Group Colorings and Bernoulli Subflows Su Gao University of North - - PowerPoint PPT Presentation

group colorings and bernoulli subflows
SMART_READER_LITE
LIVE PREVIEW

Group Colorings and Bernoulli Subflows Su Gao University of North - - PowerPoint PPT Presentation

Group Colorings and Bernoulli Subflows Su Gao University of North Texas Steve Jackson University of North Texas Brandon Seward* University of Michigan RTG Logic and Dynamics Conference June 48, 2012 Brandon Seward () Group Colorings and


slide-1
SLIDE 1

Group Colorings and Bernoulli Subflows

Su Gao University of North Texas Steve Jackson University of North Texas Brandon Seward* University of Michigan RTG Logic and Dynamics Conference June 4–8, 2012

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 1 / 16

slide-2
SLIDE 2

Bernoulli Flows

For G a countable group and k ≥ 2 an integer, we define the Bernoulli flow kG := set of all functions G → {0, . . . , k − 1} =

  • g∈G

{0, 1, . . . , k − 1}.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 2 / 16

slide-3
SLIDE 3

Bernoulli Flows

For G a countable group and k ≥ 2 an integer, we define the Bernoulli flow kG := set of all functions G → {0, . . . , k − 1} =

  • g∈G

{0, 1, . . . , k − 1}. We give kG the product topology.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 2 / 16

slide-4
SLIDE 4

Bernoulli Flows

For G a countable group and k ≥ 2 an integer, we define the Bernoulli flow kG := set of all functions G → {0, . . . , k − 1} =

  • g∈G

{0, 1, . . . , k − 1}. We give kG the product topology. G naturally acts on kG: (g · x)(h) = x(g−1h).

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 2 / 16

slide-5
SLIDE 5

Bernoulli Flows

For G a countable group and k ≥ 2 an integer, we define the Bernoulli flow kG := set of all functions G → {0, . . . , k − 1} =

  • g∈G

{0, 1, . . . , k − 1}. We give kG the product topology. G naturally acts on kG: (g · x)(h) = x(g−1h). kG is compact and G acts continuously.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 2 / 16

slide-6
SLIDE 6

Bernoulli Flows

For G a countable group and k ≥ 2 an integer, we define the Bernoulli flow kG := set of all functions G → {0, . . . , k − 1} =

  • g∈G

{0, 1, . . . , k − 1}. We give kG the product topology. G naturally acts on kG: (g · x)(h) = x(g−1h). kG is compact and G acts continuously. For µ a probability measure on {0, 1, . . . , k − 1}, we call the product measure µG a Bernoulli measure.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 2 / 16

slide-7
SLIDE 7

Bernoulli Flows

For G a countable group and k ≥ 2 an integer, we define the Bernoulli flow kG := set of all functions G → {0, . . . , k − 1} =

  • g∈G

{0, 1, . . . , k − 1}. We give kG the product topology. G naturally acts on kG: (g · x)(h) = x(g−1h). kG is compact and G acts continuously. For µ a probability measure on {0, 1, . . . , k − 1}, we call the product measure µG a Bernoulli measure. Under a Bernoulli measure, the action is measure preserving and ergodic.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 2 / 16

slide-8
SLIDE 8

Bernoulli Flows

For G a countable group and k ≥ 2 an integer, we define the Bernoulli flow kG := set of all functions G → {0, . . . , k − 1} =

  • g∈G

{0, 1, . . . , k − 1}. We give kG the product topology. G naturally acts on kG: (g · x)(h) = x(g−1h). kG is compact and G acts continuously. For µ a probability measure on {0, 1, . . . , k − 1}, we call the product measure µG a Bernoulli measure. Under a Bernoulli measure, the action is measure preserving and ergodic. A point x ∈ kG is aperiodic (or free) if its stabilizer is trivial. We let F(kG) denote the free part of the action, i.e. the set of all aperiodic (free) points of kG.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 2 / 16

slide-9
SLIDE 9

Hyperfinite Equivalence Relations

Recall that a countable Borel equivalence relation E is hyperfinite if it can be expressed as the increasing union of finite Borel equivlance relations.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 3 / 16

slide-10
SLIDE 10

Hyperfinite Equivalence Relations

Recall that a countable Borel equivalence relation E is hyperfinite if it can be expressed as the increasing union of finite Borel equivlance relations. A long standing open problem asks if every Borel action of an amenable group induces a hyperfinite equivalence relation (via its orbits).

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 3 / 16

slide-11
SLIDE 11

Hyperfinite Equivalence Relations

Recall that a countable Borel equivalence relation E is hyperfinite if it can be expressed as the increasing union of finite Borel equivlance relations. A long standing open problem asks if every Borel action of an amenable group induces a hyperfinite equivalence relation (via its orbits). This is known to hold if G is abelian (Gao–Jackson, 2007) or finitely generated and virtually nilpotent (Jackson–Kechris–Louveau, 2002).

