Setting Work in Q p ( p -adics) Two-sorted language: Q p with - - PowerPoint PPT Presentation

setting
SMART_READER_LITE
LIVE PREVIEW

Setting Work in Q p ( p -adics) Two-sorted language: Q p with - - PowerPoint PPT Presentation

Trees of definable sets in Z p I. Halupczok Ecole Normale Superieure rue dUlm, Paris Barcelona Modnet conference 2008 Trees of definable sets in Z p 1 / 13 I. Halupczok Setting Work in Q p ( p -adics) Two-sorted language: Q p


slide-1
SLIDE 1

Trees of definable sets in Zp

  • I. Halupczok

Ecole Normale Superieure rue d’Ulm, Paris

Barcelona Modnet conference 2008

Trees of definable sets in Zp 1 / 13

  • I. Halupczok
slide-2
SLIDE 2

Setting

  • Work in Qp (p-adics)
  • Two-sorted language:
  • Qp with field language
  • value group Z with ordered group language
  • valuation v : Qp → Z ∪ {∞}
  • For tuples ¯

x ⊂ Qn

p, set v(¯

x) := min{v(xi) | 1 ≤ i ≤ n}

  • Ball around ¯

x ∈ Qn

p of (valuative) radius λ ∈ Z:

B(¯ x, λ) := {¯ x′ | v(¯ x′ − ¯ x) ≥ λ} = ¯ x + pλZn

p

  • Recall: Balls of fixed radius λ form a partition of Qn

p

Trees of definable sets in Zp 2 / 13

  • I. Halupczok
slide-3
SLIDE 3

Setting

  • Work in Qp (p-adics)
  • Two-sorted language:
  • Qp with field language
  • value group Z with ordered group language
  • valuation v : Qp → Z ∪ {∞}
  • For tuples ¯

x ⊂ Qn

p, set v(¯

x) := min{v(xi) | 1 ≤ i ≤ n}

  • Ball around ¯

x ∈ Qn

p of (valuative) radius λ ∈ Z:

B(¯ x, λ) := {¯ x′ | v(¯ x′ − ¯ x) ≥ λ} = ¯ x + pλZn

p

  • Recall: Balls of fixed radius λ form a partition of Qn

p

Trees of definable sets in Zp 2 / 13

  • I. Halupczok
slide-4
SLIDE 4

Setting

  • Work in Qp (p-adics)
  • Two-sorted language:
  • Qp with field language
  • value group Z with ordered group language
  • valuation v : Qp → Z ∪ {∞}
  • For tuples ¯

x ⊂ Qn

p, set v(¯

x) := min{v(xi) | 1 ≤ i ≤ n}

  • Ball around ¯

x ∈ Qn

p of (valuative) radius λ ∈ Z:

B(¯ x, λ) := {¯ x′ | v(¯ x′ − ¯ x) ≥ λ} = ¯ x + pλZn

p

  • Recall: Balls of fixed radius λ form a partition of Qn

p

Trees of definable sets in Zp 2 / 13

  • I. Halupczok
slide-5
SLIDE 5

Setting

  • Work in Qp (p-adics)
  • Two-sorted language:
  • Qp with field language
  • value group Z with ordered group language
  • valuation v : Qp → Z ∪ {∞}
  • For tuples ¯

x ⊂ Qn

p, set v(¯

x) := min{v(xi) | 1 ≤ i ≤ n}

  • Ball around ¯

x ∈ Qn

p of (valuative) radius λ ∈ Z:

B(¯ x, λ) := {¯ x′ | v(¯ x′ − ¯ x) ≥ λ} = ¯ x + pλZn

p

  • Recall: Balls of fixed radius λ form a partition of Qn

p

Trees of definable sets in Zp 2 / 13

  • I. Halupczok
slide-6
SLIDE 6

Setting

  • Work in Qp (p-adics)
  • Two-sorted language:
  • Qp with field language
  • value group Z with ordered group language
  • valuation v : Qp → Z ∪ {∞}
  • For tuples ¯

x ⊂ Qn

p, set v(¯

x) := min{v(xi) | 1 ≤ i ≤ n}

  • Ball around ¯

x ∈ Qn

p of (valuative) radius λ ∈ Z:

B(¯ x, λ) := {¯ x′ | v(¯ x′ − ¯ x) ≥ λ} = ¯ x + pλZn

p

  • Recall: Balls of fixed radius λ form a partition of Qn

p

Trees of definable sets in Zp 2 / 13

  • I. Halupczok
slide-7
SLIDE 7

Trees of sets in Zn

p

  • Subballs of Zn

p form a tree

  • Suppose ∅ = X ⊂ Zn
  • p. This yields a sub-tree T(X) of those

balls intersecting X:

  • Vertices: T(X) = {B = B(¯

x, λ) | ¯ x ∈ X, λ ≥ 0}

  • Root of T(X) is Zp
  • Tree structure given by inclusion
  • B ⊂ Zn

p ball, B ∩ X = ∅

TB(X) := subtree of T(X) above B (i.e. vertices are balls contained in B)

Trees of definable sets in Zp 3 / 13

  • I. Halupczok
slide-8
SLIDE 8

Trees of sets in Zn

p

  • Subballs of Zn

p form a tree

  • Suppose ∅ = X ⊂ Zn
  • p. This yields a sub-tree T(X) of those

balls intersecting X:

  • Vertices: T(X) = {B = B(¯

x, λ) | ¯ x ∈ X, λ ≥ 0}

  • Root of T(X) is Zp
  • Tree structure given by inclusion
  • B ⊂ Zn

p ball, B ∩ X = ∅

TB(X) := subtree of T(X) above B (i.e. vertices are balls contained in B)

Trees of definable sets in Zp 3 / 13

  • I. Halupczok
slide-9
SLIDE 9

Trees of sets in Zn

p

  • Subballs of Zn

p form a tree

  • Suppose ∅ = X ⊂ Zn
  • p. This yields a sub-tree T(X) of those

balls intersecting X:

  • Vertices: T(X) = {B = B(¯

x, λ) | ¯ x ∈ X, λ ≥ 0}

  • Root of T(X) is Zp
  • Tree structure given by inclusion
  • B ⊂ Zn

p ball, B ∩ X = ∅

TB(X) := subtree of T(X) above B (i.e. vertices are balls contained in B)

Trees of definable sets in Zp 3 / 13

  • I. Halupczok
slide-10
SLIDE 10

Examples of trees

  • Examples:
  • X = Zn

p Every node of T(X) has pn children

  • X finite each x ∈ X corresponds to infinite path in T(X). . .
  • Infinite path in a tree T is a sequence of vertices

{v0, v1, v2, . . . } ⊂ T with v0 = root and vi+1 is child of vi

  • We have bijection

{Infinite paths in T(X)}

1:1

← → ¯ X (= p-adic closure of X)

Trees of definable sets in Zp 4 / 13

  • I. Halupczok
slide-11
SLIDE 11

Examples of trees

  • Examples:
  • X = Zn

p Every node of T(X) has pn children

  • X finite each x ∈ X corresponds to infinite path in T(X). . .
  • Infinite path in a tree T is a sequence of vertices

