andrés villaveces universidad nacional de colombia - bogotáindependence notions continuous...

70
Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad Independence, Continuity, Tree Properties Andrés Villaveces Universidad Nacional de Colombia - Bogotá Universitat de Barcelona - Model Theory Seminar - October 2016

Upload: others

Post on 06-Jun-2021

3 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Independence, Continuity, Tree Properties

Andrés Villaveces

Universidad Nacional de Colombia - Bogotá

Universitat de Barcelona - Model Theory Seminar - October 2016

Page 2: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Contents

Independence notions

A review of canonicity of independence notions

Two recent results

Continuous world/independence

Continuous Logic / Stability / Lindström

Continuous Independence

Obstacles to continuous Chatzidakis-Pillay

The Chatzidakis-Pillay Theorem - Simplicity

Continuous Generic Predicates over Hilbert Spaces

Continuity carries Interference

Generic independence, good and bad

The theory TN is bad: it has TP2

The theory TN is good: it is NTP1 (really, NSOP1)

Page 3: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Independence notions

Let us consider the following independence notion:

Denition (coheir independence (v. Boney-Grossberg))

Let M ≺ N .

Ach|M

N

m

∀A− ⊂∗ A ∀N− ≺∗ N

∃B− ⊂ M so that B− ≡N− A−.

(By ⊂∗, ≺∗ I mean “small”...)

Page 4: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Independence notions

Let us consider the following independence notion:

Denition (coheir independence (v. Boney-Grossberg))

Let M ≺ N .

Ach|M

N

m

∀A− ⊂∗ A ∀N− ≺∗ N

∃B− ⊂ M so that B− ≡N− A−.

(By ⊂∗, ≺∗ I mean “small”...)

Page 5: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Independence notions

Let us consider the following independence notion:

Denition (coheir independence (v. Boney-Grossberg))

Let M ≺ N .

Ach|M

N

m

∀A− ⊂∗ A ∀N− ≺∗ N

∃B− ⊂ M so that B− ≡N− A−.

(By ⊂∗, ≺∗ I mean “small”...)

Page 6: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Independence notions

Let us consider the following independence notion:

Denition (coheir independence (v. Boney-Grossberg))

Let M ≺ N .

Ach|M

N

m

∀A− ⊂∗ A ∀N− ≺∗ N

∃B− ⊂ M so that B− ≡N− A−.

(By ⊂∗, ≺∗ I mean “small”...)

Page 7: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Coheir

I The previous was an example of an “abstract” independence

notion. In rst order, these presentations of independence

notions have been studied by Adler and others (and back to von

Neumann).

I The point is to compare tp(A/N ) and tp(A/M).

I In this case, the type tp(A/N ) is locally (∀A− ⊂∗ A) realizable

inside M (∃B− ⊂ M so that B− ≡N− A−

tp(B−/N−) = tp(A−/N−).

I A |chMN generalizes to tame, typeshort AECs Shelah’s

“non-forking”

Page 8: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Coheir

I The previous was an example of an “abstract” independence

notion. In rst order, these presentations of independence

notions have been studied by Adler and others (and back to von

Neumann).

I The point is to compare tp(A/N ) and tp(A/M).

I In this case, the type tp(A/N ) is locally (∀A− ⊂∗ A) realizable

inside M (∃B− ⊂ M so that B− ≡N− A−

tp(B−/N−) = tp(A−/N−).

I A |chMN generalizes to tame, typeshort AECs Shelah’s

“non-forking”

Page 9: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Coheir

I The previous was an example of an “abstract” independence

notion. In rst order, these presentations of independence

notions have been studied by Adler and others (and back to von

Neumann).

I The point is to compare tp(A/N ) and tp(A/M).

I In this case, the type tp(A/N ) is locally (∀A− ⊂∗ A) realizable

inside M (∃B− ⊂ M so that B− ≡N− A−

tp(B−/N−) = tp(A−/N−).

I A |chMN generalizes to tame, typeshort AECs Shelah’s

“non-forking”

Page 10: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Coheir

I The previous was an example of an “abstract” independence

notion. In rst order, these presentations of independence

notions have been studied by Adler and others (and back to von

Neumann).

I The point is to compare tp(A/N ) and tp(A/M).