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 3 / 16

slide-12
SLIDE 12

Hyperfinite Equivalence Relations

Recall that a countable Borel equivalence relation E is hyperfinite if it can be expressed as the increasing union of finite Borel equivlance relations. A long standing open problem asks if every Borel action of an amenable group induces a hyperfinite equivalence relation (via its orbits). This is known to hold if G is abelian (Gao–Jackson, 2007) or finitely generated and virtually nilpotent (Jackson–Kechris–Louveau, 2002). Progress on this problem hinges on finding constructions for better marker sets and better marker regions.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 3 / 16

slide-13
SLIDE 13

Hyperfinite Equivalence Relations

Recall that a countable Borel equivalence relation E is hyperfinite if it can be expressed as the increasing union of finite Borel equivlance relations. A long standing open problem asks if every Borel action of an amenable group induces a hyperfinite equivalence relation (via its orbits). This is known to hold if G is abelian (Gao–Jackson, 2007) or finitely generated and virtually nilpotent (Jackson–Kechris–Louveau, 2002). Progress on this problem hinges on finding constructions for better marker sets and better marker regions. A classical construction for marker sets comes from the following. Lemma (Slaman–Steel, 1988) There is a decreasing sequence (Sn) of Borel complete sections of F(kG) such that

n Sn = ∅.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 3 / 16

slide-14
SLIDE 14

Motivating Question

A strengthening of the Slaman–Steel lemma could lead to improved understanding of hyperfinite equivalence relations.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 4 / 16

slide-15
SLIDE 15

Motivating Question

A strengthening of the Slaman–Steel lemma could lead to improved understanding of hyperfinite equivalence relations. Question Does there exist a decreasing sequence (Sn) of complete sections of F(kG) such that each Sn is relatively closed in F(kG) and

n Sn = ∅?

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 4 / 16

slide-16
SLIDE 16

Motivating Question

A strengthening of the Slaman–Steel lemma could lead to improved understanding of hyperfinite equivalence relations. Question Does there exist a decreasing sequence (Sn) of complete sections of F(kG) such that each Sn is relatively closed in F(kG) and

n Sn = ∅?

If F(kG) where to contain a compact G-invariant subset, then by compactness the answer to the above question must be ’No.’

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 4 / 16

slide-17
SLIDE 17

Free Subflows

A subflow of a Bernoulli flow kG is a closed (hence compact) subset of kG which is invariant under the action of G. A subflow X ⊆ kG is free if X ⊆ F(kG).

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 5 / 16

slide-18
SLIDE 18

Free Subflows

A subflow of a Bernoulli flow kG is a closed (hence compact) subset of kG which is invariant under the action of G. A subflow X ⊆ kG is free if X ⊆ F(kG). A point x ∈ kG is a k-coloring if the closure of its orbit, [x], is a free subflow (equivalently [x] ⊆ F(kG)).

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 5 / 16

slide-19
SLIDE 19

Free Subflows

A subflow of a Bernoulli flow kG is a closed (hence compact) subset of kG which is invariant under the action of G. A subflow X ⊆ kG is free if X ⊆ F(kG). A point x ∈ kG is a k-coloring if the closure of its orbit, [x], is a free subflow (equivalently [x] ⊆ F(kG)). Notice that x is a k-coloring if and only if it is contained in some free

  • subflow. Thus, the set of all k-colorings coincides with the union of all free

subflows.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 5 / 16

slide-20
SLIDE 20

Free Subflows

A subflow of a Bernoulli flow kG is a closed (hence compact) subset of kG which is invariant under the action of G. A subflow X ⊆ kG is free if X ⊆ F(kG). A point x ∈ kG is a k-coloring if the closure of its orbit, [x], is a free subflow (equivalently [x] ⊆ F(kG)). Notice that x is a k-coloring if and only if it is contained in some free

  • subflow. Thus, the set of all k-colorings coincides with the union of all free

subflows. When free subflows are present, the Slaman–Steel marker lemma cannot be strengthened in the manner discussed.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 5 / 16

slide-21
SLIDE 21

A Natural Question

Question Does every Bernoulli flow contain a free subflow?