{v0, v1, v2, . . . } ⊂ T with v0 = root and vi+1 is child of vi

  • We have bijection

{Infinite paths in T(X)}

1:1

← → ¯ X (= p-adic closure of X)

Trees of definable sets in Zp 4 / 13

  • I. Halupczok
slide-12
SLIDE 12

Examples of trees

  • Examples:
  • X = Zn

p Every node of T(X) has pn children

  • X finite each x ∈ X corresponds to infinite path in T(X). . .
  • Infinite path in a tree T is a sequence of vertices

{v0, v1, v2, . . . } ⊂ T with v0 = root and vi+1 is child of vi

  • We have bijection

{Infinite paths in T(X)}

1:1

← → ¯ X (= p-adic closure of X)

Trees of definable sets in Zp 4 / 13

  • I. Halupczok
slide-13
SLIDE 13

Examples of trees

  • Examples:
  • X = Zn

p Every node of T(X) has pn children

  • X finite each x ∈ X corresponds to infinite path in T(X). . .
  • Infinite path in a tree T is a sequence of vertices

{v0, v1, v2, . . . } ⊂ T with v0 = root and vi+1 is child of vi

  • We have bijection

{Infinite paths in T(X)}

1:1

← → ¯ X (= p-adic closure of X)

Trees of definable sets in Zp 4 / 13

  • I. Halupczok
slide-14
SLIDE 14

Examples of trees

  • Examples:
  • X = Zn

p Every node of T(X) has pn children

  • X finite each x ∈ X corresponds to infinite path in T(X). . .
  • Infinite path in a tree T is a sequence of vertices

{v0, v1, v2, . . . } ⊂ T with v0 = root and vi+1 is child of vi

  • We have bijection

{Infinite paths in T(X)}

1:1

← → ¯ X (= p-adic closure of X)

Trees of definable sets in Zp 4 / 13

  • I. Halupczok
slide-15
SLIDE 15

Goal

Question

For which abstract trees T does there exist a definable X such that T ∼ = T(X)?

  • Obvious condition: T has no leaves.
  • Less obvious condition:
  • Suppose T ∼

= T(X); fix infinite path P ⊂ T .

  • This yields x ∈ ¯

X.

  • Define: Side branch of P := subtree of T starting at a vertex
  • f P, without P itself.
  • P has a side branch at depth λ ⇐

⇒ there exists x′ ∈ X such that v(x′ − x) = λ

  • Thus: set of depths of side branches is definable subset of Z
  • Goal: Find combinatorial description of the set of (abstract)

trees T(X) for X definable.

Trees of definable sets in Zp 5 / 13

  • I. Halupczok
slide-16
SLIDE 16

Goal

Question

For which abstract trees T does there exist a definable X such that T ∼ = T(X)?

  • Obvious condition: T has no leaves.
  • Less obvious condition:
  • Suppose T ∼

= T(X); fix infinite path P ⊂ T .

  • This yields x ∈ ¯

X.

  • Define: Side branch of P := subtree of T starting at a vertex
  • f P, without P itself.
  • P has a side branch at depth λ ⇐

⇒ there exists x′ ∈ X such that v(x′ − x) = λ

  • Thus: set of depths of side branches is definable subset of Z
  • Goal: Find combinatorial description of the set of (abstract)

trees T(X) for X definable.

Trees of definable sets in Zp 5 / 13

  • I. Halupczok
slide-17
SLIDE 17

Goal

Question

For which abstract trees T does there exist a definable X such that T ∼ = T(X)?

  • Obvious condition: T has no leaves.
  • Less obvious condition:
  • Suppose T ∼

= T(X); fix infinite path P ⊂ T .

  • This yields x ∈ ¯

X.

  • Define: Side branch of P := subtree of T starting at a vertex
  • f P, without P itself.
  • P has a side branch at depth λ ⇐

⇒ there exists x′ ∈ X such that v(x′ − x) = λ

  • Thus: set of depths of side branches is definable subset of Z
  • Goal: Find combinatorial description of the set of (abstract)

trees T(X) for X definable.

Trees of definable sets in Zp 5 / 13

  • I. Halupczok
slide-18
SLIDE 18

Goal

Question

For which abstract trees T does there exist a definable X such that T ∼ = T(X)?

  • Obvious condition: T has no leaves.
  • Less obvious condition:
  • Suppose T ∼

= T(X); fix infinite path P ⊂ T .

  • This yields x ∈ ¯

X.

  • Define: Side branch of P := subtree of T starting at a vertex
  • f P, without P itself.
  • P has a side branch at depth λ ⇐

⇒ there exists x′ ∈ X such that v(x′ − x) = λ

  • Thus: set of depths of side branches is definable subset of Z
  • Goal: Find combinatorial description of the set of (abstract)

trees T(X) for X definable.

Trees of definable sets in Zp 5 / 13

  • I. Halupczok
slide-19
SLIDE 19

Goal

Question

For which abstract trees T does there exist a definable X such that T ∼ = T(X)?

  • Obvious condition: T has no leaves.
  • Less obvious condition:
  • Suppose T ∼

= T(X); fix infinite path P ⊂ T .

  • This yields x ∈ ¯

X.

  • Define: Side branch of P := subtree of T starting at a vertex
  • f P, without P itself.
  • P has a side branch at depth λ ⇐

⇒ there exists x′ ∈ X such that v(x′ − x) = λ

  • Thus: set of depths of side branches is definable subset of Z
  • Goal: Find combinatorial description of the set of (abstract)

trees T(X) for X definable.

Trees of definable sets in Zp 5 / 13

  • I. Halupczok
slide-20
SLIDE 20

Goal

Question

For which abstract trees T does there exist a definable X such that T ∼ = T(X)?

  • Obvious condition: T has no leaves.
  • Less obvious condition:
  • Suppose T ∼

= T(X); fix infinite path P ⊂ T .

  • This yields x ∈ ¯

X.

  • Define: Side branch of P := subtree of T starting at a vertex
  • f P, without P itself.
  • P has a side branch at depth λ ⇐

⇒ there exists x′ ∈ X such that v(x′ − x) = λ

  • Thus: set of depths of side branches is definable subset of Z
  • Goal: Find combinatorial description of the set of (abstract)

trees T(X) for X definable.

Trees of definable sets in Zp 5 / 13

  • I. Halupczok
slide-21
SLIDE 21

Goal

Question

For which abstract trees T does there exist a definable X such that T ∼ = T(X)?

  • Obvious condition: T has no leaves.
  • Less obvious condition:
  • Suppose T ∼

= T(X); fix infinite path P ⊂ T .

  • This yields x ∈ ¯

X.

  • Define: Side branch of P := subtree of T starting at a vertex
  • f P, without P itself.
  • P has a side branch at depth λ ⇐

⇒ there exists x′ ∈ X such that v(x′ − x) = λ

  • Thus: set of depths of side branches is definable subset of Z
  • Goal: Find combinatorial description of the set of (abstract)

trees T(X) for X definable.

Trees of definable sets in Zp 5 / 13

  • I. Halupczok
slide-22
SLIDE 22

Goal

Question

For which abstract trees T does there exist a definable X such that T ∼ = T(X)?