I In this case, the type tp(A/N ) is locally (∀A− ⊂∗ A) realizable

inside M (∃B− ⊂ M so that B− ≡N− A−

tp(B−/N−) = tp(A−/N−).

I A |chMN generalizes to tame, typeshort AECs Shelah’s

“non-forking”

Page 11: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Landmarks: Canonicity of independence

I 1970: Shelah discovers the notion tp(a/B) forks over A (usually

A ⊂ B). He generalized “change in Morley Rank”, from ω-stable

theories.

I 1974: Lascar proved that nonforking independence is canonical

in superstable theories.

I 1984: Harnik-Harrington generalized this to stable theories.

I 1997: Kim-Pillay generalized this to simple theories, by using

the “independence theorem” (really, type amalgamation)

I 2015: Chernikov-Ramsey generalize Kim-Pillay to SOP1

theories

I 2016: Boney-Grossberg-Kolesnikov-VanDieren extend

Harnik-Harrington to AECs (satisfying NMM, AP, JEP)

Page 12: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Landmarks: Canonicity of independence

I 1970: Shelah discovers the notion tp(a/B) forks over A (usually

A ⊂ B). He generalized “change in Morley Rank”, from ω-stable

theories.

I 1974: Lascar proved that nonforking independence is canonical

in superstable theories.

I 1984: Harnik-Harrington generalized this to stable theories.

I 1997: Kim-Pillay generalized this to simple theories, by using

the “independence theorem” (really, type amalgamation)

I 2015: Chernikov-Ramsey generalize Kim-Pillay to SOP1

theories

I 2016: Boney-Grossberg-Kolesnikov-VanDieren extend

Harnik-Harrington to AECs (satisfying NMM, AP, JEP)

Page 13: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Landmarks: Canonicity of independence

I 1970: Shelah discovers the notion tp(a/B) forks over A (usually

A ⊂ B). He generalized “change in Morley Rank”, from ω-stable

theories.

I 1974: Lascar proved that nonforking independence is canonical

in superstable theories.

I 1984: Harnik-Harrington generalized this to stable theories.

I 1997: Kim-Pillay generalized this to simple theories, by using

the “independence theorem” (really, type amalgamation)

I 2015: Chernikov-Ramsey generalize Kim-Pillay to SOP1

theories

I 2016: Boney-Grossberg-Kolesnikov-VanDieren extend

Harnik-Harrington to AECs (satisfying NMM, AP, JEP)

Page 14: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Landmarks: Canonicity of independence

I 1970: Shelah discovers the notion tp(a/B) forks over A (usually

A ⊂ B). He generalized “change in Morley Rank”, from ω-stable

theories.

I 1974: Lascar proved that nonforking independence is canonical

in superstable theories.

I 1984: Harnik-Harrington generalized this to stable theories.

I 1997: Kim-Pillay generalized this to simple theories, by using

the “independence theorem” (really, type amalgamation)

I 2015: Chernikov-Ramsey generalize Kim-Pillay to SOP1

theories

I 2016: Boney-Grossberg-Kolesnikov-VanDieren extend

Harnik-Harrington to AECs (satisfying NMM, AP, JEP)

Page 15: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Landmarks: Canonicity of independence

I 1970: Shelah discovers the notion tp(a/B) forks over A (usually

A ⊂ B). He generalized “change in Morley Rank”, from ω-stable

theories.

I 1974: Lascar proved that nonforking independence is canonical

in superstable theories.

I 1984: Harnik-Harrington generalized this to stable theories.

I 1997: Kim-Pillay generalized this to simple theories, by using

the “independence theorem” (really, type amalgamation)

I 2015: Chernikov-Ramsey generalize Kim-Pillay to SOP1

theories

I 2016: Boney-Grossberg-Kolesnikov-VanDieren extend

Harnik-Harrington to AECs (satisfying NMM, AP, JEP)

Page 16: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Landmarks: Canonicity of independence

I 1970: Shelah discovers the notion tp(a/B) forks over A (usually

A ⊂ B). He generalized “change in Morley Rank”, from ω-stable

theories.

I 1974: Lascar proved that nonforking independence is canonical

in superstable theories.

I 1984: Harnik-Harrington generalized this to stable theories.