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 6 / 16

slide-22
SLIDE 22

A Natural Question

Question Does every Bernoulli flow contain a free subflow? It is well known that F(kG) is always comeager and always has full measure under every Bernoulli measure.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 6 / 16

slide-23
SLIDE 23

A Natural Question

Question Does every Bernoulli flow contain a free subflow? It is well known that F(kG) is always comeager and always has full measure under every Bernoulli measure. Lemma (Gao–Jackson–S) The set of k-colorings (i.e. the union of all free subflows) is always meager and always has measure zero under every Bernoulli measure. Thus measure and Baire category arguments seem incapable of addressing the above question.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 6 / 16

slide-24
SLIDE 24

A Natural Question

Question Does every Bernoulli flow contain a free subflow? It is well known that F(kG) is always comeager and always has full measure under every Bernoulli measure. Lemma (Gao–Jackson–S) The set of k-colorings (i.e. the union of all free subflows) is always meager and always has measure zero under every Bernoulli measure. Thus measure and Baire category arguments seem incapable of addressing the above question. If the answer is to be ‘yes,’ a constructive proof is needed.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 6 / 16

slide-25
SLIDE 25

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring:

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-26
SLIDE 26

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-27
SLIDE 27

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0 0.1

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-28
SLIDE 28

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0 0.1 01.10

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-29
SLIDE 29

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0 0.1 01.10 0110.1001

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-30
SLIDE 30

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0 0.1 01.10 0110.1001 01101001.10010110

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-31
SLIDE 31

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0 0.1 01.10 0110.1001 01101001.10010110 0110100110010110.1001011001101001

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-32
SLIDE 32

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0 0.1 01.10 0110.1001 01101001.10010110 0110100110010110.1001011001101001 Why? Any point-wise limit x has the following property:

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-33
SLIDE 33

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0 0.1 01.10 0110.1001 01101001.10010110 0110100110010110.1001011001101001 Why? Any point-wise limit x has the following property: if w is any finite word in the alphabet {0, 1}, then the three-fold concatenation www does not appear in x.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-34
SLIDE 34

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0 0.1 01.10 0110.1001 01101001.10010110 0110100110010110.1001011001101001 Why? Any point-wise limit x has the following property: if w is any finite word in the alphabet {0, 1}, then the three-fold concatenation www does not appear in x. This is a local–global property

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-35
SLIDE 35

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0 0.1 01.10 0110.1001 01101001.10010110 0110100110010110.1001011001101001 Why? Any point-wise limit x has the following property: if w is any finite word in the alphabet {0, 1}, then the three-fold concatenation www does not appear in x. This is a local–global property which extends to all points in [x].

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-36
SLIDE 36

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0 0.1 01.10 0110.1001 01101001.10010110 0110100110010110.1001011001101001 Why? Any point-wise limit x has the following property: if w is any finite word in the alphabet {0, 1}, then the three-fold concatenation www does not appear in x. This is a local–global property which extends to all points in [x]. Periodic points do not have this property.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-37
SLIDE 37

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0 0.1 01.10 0110.1001 01101001.10010110 0110100110010110.1001011001101001 Why? Any point-wise limit x has the following property: if w is any finite word in the alphabet {0, 1}, then the three-fold concatenation www does not appear in x. This is a local–global property which extends to all points in [x]. Periodic points do not have this property. Thus [x] ⊆ F(2Z) and x is a 2-coloring.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-38
SLIDE 38

A 2-coloring on Z

Any point-wise limit of the following sequence is a 2-coloring: .0 0.1 01.10 0110.1001 01101001.10010110 0110100110010110.1001011001101001 Why? Any point-wise limit x has the following property: if w is any finite word in the alphabet {0, 1}, then the three-fold concatenation www does not appear in x. This is a local–global property which extends to all points in [x]. Periodic points do not have this property. Thus [x] ⊆ F(2Z) and x is a 2-coloring. This is (essentially) the Morse–Thue sequence.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 7 / 16

slide-39
SLIDE 39

History

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 8 / 16

slide-40
SLIDE 40

History

Folklore Every Bernoulli flow over Z contains a free subflow.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 8 / 16

slide-41
SLIDE 41

History

Folklore Every Bernoulli flow over Z contains a free subflow. Theorem (Dranishnikov–Schroeder, 2007) If G is a torsion free hyperbolic group and k ≥ 9, then kG contains a free subflow. Their proof relied on using the Morse–Thue sequence along certain geodesic rays in G.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 8 / 16

slide-42
SLIDE 42

History

Folklore Every Bernoulli flow over Z contains a free subflow. Theorem (Dranishnikov–Schroeder, 2007) If G is a torsion free hyperbolic group and k ≥ 9, then kG contains a free subflow. Their proof relied on using the Morse–Thue sequence along certain geodesic rays in G. Theorem (Glasner–Uspenskij, 2007) If G is abelian or residually finite and k ≥ 2, then kG contains a free subflow. In their proof they constructed a free continuous action of G on some compact space X and argued that there must be a subflow of kG factoring