  • Obvious condition: T has no leaves.
  • Less obvious condition:
  • Suppose T ∼

= T(X); fix infinite path P ⊂ T .

  • This yields x ∈ ¯

X.

  • Define: Side branch of P := subtree of T starting at a vertex
  • f P, without P itself.
  • P has a side branch at depth λ ⇐

⇒ there exists x′ ∈ X such that v(x′ − x) = λ

  • Thus: set of depths of side branches is definable subset of Z
  • Goal: Find combinatorial description of the set of (abstract)

trees T(X) for X definable.

Trees of definable sets in Zp 5 / 13

  • I. Halupczok
slide-23
SLIDE 23

Goal

Question

For which abstract trees T does there exist a definable X such that T ∼ = T(X)?

  • Obvious condition: T has no leaves.
  • Less obvious condition:
  • Suppose T ∼

= T(X); fix infinite path P ⊂ T .

  • This yields x ∈ ¯

X.

  • Define: Side branch of P := subtree of T starting at a vertex
  • f P, without P itself.
  • P has a side branch at depth λ ⇐

⇒ there exists x′ ∈ X such that v(x′ − x) = λ

  • Thus: set of depths of side branches is definable subset of Z
  • Goal: Find combinatorial description of the set of (abstract)

trees T(X) for X definable.

Trees of definable sets in Zp 5 / 13

  • I. Halupczok
slide-24
SLIDE 24

Motivation: isometry

  • T(X) = T(¯

X); so from now on suppose X = ¯ X

Lemma

X ⊂ Zn

p, X ′ ⊂ Zn′ p Then:

{bijective isometries X → X ′}

1:1

← → {T(X)

− → T(X ′)} Sketch of proof:

  • φ: X → X ′ yields T(X) → T(X ′), B(¯

x, λ) → B(φ(¯ x), λ) for ¯ x ∈ X. Well-defined as φ isometry.

  • ψ: T(X) → T(X ′) induces map on infinite paths. . .

Analogously: {bij. isomet. X ∩ B → X ′ ∩ B}

1:1

↔ {TB(X)

− → TB(X ′)}

Trees of definable sets in Zp 6 / 13

  • I. Halupczok
slide-25
SLIDE 25

Motivation: isometry

  • T(X) = T(¯

X); so from now on suppose X = ¯ X

Lemma

X ⊂ Zn

p, X ′ ⊂ Zn′ p Then:

{bijective isometries X → X ′}

1:1

← → {T(X)

− → T(X ′)} Sketch of proof:

  • φ: X → X ′ yields T(X) → T(X ′), B(¯

x, λ) → B(φ(¯ x), λ) for ¯ x ∈ X. Well-defined as φ isometry.

  • ψ: T(X) → T(X ′) induces map on infinite paths. . .

Analogously: {bij. isomet. X ∩ B → X ′ ∩ B}

1:1

↔ {TB(X)

− → TB(X ′)}

Trees of definable sets in Zp 6 / 13

  • I. Halupczok
slide-26
SLIDE 26

Motivation: isometry

  • T(X) = T(¯

X); so from now on suppose X = ¯ X

Lemma

X ⊂ Zn

p, X ′ ⊂ Zn′ p Then:

{bijective isometries X → X ′}

1:1

← → {T(X)

− → T(X ′)} Sketch of proof:

  • φ: X → X ′ yields T(X) → T(X ′), B(¯

x, λ) → B(φ(¯ x), λ) for ¯ x ∈ X. Well-defined as φ isometry.

  • ψ: T(X) → T(X ′) induces map on infinite paths. . .

Analogously: {bij. isomet. X ∩ B → X ′ ∩ B}

1:1

↔ {TB(X)

− → TB(X ′)}

Trees of definable sets in Zp 6 / 13

  • I. Halupczok
slide-27
SLIDE 27

Motivation: isometry

  • T(X) = T(¯

X); so from now on suppose X = ¯ X

Lemma

X ⊂ Zn

p, X ′ ⊂ Zn′ p Then:

{bijective isometries X → X ′}

1:1

← → {T(X)

− → T(X ′)} Sketch of proof:

  • φ: X → X ′ yields T(X) → T(X ′), B(¯

x, λ) → B(φ(¯ x), λ) for ¯ x ∈ X. Well-defined as φ isometry.

  • ψ: T(X) → T(X ′) induces map on infinite paths. . .

Analogously: {bij. isomet. X ∩ B → X ′ ∩ B}

1:1

↔ {TB(X)

− → TB(X ′)}

Trees of definable sets in Zp 6 / 13

  • I. Halupczok
slide-28
SLIDE 28

Motivation: isometry

  • T(X) = T(¯

X); so from now on suppose X = ¯ X

Lemma

X ⊂ Zn

p, X ′ ⊂ Zn′ p Then:

{bijective isometries X → X ′}

1:1

← → {T(X)

− → T(X ′)} Sketch of proof:

  • φ: X → X ′ yields T(X) → T(X ′), B(¯

x, λ) → B(φ(¯ x), λ) for ¯ x ∈ X. Well-defined as φ isometry.

  • ψ: T(X) → T(X ′) induces map on infinite paths. . .

Analogously: {bij. isomet. X ∩ B → X ′ ∩ B}

1:1

↔ {TB(X)

− → TB(X ′)}

Trees of definable sets in Zp 6 / 13

  • I. Halupczok
slide-29
SLIDE 29

Motivation: isometry

  • T(X) = T(¯

X); so from now on suppose X = ¯ X

Lemma

X ⊂ Zn

p, X ′ ⊂ Zn′ p Then:

{bijective isometries X → X ′}

1:1

← → {T(X)

− → T(X ′)} Sketch of proof:

  • φ: X → X ′ yields T(X) → T(X ′), B(¯

x, λ) → B(φ(¯ x), λ) for ¯ x ∈ X. Well-defined as φ isometry.

  • ψ: T(X) → T(X ′) induces map on infinite paths. . .

Analogously: {bij. isomet. X ∩ B → X ′ ∩ B}

1:1

↔ {TB(X)

− → TB(X ′)}

Trees of definable sets in Zp 6 / 13

  • I. Halupczok
slide-30
SLIDE 30

Motivation: isometry

  • T(X) = T(¯

X); so from now on suppose X = ¯ X

Lemma

X ⊂ Zn

p, X ′ ⊂ Zn′ p Then:

{bijective isometries X → X ′}

1:1

← → {T(X)

− → T(X ′)} Sketch of proof:

  • φ: X → X ′ yields T(X) → T(X ′), B(¯

x, λ) → B(φ(¯ x), λ) for ¯ x ∈ X. Well-defined as φ isometry.

  • ψ: T(X) → T(X ′) induces map on infinite paths. . .

Analogously: {bij. isomet. X ∩ B → X ′ ∩ B}

1:1

↔ {TB(X)

− → TB(X ′)}

Trees of definable sets in Zp 6 / 13

  • I. Halupczok
slide-31
SLIDE 31

The conjecture

  • For X ⊂ Qn

p definable, Scowcroft and van den Dries defined:

dim X := dimension of Zariski closure of X in ˜ Qn

p.