I 1997: Kim-Pillay generalized this to simple theories, by using

the “independence theorem” (really, type amalgamation)

I 2015: Chernikov-Ramsey generalize Kim-Pillay to SOP1

theories

I 2016: Boney-Grossberg-Kolesnikov-VanDieren extend

Harnik-Harrington to AECs (satisfying NMM, AP, JEP)

Page 17: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Map of the Universe (à la façon FO Mod Th)

from Gabriel Conant’s interactive website -

http://forkinganddividing.com/

Page 18: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Two extensions, inside and orthogonal to “the map”

Theorem (2015 - Chernikov, Ramsey - guaranteeing SOP1)

Let | be an Aut(M)-invariant independence relation on small subsetsof M (the monster M |= T) such that for all M ≺∗ M

1. [strong nite character] if a 6 |Mb then there is a formula

ϕ(x, b,m) ∈ tp(a/bM) such that for any a′ |= ϕ(x, b,m),a′ 6 |

Mb

2. [existence over models] M |= T implies a |MM for all a

3. [monotonicity] aa′ |Mbb′ implies a |

Mb

4. [symmetry] a |Mb⇔ b |

Ma

5. [independent amalgamation] c0 | M c1, b0 | M c0, b1 | M c1,b0 ≡M b1 implies there exists b with b ≡c0M b0, b ≡c1M b1.

Then, T is NSOP1.

Page 19: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Two extensions, inside and orthogonal to “the map”

Theorem (2015 - Chernikov, Ramsey - guaranteeing SOP1)

Let | be an Aut(M)-invariant independence relation on small subsetsof M (the monster M |= T) such that for all M ≺∗ M

1. [strong nite character] if a 6 |Mb then there is a formula

ϕ(x, b,m) ∈ tp(a/bM) such that for any a′ |= ϕ(x, b,m),a′ 6 |

Mb

2. [existence over models] M |= T implies a |MM for all a

3. [monotonicity] aa′ |Mbb′ implies a |

Mb

4. [symmetry] a |Mb⇔ b |

Ma

5. [independent amalgamation] c0 | M c1, b0 | M c0, b1 | M c1,b0 ≡M b1 implies there exists b with b ≡c0M b0, b ≡c1M b1.

Then, T is NSOP1.

Page 20: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Two extensions, inside and orthogonal to “the map”

Theorem (2015 - Chernikov, Ramsey - guaranteeing SOP1)

Let | be an Aut(M)-invariant independence relation on small subsetsof M (the monster M |= T) such that for all M ≺∗ M

1. [strong nite character] if a 6 |Mb then there is a formula

ϕ(x, b,m) ∈ tp(a/bM) such that for any a′ |= ϕ(x, b,m),a′ 6 |

Mb

2. [existence over models] M |= T implies a |MM for all a

3. [monotonicity] aa′ |Mbb′ implies a |

Mb

4. [symmetry] a |Mb⇔ b |

Ma

5. [independent amalgamation] c0 | M c1, b0 | M c0, b1 | M c1,b0 ≡M b1 implies there exists b with b ≡c0M b0, b ≡c1M b1.

Then, T is NSOP1.

Page 21: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Two extensions, inside and orthogonal to “the map”

Theorem (2015 - Chernikov, Ramsey - guaranteeing SOP1)

Let | be an Aut(M)-invariant independence relation on small subsetsof M (the monster M |= T) such that for all M ≺∗ M

1. [strong nite character] if a 6 |Mb then there is a formula

ϕ(x, b,m) ∈ tp(a/bM) such that for any a′ |= ϕ(x, b,m),a′ 6 |

Mb

2. [existence over models] M |= T implies a |MM for all a

3. [monotonicity] aa′ |Mbb′ implies a |

Mb

4. [symmetry] a |Mb⇔ b |

Ma

5. [independent amalgamation] c0 | M c1, b0 | M c0, b1 | M c1,b0 ≡M b1 implies there exists b with b ≡c0M b0, b ≡c1M b1.

Then, T is NSOP1.

Page 22: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Two extensions, inside and orthogonal to “the map”

Theorem (2015 - Chernikov, Ramsey - guaranteeing SOP1)

Let | be an Aut(M)-invariant independence relation on small subsetsof M (the monster M |= T) such that for all M ≺∗ M

1. [strong nite character] if a 6 |Mb then there is a formula

ϕ(x, b,m) ∈ tp(a/bM) such that for any a′ |= ϕ(x, b,m),a′ 6 |

Mb

2. [existence over models] M |= T implies a |MM for all a

3. [monotonicity] aa′ |Mbb′ implies a |

Mb

4. [symmetry] a |Mb⇔ b |

Ma

5. [independent amalgamation] c0 | M c1, b0 | M c0, b1 | M c1,b0 ≡M b1 implies there exists b with b ≡c0M b0, b ≡c1M b1.