  • nto X. Such a subflow must be free.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 8 / 16

slide-43
SLIDE 43

Free Subflows Always Exist

Theorem (Gao–Jackson–S, 2007) For every countable group G and every k ≥ 2, kG contains a free subflow.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 9 / 16

slide-44
SLIDE 44

Free Subflows Always Exist

Theorem (Gao–Jackson–S, 2007) For every countable group G and every k ≥ 2, kG contains a free subflow. In fact, kG contains uncountably many pairwise disjoint free subflows.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 9 / 16

slide-45
SLIDE 45

Free Subflows Always Exist

Theorem (Gao–Jackson–S, 2007) For every countable group G and every k ≥ 2, kG contains a free subflow. In fact, kG contains uncountably many pairwise disjoint free subflows. Thus the Slaman–Steel marker lemma can not be strengthened to having relatively closed complete sections with empty intersection.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 9 / 16

slide-46
SLIDE 46

Free Subflows Always Exist

Theorem (Gao–Jackson–S, 2007) For every countable group G and every k ≥ 2, kG contains a free subflow. In fact, kG contains uncountably many pairwise disjoint free subflows. Thus the Slaman–Steel marker lemma can not be strengthened to having relatively closed complete sections with empty intersection. The proof is entirely constructive. One ingredient in the proof is the following local–global combinatorial characterization of k-colorings.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 9 / 16

slide-47
SLIDE 47

Free Subflows Always Exist

Theorem (Gao–Jackson–S, 2007) For every countable group G and every k ≥ 2, kG contains a free subflow. In fact, kG contains uncountably many pairwise disjoint free subflows. Thus the Slaman–Steel marker lemma can not be strengthened to having relatively closed complete sections with empty intersection. The proof is entirely constructive. One ingredient in the proof is the following local–global combinatorial characterization of k-colorings. Lemma (Gao–Jackson–S; Pestov) x ∈ kG is a k-coloring if and only if for every 1G = s ∈ G there is a finite T ⊆ G so that ∀g ∈ G ∃t ∈ T x(gst) = x(gt).

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 9 / 16

slide-48
SLIDE 48

Free Subflows Always Exist

Theorem (Gao–Jackson–S, 2007) For every countable group G and every k ≥ 2, kG contains a free subflow. In fact, kG contains uncountably many pairwise disjoint free subflows. Thus the Slaman–Steel marker lemma can not be strengthened to having relatively closed complete sections with empty intersection. The proof is entirely constructive. One ingredient in the proof is the following local–global combinatorial characterization of k-colorings. Lemma (Gao–Jackson–S; Pestov) x ∈ kG is a k-coloring if and only if for every 1G = s ∈ G there is a finite T ⊆ G so that ∀g ∈ G ∃t ∈ T x(gst) = x(gt). Our construction of k-colorings relies on something we call The Fundamental Method.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 9 / 16

slide-49
SLIDE 49

The Fundamental Method

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 10 / 16

slide-50
SLIDE 50

The Fundamental Method

The Fundamental Method is a collection of interlocking tools and constructions which when used collectively provide powerful, highly customizable constructions of points in kG.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 10 / 16

slide-51
SLIDE 51

The Fundamental Method

The Fundamental Method is a collection of interlocking tools and constructions which when used collectively provide powerful, highly customizable constructions of points in kG. Roughly speaking, the Fundamental Method gives you a partial function c with domain a subset of G and range {0, 1, . . . , k − 1} and a sequence (∆n) of subsets of G such that

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 10 / 16

slide-52
SLIDE 52

The Fundamental Method

The Fundamental Method is a collection of interlocking tools and constructions which when used collectively provide powerful, highly customizable constructions of points in kG. Roughly speaking, the Fundamental Method gives you a partial function c with domain a subset of G and range {0, 1, . . . , k − 1} and a sequence (∆n) of subsets of G such that

1 For each n, ∆n is uniformly spread out in G; Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 10 / 16

slide-53
SLIDE 53

The Fundamental Method

The Fundamental Method is a collection of interlocking tools and constructions which when used collectively provide powerful, highly customizable constructions of points in kG. Roughly speaking, the Fundamental Method gives you a partial function c with domain a subset of G and range {0, 1, . . . , k − 1} and a sequence (∆n) of subsets of G such that

1 For each n, ∆n is uniformly spread out in G; 2 For each n there is a finite set F such that the behavoir of c on gF

determines if g ∈ ∆n;

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 10 / 16

slide-54
SLIDE 54

The Fundamental Method

The Fundamental Method is a collection of interlocking tools and constructions which when used collectively provide powerful, highly customizable constructions of points in kG. Roughly speaking, the Fundamental Method gives you a partial function c with domain a subset of G and range {0, 1, . . . , k − 1} and a sequence (∆n) of subsets of G such that