  • We will define trees of level d (purely combinatorial).

Conjecture (H.)

Suppose T is an (abstract) tree. Then: T is of level d ⇐ ⇒ there exists definable X ⊂ Zn

p with

dim X = d such that T ∼ = T(X)

Theorem (H.)

“= ⇒” holds. “⇐ =” holds if X ⊂ Z2

p

  • r if dim X ≤ 1
  • r if X is algebraic without singularities.

In this talk: consider mainly “⇐ =”.

Trees of definable sets in Zp 7 / 13

  • I. Halupczok
slide-32
SLIDE 32

The conjecture

  • For X ⊂ Qn

p definable, Scowcroft and van den Dries defined:

dim X := dimension of Zariski closure of X in ˜ Qn

p.

  • We will define trees of level d (purely combinatorial).

Conjecture (H.)

Suppose T is an (abstract) tree. Then: T is of level d ⇐ ⇒ there exists definable X ⊂ Zn

p with

dim X = d such that T ∼ = T(X)

Theorem (H.)

“= ⇒” holds. “⇐ =” holds if X ⊂ Z2

p

  • r if dim X ≤ 1
  • r if X is algebraic without singularities.

In this talk: consider mainly “⇐ =”.

Trees of definable sets in Zp 7 / 13

  • I. Halupczok
slide-33
SLIDE 33

The conjecture

  • For X ⊂ Qn

p definable, Scowcroft and van den Dries defined:

dim X := dimension of Zariski closure of X in ˜ Qn

p.

  • We will define trees of level d (purely combinatorial).

Conjecture (H.)

Suppose T is an (abstract) tree. Then: T is of level d ⇐ ⇒ there exists definable X ⊂ Zn

p with

dim X = d such that T ∼ = T(X)

Theorem (H.)

“= ⇒” holds. “⇐ =” holds if X ⊂ Z2

p

  • r if dim X ≤ 1
  • r if X is algebraic without singularities.

In this talk: consider mainly “⇐ =”.

Trees of definable sets in Zp 7 / 13

  • I. Halupczok
slide-34
SLIDE 34

The conjecture

  • For X ⊂ Qn

p definable, Scowcroft and van den Dries defined:

dim X := dimension of Zariski closure of X in ˜ Qn

p.

  • We will define trees of level d (purely combinatorial).

Conjecture (H.)

Suppose T is an (abstract) tree. Then: T is of level d ⇐ ⇒ there exists definable X ⊂ Zn

p with

dim X = d such that T ∼ = T(X)

Theorem (H.)

“= ⇒” holds. “⇐ =” holds if X ⊂ Z2

p

  • r if dim X ≤ 1
  • r if X is algebraic without singularities.

In this talk: consider mainly “⇐ =”.

Trees of definable sets in Zp 7 / 13

  • I. Halupczok
slide-35
SLIDE 35

The conjecture

  • For X ⊂ Qn

p definable, Scowcroft and van den Dries defined:

dim X := dimension of Zariski closure of X in ˜ Qn

p.

  • We will define trees of level d (purely combinatorial).

Conjecture (H.)

Suppose T is an (abstract) tree. Then: T is of level d ⇐ ⇒ there exists definable X ⊂ Zn

p with

dim X = d such that T ∼ = T(X)

Theorem (H.)

“= ⇒” holds. “⇐ =” holds if X ⊂ Z2

p

  • r if dim X ≤ 1
  • r if X is algebraic without singularities.

In this talk: consider mainly “⇐ =”.

Trees of definable sets in Zp 7 / 13

  • I. Halupczok
slide-36
SLIDE 36

The conjecture

  • For X ⊂ Qn

p definable, Scowcroft and van den Dries defined:

dim X := dimension of Zariski closure of X in ˜ Qn

p.

  • We will define trees of level d (purely combinatorial).

Conjecture (H.)

Suppose T is an (abstract) tree. Then: T is of level d ⇐ ⇒ there exists definable X ⊂ Zn

p with

dim X = d such that T ∼ = T(X)

Theorem (H.)

“= ⇒” holds. “⇐ =” holds if X ⊂ Z2

p

  • r if dim X ≤ 1
  • r if X is algebraic without singularities.

In this talk: consider mainly “⇐ =”.

Trees of definable sets in Zp 7 / 13

  • I. Halupczok
slide-37
SLIDE 37

Definition of level 0 trees

  • Definition of level d trees is recursive. Start with level 0:
  • Define: T is of level 0 ⇐

⇒ T has no leaves and only finitely many bifurcations.

  • These are exactly the trees of finite sets.

⇒ Conjecture holds for d = 0:

  • For d ≥ 1 we will define when T is of level ≤ d
  • Then: T is of level = d if it is of level ≤ d but not of level

≤ d − 1

Trees of definable sets in Zp 8 / 13

  • I. Halupczok
slide-38
SLIDE 38

Definition of level 0 trees

  • Definition of level d trees is recursive. Start with level 0:
  • Define: T is of level 0 ⇐

⇒ T has no leaves and only finitely many bifurcations.

  • These are exactly the trees of finite sets.

⇒ Conjecture holds for d = 0:

  • For d ≥ 1 we will define when T is of level ≤ d
  • Then: T is of level = d if it is of level ≤ d but not of level

≤ d − 1

Trees of definable sets in Zp 8 / 13

  • I. Halupczok
slide-39
SLIDE 39

Definition of level 0 trees

  • Definition of level d trees is recursive. Start with level 0:
  • Define: T is of level 0 ⇐

⇒ T has no leaves and only finitely many bifurcations.

  • These are exactly the trees of finite sets.

⇒ Conjecture holds for d = 0:

  • For d ≥ 1 we will define when T is of level ≤ d
  • Then: T is of level = d if it is of level ≤ d but not of level

≤ d − 1

Trees of definable sets in Zp 8 / 13

  • I. Halupczok
slide-40
SLIDE 40

Definition of level 0 trees

  • Definition of level d trees is recursive. Start with level 0:
  • Define: T is of level 0 ⇐

⇒ T has no leaves and only finitely many bifurcations.

  • These are exactly the trees of finite sets.

⇒ Conjecture holds for d = 0:

  • For d ≥ 1 we will define when T is of level ≤ d
  • Then: T is of level = d if it is of level ≤ d but not of level

≤ d − 1

Trees of definable sets in Zp 8 / 13

  • I. Halupczok
slide-41
SLIDE 41

Definition of level 0 trees

  • Definition of level d trees is recursive. Start with level 0:
  • Define: T is of level 0 ⇐

⇒ T has no leaves and only finitely many bifurcations.

  • These are exactly the trees of finite sets.

⇒ Conjecture holds for d = 0:

  • For d ≥ 1 we will define when T is of level ≤ d
  • Then: T is of level = d if it is of level ≤ d but not of level

≤ d − 1

Trees of definable sets in Zp 8 / 13

  • I. Halupczok
slide-42
SLIDE 42

Definition of level d trees, d ≥ 1 (1)

Fix d ≥ 1. Define: T of level ≤ d ⇐ ⇒ There exists S0 finite set of infinite paths P ⊂ T such that:

1 For each depth λ ∈ N:

  • Cut each path P ∈ S0 at depth λ
  • The remaining tree has to satisfy:
  • It consists of a finite tree F with trees T1, . . . , Tk attached to

it.

