triple point is the limit of two kinds of
smooth real 3-point configurations, namely 3 distinct real points, and one real point plus 1
complex conjugate pair.
IAAI’s,
. – p.19/22
Why Points in ∆R0,1 Are Not Hyerpbolic
A point in ∆0,1
R can be locally described by p0,0 (x), where
pa0 ,a1 (x) = (x3 + a1 x + a0 ) · r(x), a0 , a1 ∈ R.
We thus examine the vanishing (σ 2 = −1)-homology of
y 4 = x3 + a1 x + a0 , as a0 , a1 → 0,
preserved by the action induced by x 7→ x.
2
Computations show that Λ0 = 4
−2
√
1 − −1
1+
√
−1
−2
3
5 has two conjugacy classes of
say χ1 and χ2 . We expect this since a real triple point is the limit of two kinds of
smooth real 3-point configurations, namely 3 distinct real points, and one real point plus 1
complex conjugate pair. So, the local geometry of PΓ\Ks at a point in ∆0,1
R is given by
IAAI’s,
Fix(χ1 )
StabIsom(Λ0 ) (Fix(χ1 ))
[
Fix(χ2 )
StabIsom(Λ0 ) (Fix(χ2 ))
, subject to certain gluing.
. – p.19/22
Why Points in ∆R0,1 Are Not Hyperpbolic (Cont’d)
. – p.20/22
Why Points in ∆R0,1 Are Not Hyperpbolic (Cont’d)
Local quotient at a point in ∆0,1
R is given by:
Fix(χ1 )
StabIsom(Λ0 ) (Fix(χ1 ))
[
Fix(χ2 )
StabIsom(Λ0 ) (Fix(χ2 ))
,
. – p.20/22
Why Points in ∆R0,1 Are Not Hyperpbolic (Cont’d)
Local quotient at a point in ∆0,1
R is given by:
Fix(χ1 )
StabIsom(Λ0 ) (Fix(χ1 ))
[
Fix(χ2 )
StabIsom(Λ0 ) (Fix(χ2 ))
,
1. The two individual quotients above are
R2 /(Z/2 × Z/2) = a 90◦ -wedge,
and
R2 /D4 = a 45◦ -wedge.
. – p.20/22
Why Points in ∆R0,1 Are Not Hyperpbolic (Cont’d)
Local quotient at a point in ∆0,1
R is given by:
Fix(χ1 )
StabIsom(Λ0 ) (Fix(χ1 ))
[
Fix(χ2 )
StabIsom(Λ0 ) (Fix(χ2 ))
,
1. The two individual quotients above are
R2 /(Z/2 × Z/2) = a 90◦ -wedge,
and
R2 /D4 = a 45◦ -wedge.
2. The edges of the above wedges glue “pairwise.”
. – p.20/22
Why Points in ∆R0,1 Are Not Hyperpbolic (Cont’d)
Local quotient at a point in ∆0,1
R is given by:
Fix(χ1 )
StabIsom(Λ0 ) (Fix(χ1 ))
[
Fix(χ2 )
StabIsom(Λ0 ) (Fix(χ2 ))
,
1. The two individual quotients above are
R2 /(Z/2 × Z/2) = a 90◦ -wedge,
and
R2 /D4 = a 45◦ -wedge.
2. The edges of the above wedges glue “pairwise.” =⇒ local angle is
135◦ = 3π/4.
. – p.20/22
Why Points in ∆R0,1 Are Not Hyperpbolic (Cont’d)
Local quotient at a point in ∆0,1
R is given by:
Fix(χ1 )
StabIsom(Λ0 ) (Fix(χ1 ))
[
Fix(χ2 )
StabIsom(Λ0 ) (Fix(χ2 ))
,
1. The two individual quotients above are
R2 /(Z/2 × Z/2) = a 90◦ -wedge,
and
R2 /D4 = a 45◦ -wedge.
2. The edges of the above wedges glue “pairwise.” =⇒ local angle is
135◦ = 3π/4.
OBSERVATION: Points in ∆0,1
R can NOT be real hyperbolic because the
local anlge does not add up to 2π/n, for some integer n > 0.