1 For each n, ∆n is uniformly spread out in G; 2 For each n there is a finite set F such that the behavoir of c on gF

determines if g ∈ ∆n;

3 Each δ ∈ ∆n owns nearby proprietary points on which c is undefined; Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 10 / 16

slide-55
SLIDE 55

The Fundamental Method

The Fundamental Method is a collection of interlocking tools and constructions which when used collectively provide powerful, highly customizable constructions of points in kG. Roughly speaking, the Fundamental Method gives you a partial function c with domain a subset of G and range {0, 1, . . . , k − 1} and a sequence (∆n) of subsets of G such that

1 For each n, ∆n is uniformly spread out in G; 2 For each n there is a finite set F such that the behavoir of c on gF

determines if g ∈ ∆n;

3 Each δ ∈ ∆n owns nearby proprietary points on which c is undefined; 4 The location, relative to δ, of these nearby proprietary points depends

  • nly on n;

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 10 / 16

slide-56
SLIDE 56

The Fundamental Method

The Fundamental Method is a collection of interlocking tools and constructions which when used collectively provide powerful, highly customizable constructions of points in kG. Roughly speaking, the Fundamental Method gives you a partial function c with domain a subset of G and range {0, 1, . . . , k − 1} and a sequence (∆n) of subsets of G such that

1 For each n, ∆n is uniformly spread out in G; 2 For each n there is a finite set F such that the behavoir of c on gF

determines if g ∈ ∆n;

3 Each δ ∈ ∆n owns nearby proprietary points on which c is undefined; 4 The location, relative to δ, of these nearby proprietary points depends

  • nly on n;

5 The number of proprietary points can be made large. Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 10 / 16

slide-57
SLIDE 57

The Fundamental Method

The Fundamental Method is a collection of interlocking tools and constructions which when used collectively provide powerful, highly customizable constructions of points in kG. Roughly speaking, the Fundamental Method gives you a partial function c with domain a subset of G and range {0, 1, . . . , k − 1} and a sequence (∆n) of subsets of G such that

1 For each n, ∆n is uniformly spread out in G; 2 For each n there is a finite set F such that the behavoir of c on gF

determines if g ∈ ∆n;

3 Each δ ∈ ∆n owns nearby proprietary points on which c is undefined; 4 The location, relative to δ, of these nearby proprietary points depends

  • nly on n;

5 The number of proprietary points can be made large.

Using the undefined points of c, one can encode data and extend c to c′ ∈ kG. The properties above imply that this data is uniquely readable from c′.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 10 / 16

slide-58
SLIDE 58

Construction of k-Colorings

1 Use the fundamental method (c, (∆n)). Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 11 / 16

slide-59
SLIDE 59

Construction of k-Colorings

1 Use the fundamental method (c, (∆n)). 2 Enumerate G \ {1G}: s1, s2, . . .. Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 11 / 16

slide-60
SLIDE 60

Construction of k-Colorings

1 Use the fundamental method (c, (∆n)). 2 Enumerate G \ {1G}: s1, s2, . . .. 3 Find functions Ln : ∆n → {0, . . . , k − 1}m(n) so that nearby points of

∆n have distinct Ln values.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 11 / 16

slide-61
SLIDE 61

Construction of k-Colorings

1 Use the fundamental method (c, (∆n)). 2 Enumerate G \ {1G}: s1, s2, . . .. 3 Find functions Ln : ∆n → {0, . . . , k − 1}m(n) so that nearby points of

∆n have distinct Ln values.

4 Use undefined points of c to encode the values of the Ln’s. This gives

us x ∈ kG.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 11 / 16

slide-62
SLIDE 62

Construction of k-Colorings

1 Use the fundamental method (c, (∆n)). 2 Enumerate G \ {1G}: s1, s2, . . .. 3 Find functions Ln : ∆n → {0, . . . , k − 1}m(n) so that nearby points of

∆n have distinct Ln values.

4 Use undefined points of c to encode the values of the Ln’s. This gives

us x ∈ kG.

5 Fix sn and consider g ∈ G. Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 11 / 16

slide-63
SLIDE 63

Construction of k-Colorings

1 Use the fundamental method (c, (∆n)). 2 Enumerate G \ {1G}: s1, s2, . . .. 3 Find functions Ln : ∆n → {0, . . . , k − 1}m(n) so that nearby points of

∆n have distinct Ln values.

4 Use undefined points of c to encode the values of the Ln’s. This gives

us x ∈ kG.