Then, T is NSOP1.

Page 23: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Two extensions, inside and orthogonal to “the map”

Theorem (2015 - Chernikov, Ramsey - guaranteeing SOP1)

Let | be an Aut(M)-invariant independence relation on small subsetsof M (the monster M |= T) such that for all M ≺∗ M

1. [strong nite character] if a 6 |Mb then there is a formula

ϕ(x, b,m) ∈ tp(a/bM) such that for any a′ |= ϕ(x, b,m),a′ 6 |

Mb

2. [existence over models] M |= T implies a |MM for all a

3. [monotonicity] aa′ |Mbb′ implies a |

Mb

4. [symmetry] a |Mb⇔ b |

Ma

5. [independent amalgamation] c0 | M c1, b0 | M c0, b1 | M c1,b0 ≡M b1 implies there exists b with b ≡c0M b0, b ≡c1M b1.

Then, T is NSOP1.

Page 24: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Canonicity for |ch on AEC’s

I At most one abstract independence relation satises existence,

extension, uniqueness and local character (under NMM, AP,

JEP)

I |ch equals non-forking if furthermore the AEC is tame and

typeshort

Page 25: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Continuous Model Theory - Origins

I Remote origins of Continuous

Model Theory: von Neumann

(Continuous Geometry),

I Chang & Keisler (1966),

I Ben Yaacov, Berenstein, Henson,

Usvyatsov: a monograph that

summarizes stability theory for

continuous logic (around 2004)

I More recently, Boney, Caicedo,

Eagle, Iovino, etc. have

generalized continuous logic

I Hirvonen, Hyttinen - V.,

Zambrano: categoricity and

superstability in metric AECs

Page 26: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Continuous predicates and functions

Denition (Interpreting in a metric structure)

Fix (M, d) a bounded metric space. A continuous n-ary predicate is a

uniformly continuous function

P : Mn → [0, 1].

A continuous n-ary function is a uniformly continuous function

f : Mn → M.

Page 27: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Metric structures

Metric structures are of the form

M =(M, d, (fi)i∈I , (Rj)j∈J , (ak)k∈K

)

where the Ri and the fj are (uniformly) continuous functions with

values in [0, 1], the ak are distinguished elements of M .

M is a bounded metric space. Each function, relation must be

endowed with a modulus of uniform continuity.

Page 28: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Metric structures

Metric structures are of the form

M =(M, d, (fi)i∈I , (Rj)j∈J , (ak)k∈K

)where the Ri and the fj are (uniformly) continuous functions with

values in [0, 1], the ak are distinguished elements of M .

M is a bounded metric space.

Each function, relation must be

endowed with a modulus of uniform continuity.

Page 29: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Metric structures

Metric structures are of the form

M =(M, d, (fi)i∈I , (Rj)j∈J , (ak)k∈K

)where the Ri and the fj are (uniformly) continuous functions with

values in [0, 1], the ak are distinguished elements of M .

M is a bounded metric space. Each function, relation must be

endowed with a modulus of uniform continuity.

Page 30: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Examples of FO metric structures

Example

I Any FO structure, endowed with the discrete metric.

I Banach algebras (bounding them).

I Hilbert spaces with inner product as a binary predicate.

I For a probability space (Ω,B, µ), construct a metric structureM based on

the usual measure algebra of (Ω,B, µ).

I Representations of C∗-algebras (Argoty, Berenstein, Ben Yaacov, V.).

I Farah, Hart: pathological properties of operator algebras

Page 31: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Examples of FO metric structures

Example

I Any FO structure, endowed with the discrete metric.

I Banach algebras (bounding them).

I Hilbert spaces with inner product as a binary predicate.

I For a probability space (Ω,B, µ), construct a metric structureM based on

the usual measure algebra of (Ω,B, µ).

I Representations of C∗-algebras (Argoty, Berenstein, Ben Yaacov, V.).

I Farah, Hart: pathological properties of operator algebras

Page 32: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Examples of FO metric structures

Example

I Any FO structure, endowed with the discrete metric.