  • Each Ti is of the form T ′

i × T(Zp), with T ′ i of level ≤ d − 1.

2 Uniformity condition when walking up a path P ∈ S0

[On next slide]

Examples: X = Zp 1 X = Zn

p

n X = {x ∈ Zp | v(x) even} 1

Trees of definable sets in Zp 9 / 13

  • I. Halupczok
slide-43
SLIDE 43

Definition of level d trees, d ≥ 1 (1)

Fix d ≥ 1. Define: T of level ≤ d ⇐ ⇒ There exists S0 finite set of infinite paths P ⊂ T such that:

1 For each depth λ ∈ N:

  • Cut each path P ∈ S0 at depth λ
  • The remaining tree has to satisfy:
  • It consists of a finite tree F with trees T1, . . . , Tk attached to

it.

  • Each Ti is of the form T ′

i × T(Zp), with T ′ i of level ≤ d − 1.

2 Uniformity condition when walking up a path P ∈ S0

[On next slide]

Examples: X = Zp 1 X = Zn

p

n X = {x ∈ Zp | v(x) even} 1

Trees of definable sets in Zp 9 / 13

  • I. Halupczok
slide-44
SLIDE 44

Definition of level d trees, d ≥ 1 (1)

Fix d ≥ 1. Define: T of level ≤ d ⇐ ⇒ There exists S0 finite set of infinite paths P ⊂ T such that:

1 For each depth λ ∈ N:

  • Cut each path P ∈ S0 at depth λ
  • The remaining tree has to satisfy:
  • It consists of a finite tree F with trees T1, . . . , Tk attached to

it.

  • Each Ti is of the form T ′

i × T(Zp), with T ′ i of level ≤ d − 1.

2 Uniformity condition when walking up a path P ∈ S0

[On next slide]

Examples: X = Zp 1 X = Zn

p

n X = {x ∈ Zp | v(x) even} 1

Trees of definable sets in Zp 9 / 13

  • I. Halupczok
slide-45
SLIDE 45

Definition of level d trees, d ≥ 1 (1)

Fix d ≥ 1. Define: T of level ≤ d ⇐ ⇒ There exists S0 finite set of infinite paths P ⊂ T such that:

1 For each depth λ ∈ N:

  • Cut each path P ∈ S0 at depth λ
  • The remaining tree has to satisfy:
  • It consists of a finite tree F with trees T1, . . . , Tk attached to

it.

  • Each Ti is of the form T ′

i × T(Zp), with T ′ i of level ≤ d − 1.

2 Uniformity condition when walking up a path P ∈ S0

[On next slide]

Examples: X = Zp 1 X = Zn

p

n X = {x ∈ Zp | v(x) even} 1

Trees of definable sets in Zp 9 / 13

  • I. Halupczok
slide-46
SLIDE 46

Definition of level d trees, d ≥ 1 (1)

Fix d ≥ 1. Define: T of level ≤ d ⇐ ⇒ There exists S0 finite set of infinite paths P ⊂ T such that:

1 For each depth λ ∈ N:

  • Cut each path P ∈ S0 at depth λ
  • The remaining tree has to satisfy:
  • It consists of a finite tree F with trees T1, . . . , Tk attached to

it.

  • Each Ti is of the form T ′

i × T(Zp), with T ′ i of level ≤ d − 1.

2 Uniformity condition when walking up a path P ∈ S0

[On next slide]

Examples: X = Zp 1 X = Zn

p

n X = {x ∈ Zp | v(x) even} 1

Trees of definable sets in Zp 9 / 13

  • I. Halupczok
slide-47
SLIDE 47

Definition of level d trees, d ≥ 1 (1)

Fix d ≥ 1. Define: T of level ≤ d ⇐ ⇒ There exists S0 finite set of infinite paths P ⊂ T such that:

1 For each depth λ ∈ N:

  • Cut each path P ∈ S0 at depth λ
  • The remaining tree has to satisfy:
  • It consists of a finite tree F with trees T1, . . . , Tk attached to

it.

  • Each Ti is of the form T ′

i × T(Zp), with T ′ i of level ≤ d − 1.

2 Uniformity condition when walking up a path P ∈ S0

[On next slide]

Examples: X = Zp 1 X = Zn

p

n X = {x ∈ Zp | v(x) even} 1

Trees of definable sets in Zp 9 / 13

  • I. Halupczok
slide-48
SLIDE 48

Definition of level d trees, d ≥ 1 (1)

Fix d ≥ 1. Define: T of level ≤ d ⇐ ⇒ There exists S0 finite set of infinite paths P ⊂ T such that:

1 For each depth λ ∈ N:

  • Cut each path P ∈ S0 at depth λ
  • The remaining tree has to satisfy:
  • It consists of a finite tree F with trees T1, . . . , Tk attached to

it.

  • Each Ti is of the form T ′

i × T(Zp), with T ′ i of level ≤ d − 1.

2 Uniformity condition when walking up a path P ∈ S0

[On next slide]

Examples: X = Zp 1 X = Zn

p

n X = {x ∈ Zp | v(x) even} 1

Trees of definable sets in Zp 9 / 13

  • I. Halupczok
slide-49
SLIDE 49

Definition of level d trees, d ≥ 1 (1)

Fix d ≥ 1. Define: T of level ≤ d ⇐ ⇒ There exists S0 finite set of infinite paths P ⊂ T such that:

1 For each depth λ ∈ N:

  • Cut each path P ∈ S0 at depth λ
  • The remaining tree has to satisfy:
  • It consists of a finite tree F with trees T1, . . . , Tk attached to

it.

  • Each Ti is of the form T ′

i × T(Zp), with T ′ i of level ≤ d − 1.

2 Uniformity condition when walking up a path P ∈ S0

[On next slide]

Examples: X = Zp 1 X = Zn

p

n X = {x ∈ Zp | v(x) even} 1

Trees of definable sets in Zp 9 / 13

  • I. Halupczok
slide-50
SLIDE 50

Definition of level d trees, d ≥ 1 (1)

Fix d ≥ 1. Define: T of level ≤ d ⇐ ⇒ There exists S0 finite set of infinite paths P ⊂ T such that:

1 For each depth λ ∈ N:

  • Cut each path P ∈ S0 at depth λ
  • The remaining tree has to satisfy:
  • It consists of a finite tree F with trees T1, . . . , Tk attached to

it.

  • Each Ti is of the form T ′

i × T(Zp), with T ′ i of level ≤ d − 1.

2 Uniformity condition when walking up a path P ∈ S0

[On next slide]

Examples: X = Zp 1 X = Zn

p

n X = {x ∈ Zp | v(x) even} 1

Trees of definable sets in Zp 9 / 13

  • I. Halupczok
slide-51
SLIDE 51

Definition of level d trees, d ≥ 1 (1)

Fix d ≥ 1. Define: T of level ≤ d ⇐ ⇒ There exists S0 finite set of infinite paths P ⊂ T such that:

1 For each depth λ ∈ N:

  • Cut each path P ∈ S0 at depth λ
  • The remaining tree has to satisfy:
  • It consists of a finite tree F with trees T1, . . . , Tk attached to

it.