5 Fix sn and consider g ∈ G. 6 ∆n is uniformly spread out, so there is u ∈ G not too big with

gu ∈ ∆n.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 11 / 16

slide-64
SLIDE 64

Construction of k-Colorings

1 Use the fundamental method (c, (∆n)). 2 Enumerate G \ {1G}: s1, s2, . . .. 3 Find functions Ln : ∆n → {0, . . . , k − 1}m(n) so that nearby points of

∆n have distinct Ln values.

4 Use undefined points of c to encode the values of the Ln’s. This gives

us x ∈ kG.

5 Fix sn and consider g ∈ G. 6 ∆n is uniformly spread out, so there is u ∈ G not too big with

gu ∈ ∆n.

7 We are done if gsnu ∈ ∆n. Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 11 / 16

slide-65
SLIDE 65

Construction of k-Colorings

1 Use the fundamental method (c, (∆n)). 2 Enumerate G \ {1G}: s1, s2, . . .. 3 Find functions Ln : ∆n → {0, . . . , k − 1}m(n) so that nearby points of

∆n have distinct Ln values.

4 Use undefined points of c to encode the values of the Ln’s. This gives

us x ∈ kG.

5 Fix sn and consider g ∈ G. 6 ∆n is uniformly spread out, so there is u ∈ G not too big with

gu ∈ ∆n.

7 We are done if gsnu ∈ ∆n. 8 If gsnu ∈ ∆n, then gu and gsnu are close and hence have distinct Ln

values.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 11 / 16

slide-66
SLIDE 66

Construction of k-Colorings

1 Use the fundamental method (c, (∆n)). 2 Enumerate G \ {1G}: s1, s2, . . .. 3 Find functions Ln : ∆n → {0, . . . , k − 1}m(n) so that nearby points of

∆n have distinct Ln values.

4 Use undefined points of c to encode the values of the Ln’s. This gives

us x ∈ kG.

5 Fix sn and consider g ∈ G. 6 ∆n is uniformly spread out, so there is u ∈ G not too big with

gu ∈ ∆n.

7 We are done if gsnu ∈ ∆n. 8 If gsnu ∈ ∆n, then gu and gsnu are close and hence have distinct Ln

values.

9 Since the Ln’s can be “decoded” from x, there is a small v ∈ G with

x(gsnuv) = x(guv).

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 11 / 16

slide-67
SLIDE 67

Stronger Results

Recall that a subflow is minimal if every orbit is dense.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 12 / 16

slide-68
SLIDE 68

Stronger Results

Recall that a subflow is minimal if every orbit is dense. Theorem (Gao–Jackson–S) If U ⊆ kG is open and non-empty then there are continuum-many pairwise disjoint minimal free subflows intersecting U.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 12 / 16

slide-69
SLIDE 69

Stronger Results

Recall that a subflow is minimal if every orbit is dense. Theorem (Gao–Jackson–S) If U ⊆ kG is open and non-empty then there are continuum-many pairwise disjoint minimal free subflows intersecting U. The above theorem says in particular that if A ⊆ G is finite and y : A → {0, . . . , k − 1} is any function, then there are continuum-many k-colorings which extend y and have pairwise disjoint orbit closures.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 12 / 16

slide-70
SLIDE 70

Stronger Results

Recall that a subflow is minimal if every orbit is dense. Theorem (Gao–Jackson–S) If U ⊆ kG is open and non-empty then there are continuum-many pairwise disjoint minimal free subflows intersecting U. The above theorem says in particular that if A ⊆ G is finite and y : A → {0, . . . , k − 1} is any function, then there are continuum-many k-colorings which extend y and have pairwise disjoint orbit closures. Theorem (Gao–Jackson–S) For A ⊆ G the following are equivalent:

1 There is a finite T ⊆ G so that for all g ∈ G there is t ∈ T with

gt ∈ A;

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 12 / 16

slide-71
SLIDE 71

Stronger Results

Recall that a subflow is minimal if every orbit is dense. Theorem (Gao–Jackson–S) If U ⊆ kG is open and non-empty then there are continuum-many pairwise disjoint minimal free subflows intersecting U. The above theorem says in particular that if A ⊆ G is finite and y : A → {0, . . . , k − 1} is any function, then there are continuum-many k-colorings which extend y and have pairwise disjoint orbit closures. Theorem (Gao–Jackson–S) For A ⊆ G the following are equivalent:

1 There is a finite T ⊆ G so that for all g ∈ G there is t ∈ T with

gt ∈ A;

2 For every y ∈ kA there is a k-coloring x extending y; Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 12 / 16

slide-72
SLIDE 72

Stronger Results

Recall that a subflow is minimal if every orbit is dense. Theorem (Gao–Jackson–S) If U ⊆ kG is open and non-empty then there are continuum-many pairwise disjoint minimal free subflows intersecting U. The above theorem says in particular that if A ⊆ G is finite and y : A → {0, . . . , k − 1} is any function, then there are continuum-many k-colorings which extend y and have pairwise disjoint orbit closures. Theorem (Gao–Jackson–S) For A ⊆ G the following are equivalent:

1 There is a finite T ⊆ G so that for all g ∈ G there is t ∈ T with

gt ∈ A;

2 For every y ∈ kA there is a k-coloring x extending y; 3 For every y ∈ kA there are continuum-many k-colorings which extend

y and have pairwise disjoint orbit closures.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 12 / 16

slide-73
SLIDE 73

Descriptive Complexity

A set is Σ0

2 if it is the countable union of closed sets (i.e. Fσ). A set is Π0 3

if it is the countable intersection of Σ0

2 sets.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 13 / 16

slide-74
SLIDE 74

Descriptive Complexity

A set is Σ0

2 if it is the countable union of closed sets (i.e. Fσ). A set is Π0 3

if it is the countable intersection of Σ0

2 sets.

Let X be a Polish space and let A ⊆ X be Σ0

  • 2. Recall that A is

Σ0

2-complete if for every Polish space Y and every Σ0 2 set B ⊆ Y there is a

continuous function Y → X with B = f −1(A). A similar definition applies to Π0

3-complete.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 13 / 16

slide-75
SLIDE 75

Descriptive Complexity

A set is Σ0

2 if it is the countable union of closed sets (i.e. Fσ). A set is Π0 3

if it is the countable intersection of Σ0

2 sets.

Let X be a Polish space and let A ⊆ X be Σ0

  • 2. Recall that A is

Σ0

2-complete if for every Polish space Y and every Σ0 2 set B ⊆ Y there is a

continuous function Y → X with B = f −1(A). A similar definition applies to Π0

3-complete.

We call G flecc if there is a finite A ⊆ G \ {1G} such that every g ∈ G has a power which is conjugate to an element of A, i.e. ∀g ∈ G ∃n ∈ N ∃h ∈ G hgnh−1 ∈ A.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 13 / 16

slide-76
SLIDE 76

Descriptive Complexity

A set is Σ0

2 if it is the countable union of closed sets (i.e. Fσ). A set is Π0 3

if it is the countable intersection of Σ0

2 sets.

Let X be a Polish space and let A ⊆ X be Σ0

  • 2. Recall that A is

Σ0

2-complete if for every Polish space Y and every Σ0 2 set B ⊆ Y there is a

continuous function Y → X with B = f −1(A). A similar definition applies to Π0

3-complete.

We call G flecc if there is a finite A ⊆ G \ {1G} such that every g ∈ G has a power which is conjugate to an element of A, i.e. ∀g ∈ G ∃n ∈ N ∃h ∈ G hgnh−1 ∈ A. Theorem (Gao–Jackson–S) If G is flecc then the set of k-colorings is Σ0

2-complete. If G is not flecc

then the set of k-colorings is Π0

3-complete.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 13 / 16

slide-77
SLIDE 77

The Topological Conjugacy Relation

Two subflows X, Y ⊆ kG are topologically conjugate if there is a homeomorphism φ : X → Y such that φ(g · x) = g · φ(x) for every g ∈ G and every x ∈ X. Let TC(kG) denote the topological conjugacy relation

  • n subflows of kG.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 14 / 16

slide-78
SLIDE 78

The Topological Conjugacy Relation

Two subflows X, Y ⊆ kG are topologically conjugate if there is a homeomorphism φ : X → Y such that φ(g · x) = g · φ(x) for every g ∈ G and every x ∈ X. Let TC(kG) denote the topological conjugacy relation

  • n subflows of kG.

Let S(kG) be the set of all subflows of kG, let SF(kG) be the set of all free subflows, and SMF(kG) be the set of all minimal free subflows.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 14 / 16

slide-79
SLIDE 79

The Topological Conjugacy Relation

Two subflows X, Y ⊆ kG are topologically conjugate if there is a homeomorphism φ : X → Y such that φ(g · x) = g · φ(x) for every g ∈ G and every x ∈ X. Let TC(kG) denote the topological conjugacy relation

  • n subflows of kG.

Let S(kG) be the set of all subflows of kG, let SF(kG) be the set of all free subflows, and SMF(kG) be the set of all minimal free subflows. With the Vietoris topology (or equivalently the Hausdorff distance), S(kG) and SF(kG) are Polish. Furthermore, SMF(kG) is a Borel subset of S(kG) and is thus a standard Borel space.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 14 / 16

slide-80
SLIDE 80

The Topological Conjugacy Relation

Two subflows X, Y ⊆ kG are topologically conjugate if there is a homeomorphism φ : X → Y such that φ(g · x) = g · φ(x) for every g ∈ G and every x ∈ X. Let TC(kG) denote the topological conjugacy relation

  • n subflows of kG.