. – p.20/22
Ongoing Work & Future Directions ...
. – p.21/22
Ongoing Work & Future Directions ...
1. Prove that two representatives of IAAI’s of Λ whose fixed lattice are
isometric to Fix(χ2 ) are in fact conjugates.
. – p.21/22
Ongoing Work & Future Directions ...
1. Prove that two representatives of IAAI’s of Λ whose fixed lattice are
isometric to Fix(χ2 ) are in fact conjugates.
2. Identify PΓR
4.
. – p.21/22
Ongoing Work & Future Directions ...
1. Prove that two representatives of IAAI’s of Λ whose fixed lattice are
isometric to Fix(χ2 ) are in fact conjugates.
2. Identify PΓR
4.
3. Study the topology of MR
0,i : fundamental and higher homotopy
groups.
. – p.21/22
Ongoing Work & Future Directions ...
1. Prove that two representatives of IAAI’s of Λ whose fixed lattice are
isometric to Fix(χ2 ) are in fact conjugates.
2. Identify PΓR
4.
3. Study the topology of MR
0,i : fundamental and higher homotopy
groups.
4. Complete the examination of the local geometry of PΓ\Ks .
. – p.21/22
Ongoing Work & Future Directions ...
1. Prove that two representatives of IAAI’s of Λ whose fixed lattice are
isometric to Fix(χ2 ) are in fact conjugates.
2. Identify PΓR
4.
3. Study the topology of MR
0,i : fundamental and higher homotopy
groups.
4. Complete the examination of the local geometry of PΓ\Ks .
We do know that Ks is (obviously) a metric space and PΓ acts on it by isometries,
. – p.21/22
Ongoing Work & Future Directions ...
1. Prove that two representatives of IAAI’s of Λ whose fixed lattice are
isometric to Fix(χ2 ) are in fact conjugates.
2. Identify PΓR
4.
3. Study the topology of MR
0,i : fundamental and higher homotopy
groups.
4. Complete the examination of the local geometry of PΓ\Ks .
We do know that Ks is (obviously) a metric space and PΓ acts on it by isometries,
properly discontinuously,
. – p.21/22
Ongoing Work & Future Directions ...
1. Prove that two representatives of IAAI’s of Λ whose fixed lattice are
isometric to Fix(χ2 ) are in fact conjugates.
2. Identify PΓR
4.
3. Study the topology of MR
0,i : fundamental and higher homotopy
groups.
4. Complete the examination of the local geometry of PΓ\Ks .
We do know that Ks is (obviously) a metric space and PΓ acts on it by isometries,
properly discontinuously, hence with closed orbits.
. – p.21/22
Ongoing Work & Future Directions ...
1. Prove that two representatives of IAAI’s of Λ whose fixed lattice are
isometric to Fix(χ2 ) are in fact conjugates.
2. Identify PΓR
4.
3. Study the topology of MR
0,i : fundamental and higher homotopy
groups.
4. Complete the examination of the local geometry of PΓ\Ks .
We do know that Ks is (obviously) a metric space and PΓ acts on it by isometries,
properly discontinuously, hence with closed orbits. PΓ\Ks is thus itself a metric space.
. – p.21/22
Ongoing Work & Future Directions ...
1. Prove that two representatives of IAAI’s of Λ whose fixed lattice are
isometric to Fix(χ2 ) are in fact conjugates.
2. Identify PΓR
4.
3. Study the topology of MR
0,i : fundamental and higher homotopy
groups.
4. Complete the examination of the local geometry of PΓ\Ks .
We do know that Ks is (obviously) a metric space and PΓ acts on it by isometries,
properly discontinuously, hence with closed orbits. PΓ\Ks is thus itself a metric space.
Speculation:
PΓ\Ks is some kind of an orbit space by a negatively curved, non-locally-symmetric space. These were
once conjectured not to exist. But Mostow-Siu [1980] first constructed such compact Käher (complex)
surface (hence of real dimension 4). Gromov-Thurston [1987] constructed examples of any real
dimension ≥
4.
. – p.21/22
THE END
THANK YOU!
. – p.22/22
Download