  • Each Ti is of the form T ′

i × T(Zp), with T ′ i of level ≤ d − 1.

2 Uniformity condition when walking up a path P ∈ S0

[On next slide]

Examples: X = Zp 1 X = Zn

p

n X = {x ∈ Zp | v(x) even} 1

Trees of definable sets in Zp 9 / 13

  • I. Halupczok
slide-52
SLIDE 52

Definition of level d trees, d ≥ 1 (1)

Fix d ≥ 1. Define: T of level ≤ d ⇐ ⇒ There exists S0 finite set of infinite paths P ⊂ T such that:

1 For each depth λ ∈ N:

  • Cut each path P ∈ S0 at depth λ
  • The remaining tree has to satisfy:
  • It consists of a finite tree F with trees T1, . . . , Tk attached to

it.

  • Each Ti is of the form T ′

i × T(Zp), with T ′ i of level ≤ d − 1.

2 Uniformity condition when walking up a path P ∈ S0

[On next slide]

Examples: X = Zp 1 X = Zn

p

n X = {x ∈ Zp | v(x) even} 1

Trees of definable sets in Zp 9 / 13

  • I. Halupczok
slide-53
SLIDE 53

Definition of level d trees, d ≥ 1 (1)

Fix d ≥ 1. Define: T of level ≤ d ⇐ ⇒ There exists S0 finite set of infinite paths P ⊂ T such that:

1 For each depth λ ∈ N:

  • Cut each path P ∈ S0 at depth λ
  • The remaining tree has to satisfy:
  • It consists of a finite tree F with trees T1, . . . , Tk attached to

it.

  • Each Ti is of the form T ′

i × T(Zp), with T ′ i of level ≤ d − 1.

2 Uniformity condition when walking up a path P ∈ S0

[On next slide]

Examples: X = Zp 1 X = Zn

p

n X = {x ∈ Zp | v(x) even} 1

Trees of definable sets in Zp 9 / 13

  • I. Halupczok
slide-54
SLIDE 54

Definition of level d trees, d ≥ 1 (2)

Uniformity condition when walking up a path P ∈ S0:

2 For each path P ∈ S0:

  • Let Bµ be the side branch of P at depth µ.
  • There are λ, ρ ∈ N such that:
  • Consider all Bµ with µ ≥ λ, µ ≡ a mod ρ for some fixed a.
  • For these:
  • The finite tree at the beginning of Bµ is the same for all µ
  • The trees the T ′

i of level d − 1 appearing inside Bµ are

“uniformly (in µ) of level d − 1”

Trees of definable sets in Zp 10 / 13

  • I. Halupczok
slide-55
SLIDE 55

Definition of level d trees, d ≥ 1 (2)

Uniformity condition when walking up a path P ∈ S0:

2 For each path P ∈ S0:

  • Let Bµ be the side branch of P at depth µ.
  • There are λ, ρ ∈ N such that:
  • Consider all Bµ with µ ≥ λ, µ ≡ a mod ρ for some fixed a.
  • For these:
  • The finite tree at the beginning of Bµ is the same for all µ
  • The trees the T ′

i of level d − 1 appearing inside Bµ are

“uniformly (in µ) of level d − 1”

Trees of definable sets in Zp 10 / 13

  • I. Halupczok
slide-56
SLIDE 56

Definition of level d trees, d ≥ 1 (2)

Uniformity condition when walking up a path P ∈ S0:

2 For each path P ∈ S0:

  • Let Bµ be the side branch of P at depth µ.
  • There are λ, ρ ∈ N such that:
  • Consider all Bµ with µ ≥ λ, µ ≡ a mod ρ for some fixed a.
  • For these:
  • The finite tree at the beginning of Bµ is the same for all µ
  • The trees the T ′

i of level d − 1 appearing inside Bµ are

“uniformly (in µ) of level d − 1”

Trees of definable sets in Zp 10 / 13

  • I. Halupczok
slide-57
SLIDE 57

Definition of level d trees, d ≥ 1 (2)

Uniformity condition when walking up a path P ∈ S0:

2 For each path P ∈ S0:

  • Let Bµ be the side branch of P at depth µ.
  • There are λ, ρ ∈ N such that:
  • Consider all Bµ with µ ≥ λ, µ ≡ a mod ρ for some fixed a.
  • For these:
  • The finite tree at the beginning of Bµ is the same for all µ
  • The trees the T ′

i of level d − 1 appearing inside Bµ are

“uniformly (in µ) of level d − 1”

Trees of definable sets in Zp 10 / 13

  • I. Halupczok
slide-58
SLIDE 58

Definition of level d trees, d ≥ 1 (2)

Uniformity condition when walking up a path P ∈ S0:

2 For each path P ∈ S0:

  • Let Bµ be the side branch of P at depth µ.
  • There are λ, ρ ∈ N such that:
  • Consider all Bµ with µ ≥ λ, µ ≡ a mod ρ for some fixed a.
  • For these:
  • The finite tree at the beginning of Bµ is the same for all µ
  • The trees the T ′

i of level d − 1 appearing inside Bµ are

“uniformly (in µ) of level d − 1”

Trees of definable sets in Zp 10 / 13

  • I. Halupczok
slide-59
SLIDE 59

Definition of level d trees, d ≥ 1 (2)

Uniformity condition when walking up a path P ∈ S0:

2 For each path P ∈ S0:

  • Let Bµ be the side branch of P at depth µ.
  • There are λ, ρ ∈ N such that:
  • Consider all Bµ with µ ≥ λ, µ ≡ a mod ρ for some fixed a.
  • For these:
  • The finite tree at the beginning of Bµ is the same for all µ
  • The trees the T ′

i of level d − 1 appearing inside Bµ are

“uniformly (in µ) of level d − 1”

Trees of definable sets in Zp 10 / 13

  • I. Halupczok
slide-60
SLIDE 60

Meaning of 1 for the definable set

What does “T(X) satisfies 1” mean for X?

Lemma

X ⊂ Zn

  • p. (Recall: X closed!) Then T(X) satisfies 1 ⇐

⇒ There is a finite X0 ⊂ X such that: Around each ¯ x ∈ X \ X0 there exists ball B¯

x

such that TB¯

x(X) ∼

= T ′ × T(Zp), where T ′ of level ≤ d − 1 Idea of proof:

  • X0 = set of points corresponding to paths S0
  • “=

⇒”: easy.

  • “⇐

=”: use that X is compact.

Trees of definable sets in Zp 11 / 13

  • I. Halupczok
slide-61
SLIDE 61

Meaning of 1 for the definable set

What does “T(X) satisfies 1” mean for X?

Lemma

X ⊂ Zn

  • p. (Recall: X closed!) Then T(X) satisfies 1 ⇐

⇒ There is a finite X0 ⊂ X such that: Around each ¯ x ∈ X \ X0 there exists ball B¯

x

such that TB¯

x(X) ∼

= T ′ × T(Zp), where T ′ of level ≤ d − 1 Idea of proof:

  • X0 = set of points corresponding to paths S0
  • “=

⇒”: easy.