I Banach algebras (bounding them).

I Hilbert spaces with inner product as a binary predicate.

I For a probability space (Ω,B, µ), construct a metric structureM based on

the usual measure algebra of (Ω,B, µ).

I Representations of C∗-algebras (Argoty, Berenstein, Ben Yaacov, V.).

I Farah, Hart: pathological properties of operator algebras

Page 33: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Examples of FO metric structures

Example

I Any FO structure, endowed with the discrete metric.

I Banach algebras (bounding them).

I Hilbert spaces with inner product as a binary predicate.

I For a probability space (Ω,B, µ), construct a metric structureM based on

the usual measure algebra of (Ω,B, µ).

I Representations of C∗-algebras (Argoty, Berenstein, Ben Yaacov, V.).

I Farah, Hart: pathological properties of operator algebras

Page 34: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Examples of FO metric structures

Example

I Any FO structure, endowed with the discrete metric.

I Banach algebras (bounding them).

I Hilbert spaces with inner product as a binary predicate.

I For a probability space (Ω,B, µ), construct a metric structureM based on

the usual measure algebra of (Ω,B, µ).

I Representations of C∗-algebras (Argoty, Berenstein, Ben Yaacov, V.).

I Farah, Hart: pathological properties of operator algebras

Page 35: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

The syntax

1. Terms: as usual.

2. Atomic formulas: d(t1, tn) and R(t1, · · · , tn), if the ti are terms.

Formulas are then interpreted as functions into [0, 1].

3. Connectives: continuous functions from [0, 1]n → [0, 1].Therefore, applying connectives to formulas gives new

formulas.

4. Quantiers: supx ϕ(x) (universal) and infx ϕ(x) (existential).

Page 36: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Interpretation

The logical distance between ϕ(x) and ψ(x) is

supa∈M |ϕM(a)− ψM(a)|.The satisfaction relation is dened on conditions rather than on

formulas.

Conditions are expressions of the form ϕ(x) ≤ ψ(y), ϕ(x) ≤ ψ(y),

ϕ(x) ≥ ψ(y), etc.

Notice also that the set of connectives is too large, but it may be

“densely” and uniformly generated by 0, 1, x/2,.−: for every ε, for

every connective f (t1, · · · , tn) there exists a connective g(t1, · · · , tn)generated by these four by composition such that |f (~t)− g(~t)| < ε.

Page 37: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Interpretation

The logical distance between ϕ(x) and ψ(x) is

supa∈M |ϕM(a)− ψM(a)|.The satisfaction relation is dened on conditions rather than on

formulas.

Conditions are expressions of the form ϕ(x) ≤ ψ(y), ϕ(x) ≤ ψ(y),

ϕ(x) ≥ ψ(y), etc.

Notice also that the set of connectives is too large, but it may be

“densely” and uniformly generated by 0, 1, x/2,.−: for every ε, for

every connective f (t1, · · · , tn) there exists a connective g(t1, · · · , tn)generated by these four by composition such that |f (~t)− g(~t)| < ε.

Page 38: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Interpretation

The logical distance between ϕ(x) and ψ(x) is

supa∈M |ϕM(a)− ψM(a)|.The satisfaction relation is dened on conditions rather than on

formulas.

Conditions are expressions of the form ϕ(x) ≤ ψ(y), ϕ(x) ≤ ψ(y),

ϕ(x) ≥ ψ(y), etc.

Notice also that the set of connectives is too large, but it may be

“densely” and uniformly generated by 0, 1, x/2,.−: for every ε, for

every connective f (t1, · · · , tn) there exists a connective g(t1, · · · , tn)generated by these four by composition such that |f (~t)− g(~t)| < ε.

Page 39: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Stability Theory

I Stability (Ben Yaacov, Iovino, etc.),

I Categoricity for countable languages (Ben Yaacov),

I ω-stability,

I Dependent theories (Ben Yaacov),

I Not much geometric stability theory: no analog to

Baldwin-Lachlan (no minimality, except some openings by

Usvyatsov and Shelah in the context of ℵ1-categorical Banach

spaces),

I NO simplicity!!! (Berenstein, Hyttinen, V.),

I Keisler measures, NIP (Hrushovski, Pillay, etc.).