Let S(kG) be the set of all subflows of kG, let SF(kG) be the set of all free subflows, and SMF(kG) be the set of all minimal free subflows. With the Vietoris topology (or equivalently the Hausdorff distance), S(kG) and SF(kG) are Polish. Furthermore, SMF(kG) is a Borel subset of S(kG) and is thus a standard Borel space. Recall that an equivalence relation is countable if every equivalence class is countable.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 14 / 16

slide-81
SLIDE 81

The Topological Conjugacy Relation

Two subflows X, Y ⊆ kG are topologically conjugate if there is a homeomorphism φ : X → Y such that φ(g · x) = g · φ(x) for every g ∈ G and every x ∈ X. Let TC(kG) denote the topological conjugacy relation

  • n subflows of kG.

Let S(kG) be the set of all subflows of kG, let SF(kG) be the set of all free subflows, and SMF(kG) be the set of all minimal free subflows. With the Vietoris topology (or equivalently the Hausdorff distance), S(kG) and SF(kG) are Polish. Furthermore, SMF(kG) is a Borel subset of S(kG) and is thus a standard Borel space. Recall that an equivalence relation is countable if every equivalence class is countable. Lemma (Clemens, 2009) TC and its restrictions to SF(kG) and SMF(kG) are countable Borel equivalence relations.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 14 / 16

slide-82
SLIDE 82

The Topological Conjugacy Relation

For two Borel equivalence relations E and F on X and Y , we say E is reducible to F if there is a Borel function f : X → Y such that x1 E x2 ⇐ ⇒ f (x1) F f (x2). Intuitively, F is more complicated than E if E is reducible to F.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 15 / 16

slide-83
SLIDE 83

The Topological Conjugacy Relation

For two Borel equivalence relations E and F on X and Y , we say E is reducible to F if there is a Borel function f : X → Y such that x1 E x2 ⇐ ⇒ f (x1) F f (x2). Intuitively, F is more complicated than E if E is reducible to F. We call E smooth if E is reducible to the equality equivalence relation.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 15 / 16

slide-84
SLIDE 84

The Topological Conjugacy Relation

For two Borel equivalence relations E and F on X and Y , we say E is reducible to F if there is a Borel function f : X → Y such that x1 E x2 ⇐ ⇒ f (x1) F f (x2). Intuitively, F is more complicated than E if E is reducible to F. We call E smooth if E is reducible to the equality equivalence relation. E is a universal countable Borel equivalence relation if E is countable and every other countable Borel equivalence relation is reducible to E.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 15 / 16

slide-85
SLIDE 85

The Topological Conjugacy Relation

For two Borel equivalence relations E and F on X and Y , we say E is reducible to F if there is a Borel function f : X → Y such that x1 E x2 ⇐ ⇒ f (x1) F f (x2). Intuitively, F is more complicated than E if E is reducible to F. We call E smooth if E is reducible to the equality equivalence relation. E is a universal countable Borel equivalence relation if E is countable and every other countable Borel equivalence relation is reducible to E. Theorem (Clemens, 2009) TC(kZn) is a universal countable Borel equivalence relation.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 15 / 16

slide-86
SLIDE 86

The Topological Conjugacy Relation

For two Borel equivalence relations E and F on X and Y , we say E is reducible to F if there is a Borel function f : X → Y such that x1 E x2 ⇐ ⇒ f (x1) F f (x2). Intuitively, F is more complicated than E if E is reducible to F. We call E smooth if E is reducible to the equality equivalence relation. E is a universal countable Borel equivalence relation if E is countable and every other countable Borel equivalence relation is reducible to E. Theorem (Clemens, 2009) TC(kZn) is a universal countable Borel equivalence relation. Theorem (Gao–Jackson–S; Clemens) If G is infinite and locally finite, then TC(kG) and its restriction to SF(kG) are non-smooth and hyperfinite. If G is not locally finite, then TC(kG) and its restriction to SF(kG) are universal countable Borel equivalence relations.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 15 / 16

slide-87
SLIDE 87

The Topological Conjugacy Relation

Theorem (Gao–Jackson–S) For any countably infinite group G, the restriction of TC(kG) to SMF(kG) is non-smooth.

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 16 / 16

slide-88
SLIDE 88

The Topological Conjugacy Relation

Theorem (Gao–Jackson–S) For any countably infinite group G, the restriction of TC(kG) to SMF(kG) is non-smooth. Open Question What is the complexity of the restriction of TC(kZ) to SMF(kZ)?

Brandon Seward () Group Colorings and Bernoulli Subflows Logic & Dynamics Conf. 2012 16 / 16