  • “⇐

=”: use that X is compact.

Trees of definable sets in Zp 11 / 13

  • I. Halupczok
slide-62
SLIDE 62

Meaning of 1 for the definable set

What does “T(X) satisfies 1” mean for X?

Lemma

X ⊂ Zn

  • p. (Recall: X closed!) Then T(X) satisfies 1 ⇐

⇒ There is a finite X0 ⊂ X such that: Around each ¯ x ∈ X \ X0 there exists ball B¯

x

such that TB¯

x(X) ∼

= T ′ × T(Zp), where T ′ of level ≤ d − 1 Idea of proof:

  • X0 = set of points corresponding to paths S0
  • “=

⇒”: easy.

  • “⇐

=”: use that X is compact.

Trees of definable sets in Zp 11 / 13

  • I. Halupczok
slide-63
SLIDE 63

Meaning of 1 for the definable set

What does “T(X) satisfies 1” mean for X?

Lemma

X ⊂ Zn

  • p. (Recall: X closed!) Then T(X) satisfies 1 ⇐

⇒ There is a finite X0 ⊂ X such that: Around each ¯ x ∈ X \ X0 there exists ball B¯

x

such that TB¯

x(X) ∼

= T ′ × T(Zp), where T ′ of level ≤ d − 1 Idea of proof:

  • X0 = set of points corresponding to paths S0
  • “=

⇒”: easy.

  • “⇐

=”: use that X is compact.

Trees of definable sets in Zp 11 / 13

  • I. Halupczok
slide-64
SLIDE 64

Meaning of 1 for the definable set

What does “T(X) satisfies 1” mean for X?

Lemma

X ⊂ Zn

  • p. (Recall: X closed!) Then T(X) satisfies 1 ⇐

⇒ There is a finite X0 ⊂ X such that: Around each ¯ x ∈ X \ X0 there exists ball B¯

x

such that TB¯

x(X) ∼

= T ′ × T(Zp), where T ′ of level ≤ d − 1 Idea of proof:

  • X0 = set of points corresponding to paths S0
  • “=

⇒”: easy.

  • “⇐

=”: use that X is compact.

Trees of definable sets in Zp 11 / 13

  • I. Halupczok
slide-65
SLIDE 65

Meaning of 1 for the definable set

What does “T(X) satisfies 1” mean for X?

Lemma

X ⊂ Zn

  • p. (Recall: X closed!) Then T(X) satisfies 1 ⇐

⇒ There is a finite X0 ⊂ X such that: Around each ¯ x ∈ X \ X0 there exists ball B¯

x

such that TB¯

x(X) ∼

= T ′ × T(Zp), where T ′ of level ≤ d − 1 Idea of proof:

  • X0 = set of points corresponding to paths S0
  • “=

⇒”: easy.

  • “⇐

=”: use that X is compact.

Trees of definable sets in Zp 11 / 13

  • I. Halupczok
slide-66
SLIDE 66

Meaning of 1 for the definable set

What does “T(X) satisfies 1” mean for X?

Lemma

X ⊂ Zn

  • p. (Recall: X closed!) Then T(X) satisfies 1 ⇐

⇒ There is a finite X0 ⊂ X such that: Around each ¯ x ∈ X \ X0 there exists ball B¯

x

such that TB¯

x(X) ∼

= T ′ × T(Zp), where T ′ of level ≤ d − 1 Idea of proof:

  • X0 = set of points corresponding to paths S0
  • “=

⇒”: easy.

  • “⇐

=”: use that X is compact.

Trees of definable sets in Zp 11 / 13

  • I. Halupczok
slide-67
SLIDE 67

Meaning of 1 for the definable set

What does “T(X) satisfies 1” mean for X?

Lemma

X ⊂ Zn

  • p. (Recall: X closed!) Then T(X) satisfies 1 ⇐

⇒ There is a finite X0 ⊂ X such that: Around each ¯ x ∈ X \ X0 there exists ball B¯

x

such that TB¯

x(X) ∼

= T ′ × T(Zp), where T ′ of level ≤ d − 1 Idea of proof:

  • X0 = set of points corresponding to paths S0
  • “=

⇒”: easy.

  • “⇐

=”: use that X is compact.

Trees of definable sets in Zp 11 / 13

  • I. Halupczok
slide-68
SLIDE 68

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-69
SLIDE 69

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-70
SLIDE 70

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-71
SLIDE 71

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-72
SLIDE 72

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-73
SLIDE 73

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-74
SLIDE 74

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-75
SLIDE 75

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-76
SLIDE 76

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-77
SLIDE 77

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-78
SLIDE 78

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-79
SLIDE 79

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-80
SLIDE 80

Proof of conjecture for X smooth algebraic

We prove that T(X) is of level ≤ dim X in some special cases.

  • X = {(x, φ(x)) | x ∈ Zp} with v(φ(x1)−φ(x2)) ≥ v(x1 −x2)

(*)

  • Then Zp → X, x → (x, φ(x)) is isometry.
  • By isometry lemma, T(X) ∼

= T(Zp)

  • ⇒ T(X) is of level 1.
  • X ⊂ Z2

p is smooth algebraic curve:

  • It suffices to find ball B around each (x, y) ∈ X such that

TB(X) ∼ = T(Zp)

  • Fix (x, y) ∈ X.
  • Implicit function theorem X ∩ B is graph of function φ
  • After possibly exchanging coordinates, φ satisfies (*)
  • Then as above.
  • Higher dimensions work similarly.
  • If X has only isolated singularities (e.g. any algebraic curve),

then this proves 1. (But 2 is difficult.)

Trees of definable sets in Zp 12 / 13

  • I. Halupczok
slide-81
SLIDE 81

Conjecture yields Stratifications

Suppose the conjecture holds. Then definable sets X can be stratified, i.e. decomposed into nice subsets:

  • X0 := set corresponding to paths S0 in T(X)

X0 ⊂ X is set of “0-dimensional singularities”

  • X \ X0 can be partitioned into X ∩ B for balls B, such that

TB(X) ∼ = T ′ × T(Zp) with T ′ of level ≤ d − 1

  • Apply conjecture to T ′; the paths S′

0 of T ′ yield

1-dimensional subset of X ∩ B X1 := union of these subsets X1 ⊂ X \ X0 is set of “1-dimensional singularities”

  • Inductively obtain X = X0 ˙

∪ . . . ˙ ∪ Xd This is a stratification:

  • Xi is locally isometric to Zi

p near each ¯

x ∈ Xi

  • ¯

Xi = X0 ∪ · · · ∪ Xi

  • 2 yields precise description of Xi near ¯

x ∈ ¯ Xi.