Page 40: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Stability Theory

I Stability (Ben Yaacov, Iovino, etc.),

I Categoricity for countable languages (Ben Yaacov),

I ω-stability,

I Dependent theories (Ben Yaacov),

I Not much geometric stability theory: no analog to

Baldwin-Lachlan (no minimality, except some openings by

Usvyatsov and Shelah in the context of ℵ1-categorical Banach

spaces),

I NO simplicity!!! (Berenstein, Hyttinen, V.),

I Keisler measures, NIP (Hrushovski, Pillay, etc.).

Page 41: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Stability Theory

I Stability (Ben Yaacov, Iovino, etc.),

I Categoricity for countable languages (Ben Yaacov),

I ω-stability,

I Dependent theories (Ben Yaacov),

I Not much geometric stability theory: no analog to

Baldwin-Lachlan (no minimality, except some openings by

Usvyatsov and Shelah in the context of ℵ1-categorical Banach

spaces),

I NO simplicity!!! (Berenstein, Hyttinen, V.),

I Keisler measures, NIP (Hrushovski, Pillay, etc.).

Page 42: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Stability Theory

I Stability (Ben Yaacov, Iovino, etc.),

I Categoricity for countable languages (Ben Yaacov),

I ω-stability,

I Dependent theories (Ben Yaacov),

I Not much geometric stability theory: no analog to

Baldwin-Lachlan (no minimality, except some openings by

Usvyatsov and Shelah in the context of ℵ1-categorical Banach

spaces),

I NO simplicity!!! (Berenstein, Hyttinen, V.),

I Keisler measures, NIP (Hrushovski, Pillay, etc.).

Page 43: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

"Continuous Model Theory" beyond First Order

Several contexts, some unexplored so far.

1. Metric Abstract Elementary Classes (Hirvonen, Hyttinen -

ω-stability, V. Zambrano - superstability, domination, notions of

independence): an amalgam of the power of Abstract

Elementary Classes with metric ideas.

2. Continuous Lω1ω . So far, no published results as such. There are

however “Lindström theorems” for Continuous First Order due

to Caicedo/Iovino and Eagle.

Page 44: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

"Continuous Model Theory" beyond First Order

Several contexts, some unexplored so far.

1. Metric Abstract Elementary Classes (Hirvonen, Hyttinen -

ω-stability, V. Zambrano - superstability, domination, notions of

independence): an amalgam of the power of Abstract

Elementary Classes with metric ideas.

2. Continuous Lω1ω . So far, no published results as such. There are

however “Lindström theorems” for Continuous First Order due

to Caicedo/Iovino and Eagle.

Page 45: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Independence, aware of metric and continuity

Denition (ε-coheir / the simplest)

Let M ≺ N .

Ach,ε|M

N

m

∀A− ⊂∗ A ∀N− ≺∗ N

∃B− ⊂ M so that

d(tp(B−/N−), tp(A−/N−)) < ε.

Page 46: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Independence, aware of metric and continuity

Denition (ε-coheir / the simplest)

Let M ≺ N .

Ach,ε|M

N

m

∀A− ⊂∗ A ∀N− ≺∗ N

∃B− ⊂ M so that

d(tp(B−/N−), tp(A−/N−)) < ε.

Page 47: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Independence, aware of metric and continuity

Denition (ε-coheir / the simplest)

Let M ≺ N .

Ach,ε|M

N

m

∀A− ⊂∗ A ∀N− ≺∗ N

∃B− ⊂ M so that

d(tp(B−/N−), tp(A−/N−)) < ε.

Page 48: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Independence, aware of metric and continuity

Denition (ε-coheir / the simplest)

Let M ≺ N .

Ach,ε|M

N

m

∀A− ⊂∗ A ∀N− ≺∗ N

∃B− ⊂ M so that

d(tp(B−/N−), tp(A−/N−)) < ε.

Page 49: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Generic Predicates

Theorem (Chatzidakis-Pillay)

Let T0 = ACF0, LP = L ∪ P, P a new symbol for a unary predicate.TP = T0∪“P is a generic predicate”- again the model companion ofT0∪“P is a predicate”. Then TP is a simple theory.

The existence of TP is ensured as T0 eliminates ∃∞.

