The (λ, κ)-fn and the order theory of bases in boolean algebras

Similar documents
ADDING A LOT OF COHEN REALS BY ADDING A FEW II. 1. Introduction

Axiomatization of generic extensions by homogeneous partial orderings

THE NUMBER OF UNARY CLONES CONTAINING THE PERMUTATIONS ON AN INFINITE SET

is not generally true for the local Noetherian type of p, as shown by a counterexample where χ(p, X) is singular.

being saturated Lemma 0.2 Suppose V = L[E]. Every Woodin cardinal is Woodin with.

DEPTH OF BOOLEAN ALGEBRAS SHIMON GARTI AND SAHARON SHELAH

MITCHELL S THEOREM REVISITED. Contents

Continuous images of closed sets in generalized Baire spaces ESI Workshop: Forcing and Large Cardinals

Philipp Moritz Lücke

PARTITIONS OF 2 ω AND COMPLETELY ULTRAMETRIZABLE SPACES

A relative of the approachability ideal, diamond and non-saturation

Silver type theorems for collapses.

UPWARD STABILITY TRANSFER FOR TAME ABSTRACT ELEMENTARY CLASSES

Generalising the weak compactness of ω

arxiv:math/ v1 [math.lo] 15 Jan 1991

Determinacy models and good scales at singular cardinals

CARDINALITIES OF RESIDUE FIELDS OF NOETHERIAN INTEGRAL DOMAINS

STRONGLY UNFOLDABLE CARDINALS MADE INDESTRUCTIBLE

The Semi-Weak Square Principle

Bounds on coloring numbers

Set- theore(c methods in model theory

Extender based forcings, fresh sets and Aronszajn trees

Open Problems. Problem 2. Assume PD. C 3 is the largest countable Π 1 3-set of reals. Is it true that C 3 = {x M 2 R x is. Known:

Strongly compact Magidor forcing.

Chromatic number of infinite graphs

Interpolation of κ-compactness and PCF

PERFECT TREE FORCINGS FOR SINGULAR CARDINALS

On the strengths and weaknesses of weak squares

Characterizing large cardinals in terms of layered partial orders

On almost precipitous ideals.

On Singular Stationarity II (tight stationarity and extenders-based methods)

A precipitous club guessing ideal on ω 1

A HIERARCHY OF RAMSEY-LIKE CARDINALS

On Singular Stationarity I (mutual stationarity and ideal-based methods)

Währinger Strasse 25, 1090 Vienna Austria

Cardinal arithmetic: The Silver and Galvin-Hajnal Theorems

FORCING AND THE HALPERN-LÄUCHLI THEOREM. 1. Introduction This document is a continuation of [1]. It is intended to be part of a larger paper.

GUESSING MODELS IMPLY THE SINGULAR CARDINAL HYPOTHESIS arxiv: v1 [math.lo] 25 Mar 2019

Sy D. Friedman. August 28, 2001

Chain conditions, layered partial orders and weak compactness

Strongly Unfoldable Cardinals Made Indestructible

Math 280B Winter Recursion on Well-Founded Relations. 6.1 Recall: For a binary relation R (may be a proper class): T 0 = A T n+1 = pred R (a)

6. Recursion on Well-Founded Relations

On the Splitting Number at Regular Cardinals

COMBINATORICS OF REDUCTIONS BETWEEN EQUIVALENCE RELATIONS

MODIFIED EXTENDER BASED FORCING

New tools of set-theoretic homological algebra and their applications to modules

Notes on getting presaturation from collapsing a Woodin cardinal

Combinatorics, Cardinal Characteristics of the Continuum, and the Colouring Calculus

Generic embeddings associated to an indestructibly weakly compact cardinal

Tall, Strong, and Strongly Compact Cardinals

HEIKE MILDENBERGER AND SAHARON SHELAH

SOME CONSEQUENCES OF REFLECTION ON THE APPROACHABILITY IDEAL

Two Stationary Sets with Different Gaps of the Power Function

1. Introduction. As part of his study of functions defined on product spaces, M. Hušek introduced a family of diagonal conditions in a topological

On almost precipitous ideals.