Trees of definable sets in Zp 13 / 13

  • I. Halupczok
slide-82
SLIDE 82

Conjecture yields Stratifications

Suppose the conjecture holds. Then definable sets X can be stratified, i.e. decomposed into nice subsets:

  • X0 := set corresponding to paths S0 in T(X)

X0 ⊂ X is set of “0-dimensional singularities”

  • X \ X0 can be partitioned into X ∩ B for balls B, such that

TB(X) ∼ = T ′ × T(Zp) with T ′ of level ≤ d − 1

  • Apply conjecture to T ′; the paths S′

0 of T ′ yield

1-dimensional subset of X ∩ B X1 := union of these subsets X1 ⊂ X \ X0 is set of “1-dimensional singularities”

  • Inductively obtain X = X0 ˙

∪ . . . ˙ ∪ Xd This is a stratification:

  • Xi is locally isometric to Zi

p near each ¯

x ∈ Xi

  • ¯

Xi = X0 ∪ · · · ∪ Xi

  • 2 yields precise description of Xi near ¯

x ∈ ¯ Xi.

Trees of definable sets in Zp 13 / 13

  • I. Halupczok
slide-83
SLIDE 83

Conjecture yields Stratifications

Suppose the conjecture holds. Then definable sets X can be stratified, i.e. decomposed into nice subsets:

  • X0 := set corresponding to paths S0 in T(X)

X0 ⊂ X is set of “0-dimensional singularities”

  • X \ X0 can be partitioned into X ∩ B for balls B, such that

TB(X) ∼ = T ′ × T(Zp) with T ′ of level ≤ d − 1

  • Apply conjecture to T ′; the paths S′

0 of T ′ yield

1-dimensional subset of X ∩ B X1 := union of these subsets X1 ⊂ X \ X0 is set of “1-dimensional singularities”

  • Inductively obtain X = X0 ˙

∪ . . . ˙ ∪ Xd This is a stratification:

  • Xi is locally isometric to Zi

p near each ¯

x ∈ Xi

  • ¯

Xi = X0 ∪ · · · ∪ Xi

  • 2 yields precise description of Xi near ¯

x ∈ ¯ Xi.

Trees of definable sets in Zp 13 / 13

  • I. Halupczok
slide-84
SLIDE 84

Conjecture yields Stratifications

Suppose the conjecture holds. Then definable sets X can be stratified, i.e. decomposed into nice subsets:

  • X0 := set corresponding to paths S0 in T(X)

X0 ⊂ X is set of “0-dimensional singularities”

  • X \ X0 can be partitioned into X ∩ B for balls B, such that

TB(X) ∼ = T ′ × T(Zp) with T ′ of level ≤ d − 1

  • Apply conjecture to T ′; the paths S′

0 of T ′ yield

1-dimensional subset of X ∩ B X1 := union of these subsets X1 ⊂ X \ X0 is set of “1-dimensional singularities”

  • Inductively obtain X = X0 ˙

∪ . . . ˙ ∪ Xd This is a stratification:

  • Xi is locally isometric to Zi

p near each ¯

x ∈ Xi

  • ¯

Xi = X0 ∪ · · · ∪ Xi

  • 2 yields precise description of Xi near ¯

x ∈ ¯ Xi.

Trees of definable sets in Zp 13 / 13

  • I. Halupczok
slide-85
SLIDE 85

Conjecture yields Stratifications

Suppose the conjecture holds. Then definable sets X can be stratified, i.e. decomposed into nice subsets:

  • X0 := set corresponding to paths S0 in T(X)

X0 ⊂ X is set of “0-dimensional singularities”

  • X \ X0 can be partitioned into X ∩ B for balls B, such that

TB(X) ∼ = T ′ × T(Zp) with T ′ of level ≤ d − 1

  • Apply conjecture to T ′; the paths S′

0 of T ′ yield

1-dimensional subset of X ∩ B X1 := union of these subsets X1 ⊂ X \ X0 is set of “1-dimensional singularities”

  • Inductively obtain X = X0 ˙

∪ . . . ˙ ∪ Xd This is a stratification:

  • Xi is locally isometric to Zi

p near each ¯

x ∈ Xi

  • ¯

Xi = X0 ∪ · · · ∪ Xi

  • 2 yields precise description of Xi near ¯

x ∈ ¯ Xi.

Trees of definable sets in Zp 13 / 13

  • I. Halupczok
slide-86
SLIDE 86

Conjecture yields Stratifications

Suppose the conjecture holds. Then definable sets X can be stratified, i.e. decomposed into nice subsets:

  • X0 := set corresponding to paths S0 in T(X)

X0 ⊂ X is set of “0-dimensional singularities”

  • X \ X0 can be partitioned into X ∩ B for balls B, such that

TB(X) ∼ = T ′ × T(Zp) with T ′ of level ≤ d − 1

  • Apply conjecture to T ′; the paths S′

0 of T ′ yield

1-dimensional subset of X ∩ B X1 := union of these subsets X1 ⊂ X \ X0 is set of “1-dimensional singularities”

  • Inductively obtain X = X0 ˙

∪ . . . ˙ ∪ Xd This is a stratification:

  • Xi is locally isometric to Zi

p near each ¯

x ∈ Xi

  • ¯

Xi = X0 ∪ · · · ∪ Xi

  • 2 yields precise description of Xi near ¯

x ∈ ¯ Xi.

Trees of definable sets in Zp 13 / 13

  • I. Halupczok
slide-87
SLIDE 87

Conjecture yields Stratifications

Suppose the conjecture holds. Then definable sets X can be stratified, i.e. decomposed into nice subsets:

  • X0 := set corresponding to paths S0 in T(X)

X0 ⊂ X is set of “0-dimensional singularities”

  • X \ X0 can be partitioned into X ∩ B for balls B, such that

TB(X) ∼ = T ′ × T(Zp) with T ′ of level ≤ d − 1

  • Apply conjecture to T ′; the paths S′

0 of T ′ yield

1-dimensional subset of X ∩ B X1 := union of these subsets X1 ⊂ X \ X0 is set of “1-dimensional singularities”

  • Inductively obtain X = X0 ˙

∪ . . . ˙ ∪ Xd This is a stratification:

  • Xi is locally isometric to Zi

p near each ¯

x ∈ Xi

  • ¯

Xi = X0 ∪ · · · ∪ Xi

  • 2 yields precise description of Xi near ¯

x ∈ ¯ Xi.

Trees of definable sets in Zp 13 / 13

  • I. Halupczok
slide-88
SLIDE 88

Conjecture yields Stratifications

Suppose the conjecture holds. Then definable sets X can be stratified, i.e. decomposed into nice subsets:

  • X0 := set corresponding to paths S0 in T(X)

X0 ⊂ X is set of “0-dimensional singularities”

  • X \ X0 can be partitioned into X ∩ B for balls B, such that

TB(X) ∼ = T ′ × T(Zp) with T ′ of level ≤ d − 1

  • Apply conjecture to T ′; the paths S′

0 of T ′ yield

1-dimensional subset of X ∩ B X1 := union of these subsets X1 ⊂ X \ X0 is set of “1-dimensional singularities”

  • Inductively obtain X = X0 ˙

∪ . . . ˙ ∪ Xd This is a stratification:

  • Xi is locally isometric to Zi

p near each ¯

x ∈ Xi

  • ¯

Xi = X0 ∪ · · · ∪ Xi

  • 2 yields precise description of Xi near ¯

x ∈ ¯ Xi.

Trees of definable sets in Zp 13 / 13

  • I. Halupczok