If (k, P) |= TP , A,B,C ⊂ k, C = C,

AP|CB ⇔ A |

CB

Page 50: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Generic Predicates

Theorem (Chatzidakis-Pillay)

Let T0 = ACF0, LP = L ∪ P, P a new symbol for a unary predicate.TP = T0∪“P is a generic predicate”- again the model companion ofT0∪“P is a predicate”. Then TP is a simple theory.

The existence of TP is ensured as T0 eliminates ∃∞.

If (k, P) |= TP , A,B,C ⊂ k, C = C,

AP|CB ⇔ A |

CB

Page 51: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Generic Predicates

Theorem (Chatzidakis-Pillay)

Let T0 = ACF0, LP = L ∪ P, P a new symbol for a unary predicate.TP = T0∪“P is a generic predicate”- again the model companion ofT0∪“P is a predicate”. Then TP is a simple theory.

The existence of TP is ensured as T0 eliminates ∃∞.

If (k, P) |= TP , A,B,C ⊂ k, C = C,

AP|CB ⇔ A |

CB

Page 52: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Our aim: an analog of Chatzidakis-Pillay, TP

We will next look at structures of the form

(H,+, 0, 〈〉, dN )

where dN (x) “measures” the distance to a set of “black points” which

we call N (H).

Page 53: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

TP and simplicity

This will correspond in a natural way, once we go to the model

companions, to a generic “predicate” no longer dividing the Hilbert

space into two complementary areas, but rather a “generic grey

spot” with shades of grey between black and white. . . in a generic

way. Our result also intended to produce the rst simple unstable

theories in Continuous Model Theory.

Page 54: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Basic definitions

We consider structures of the form

(H,+, 0, 〈〉, dN )

The function dN is therefore additional to the structure of Hilbert

Spaces. Let THilbert be the theory of Hilbert spaces, and let

T0 := THilbert ∪ Ax1,Ax2:1. Ax1: supx infy max|dN (x)− ‖x − y‖|, dN (y) = 0

2. Ax2: supx supy dN (x) ≤ dN (y) + ‖x − y‖

Page 55: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Basic definitions

We consider structures of the form

(H,+, 0, 〈〉, dN )

The function dN is therefore additional to the structure of Hilbert

Spaces. Let THilbert be the theory of Hilbert spaces, and let

T0 := THilbert ∪ Ax1,Ax2:1. Ax1: supx infy max|dN (x)− ‖x − y‖|, dN (y) = 0

2. Ax2: supx supy dN (x) ≤ dN (y) + ‖x − y‖

Page 56: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Amalgamation

Let (H0, d0

N ) ⊂ (Hi, diN ) where i ∈ 1, 2 and let H1 | H0

H2 be

Hilbert spaces with distance functions, all of them in

Gr = 〈H , . . . , dN 〉|dN (0) = r.

Let H3 = spanH1,H2 and let

d3

N (v) = min

√d1

N (PH1(v))2 + ‖P

H2∩H⊥

0

(v)‖2,√

d2

N (PH2(v))2 + ‖P

H1∩H⊥

0

(v)‖2

.

Then (Hi, diN ) ⊂ (H3, d3

N ) for i ∈ 1, 2, (H3, d3

N ) |= T0 and

(H3, d3

N ) ∈ Gr .

Page 57: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Fraïssé limits

Characterize

Td,0 = Thcont(lim−→Fr

(K0))

for the class K0 of all nite dimensional models of T0 such that 0 is a

black point.

Page 58: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

A typical configuration for the antecedent

Page 59: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

And the new axiom, modulo ε and ϕ

Page 60: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

EC models

Theorem (Berenstein and V.)

(M, dN ) is an existentially closed model of T0 if and only if(M, dN ) |= TN .

However,

Theorem (Berenstein and V.)

TN does not have elimination of quantiers.

Page 61: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

EC models - QE

(We take H with orthonormal basis u1, u2; N = 0, u0 + 1

4u1,

dN (x) = min1, d(x,N ). Then (H1, d1

N ) |= T0. Let a = u0,

b = u0 − 1

4u1 and c = u0 + 1

4u1. Then d′N (b) = 1

2.

Let H ′′ also have an orthonormal basis vi : i ∈ ω,N2 = x ∈ H : ‖x − v1‖ = 1

4, P

span(v1)(x) = v1 ∪ 0 y

d′′N (x) = min1, d(x,N2).

Then (span(a), d1

N span(a))F∼= (span(v1), d2

N span(v1)). But (H ′, d′N )and (H ′′, d′′N ) have no amalgam over the common part: if they had,

we would have d(F(b), v1 + 1

4vi) = d(b, v1 + 1

4vi) < 1

2for (some)

i > 1, i.e. d1

N (b) < 1

2, contradiction.)

Page 62: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

No amalgams, no QE - Interference

Page 63: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

TN , among the few theories both TP2 and NTP1

Denition (TP2)

A formula ϕ(x; y) has TP2 if there exists k < ω and there exists a

matrix of tuples

a00 a01 · · · a0i · · ·a10 a11 · · · a1i · · ·...

aα0 aα1 · · · aαi · · ·...

rows (ϕ(x, aα,i) | i < ω is k-inconsistent for each α)

functions across going down for each f ∈ ωωϕ(x, aα,f (α)) | α < ω is consistent.

Page 64: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

TP2 - bad side

Theorem (Hyttinen, V.)

TN has TP2.

Proof (sketch) Let 〈af | f : ω → ω〉 ∪ 〈bn | n < ω〉 ∪ 〈cn,i | i, n < ω〉be an orthonormal basis of a Hilbert space.

Let the “black points” consist of

af + bn + 1

2cn,f (i) | f : ω → ω, n, i < ω and let

ϕ(x, y, z) : dN (x + y − 1

2

z) ≥ 1 ∧ dN (x + y +1

2

z) ≤ 0.

This formula witnesses TP2. (Originally we had a proof of failure of

simplicity using Casanovas/Shelah’s characterization but this is far

worse!)

Page 65: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

TP2 - bad side

Theorem (Hyttinen, V.)

TN has TP2.

Proof (sketch) Let 〈af | f : ω → ω〉 ∪ 〈bn | n < ω〉 ∪ 〈cn,i | i, n < ω〉be an orthonormal basis of a Hilbert space.

Let the “black points” consist of

af + bn + 1

2cn,f (i) | f : ω → ω, n, i < ω and let

ϕ(x, y, z) : dN (x + y − 1

2

z) ≥ 1 ∧ dN (x + y +1

2

z) ≤ 0.

This formula witnesses TP2. (Originally we had a proof of failure of

simplicity using Casanovas/Shelah’s characterization but this is far

worse!)

Page 66: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

TP2 - bad side

Theorem (Hyttinen, V.)

TN has TP2.

Proof (sketch) Let 〈af | f : ω → ω〉 ∪ 〈bn | n < ω〉 ∪ 〈cn,i | i, n < ω〉be an orthonormal basis of a Hilbert space.

Let the “black points” consist of

af + bn + 1

2cn,f (i) | f : ω → ω, n, i < ω and let

ϕ(x, y, z) : dN (x + y − 1

2

z) ≥ 1 ∧ dN (x + y +1

2

z) ≤ 0.

This formula witnesses TP2. (Originally we had a proof of failure of

simplicity using Casanovas/Shelah’s characterization but this is far

worse!)

Page 67: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

NTP1... really NSOP1 - good side

Theorem (Hyttinen, V.)

TN has NSOP1.

Denition (SOP1)

ϕ(x; y) is SOP1 i there exists a tree of parameters (aη)η∈2<ω such

that

I for every η ∈ 2ω

, ϕ(x; aηn) | n < ω is consistent

I if η_0 E ν ∈ 2<ω

then ϕ(x; aη_1), ϕ(x; aν) is inconsistent.

T is NSOP1 if NO formula has SOP1 in T .

Page 68: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

NTP1... really NSOP1 - good side

Sketch of proof:

I We adapt Chernikov-Ramsey (characterization of NSOP1 in

terms of an independence notion) to the continuous setting.

I We prove that the independence property | ∗ satises the ve

properties.

Page 69: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

estions

I Why does Chatzidakis-Pillay fail so badly in the continuous

setting?

I Why the “polarization” of dividing lines in the continuous case?

I The interference and the tree property TP2 both seem to be

connected with the non-triviality of the metric. Is there a

deeper model-theoretic/logical reason for this?

Page 70: Andrés Villaveces Universidad Nacional de Colombia - BogotáIndependence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good

Independence notions Continuous world/independence Obstacles to continuous Chatzidakis-Pillay Generic independence, good and bad

Thank you! Gràcies! ¡Gracias!