Short Extenders Forcings II

2. The ultrapower construction

COMBINATORICS AT ℵ ω

Large Cardinals with Few Measures

Generalization by Collapse

Commentationes Mathematicae Universitatis Carolinae

ON THE QUOTIENT SHAPES OF VECTORIAL SPACES. Nikica Uglešić

Level by Level Inequivalence, Strong Compactness, and GCH

LOCAL CLUB CONDENSATION AND L-LIKENESS

3 The Model Existence Theorem

SUCCESSIVE FAILURES OF APPROACHABILITY

January 28, 2013 EASTON S THEOREM FOR RAMSEY AND STRONGLY RAMSEY CARDINALS

DIAGONAL PRIKRY EXTENSIONS

AN INFINITE CARDINAL-VALUED KRULL DIMENSION FOR RINGS

COLLAPSING SUCCESSORS OF SINGULARS

Notes to The Resurrection Axioms

BLOWING UP POWER OF A SINGULAR CARDINAL WIDER GAPS

LARGE CARDINALS AND L-LIKE UNIVERSES

ARONSZAJN TREES AND THE SUCCESSORS OF A SINGULAR CARDINAL. 1. Introduction

The first author was supported by FWF Project P23316-N13.

CONSECUTIVE SINGULAR CARDINALS AND THE CONTINUUM FUNCTION

Covering properties of derived models

UC Irvine UC Irvine Electronic Theses and Dissertations

The tree property for supercompactness

INDESTRUCTIBLE STRONG UNFOLDABILITY

Fat subsets of P kappa (lambda)

SHIMON GARTI AND SAHARON SHELAH

Global singularization and the failure of SCH

The Resurrection Axioms

Easton s theorem and large cardinals from the optimal hypothesis

arxiv: v3 [math.lo] 23 Jul 2018

LECTURE NOTES - ADVANCED TOPICS IN MATHEMATICAL LOGIC

Annals of Pure and Applied Logic

A Laver-like indestructibility for hypermeasurable cardinals

The Outer Model Programme

Building Infinite Processes from Regular Conditional Probability Distributions

An effective perfect-set theorem

Gödel algebras free over finite distributive lattices

Chapter 4. Cardinal Arithmetic.

arxiv: v1 [math.lo] 26 Mar 2014

Preservation theorems for Namba forcing

RVM, RVC revisited: Clubs and Lusin sets

THE FIRST MEASURABLE CARDINAL CAN BE THE FIRST UNCOUNTABLE REGULAR CARDINAL AT ANY SUCCESSOR HEIGHT

ANNALES ACADEMIÆ SCIENTIARUM FENNICÆ DIAMONDS ON LARGE CARDINALS

Transcription:

The (λ, κ)-fn and the order theory of bases in boolean algebras David Milovich Texas A&M International University david.milovich@tamiu.edu http://www.tamiu.edu/ dmilovich/ June 2, 2010 BLAST 1 / 22

The κ-freese-nation property Definition (Fuchino, Koppelberg, Shelah) A boolean algebra B has the κ-fn iff there is a κ-fn map f : B [B] <κ, i.e., {p q} B r f (p) f (q) p r q. The FN is the ω-fn. The weak FN or WFN is the ω 1 -FN. 2 / 22

The κ-freese-nation property Definition (Fuchino, Koppelberg, Shelah) A boolean algebra B has the κ-fn iff there is a κ-fn map f : B [B] <κ, i.e., {p q} B r f (p) f (q) p r q. The FN is the ω-fn. The weak FN or WFN is the ω 1 -FN. Basic facts If B κ, then B has the κ-fn. The κ-fn is preserved by coproducts. In particular, free boolean algebras have the FN. The interval algebra of κ + lacks the κ-fn. 2 / 22

The κ-fn in terms of elementary submodels Let κ = cf κ ω. Let θ = cf θ be sufficiently large. H θ = {x : x, x, x,... < θ}. Small-model version (FKS) B has the κ-fn iff, for every p B and every M with M = κ M and B M H θ, we have cf(m p) < κ and ci(m p) < κ. Big-model version B has the κ-fn iff, for every p B and every M with κ M and B M H θ, we have cf(m p) < κ and ci(m p) < κ. 3 / 22

Between the FN and the κ-fn Assume λ = cf λ > κ = cf κ ω. Small-model definition A boolean algebra B has the (λ, κ)-fn iff, for every p B and every M with λ M λ > M and B M H θ, we have cf(m p) < κ and ci(m p) < κ. Equivalent big-model definition A boolean algebra B has the (λ, κ)-fn iff, for every p B and every M with λ M or λ M λ > M and B M H θ, we have cf(m p) < κ and ci(m p) < κ. 4 / 22

Properties of the (λ, κ)-fn Every boolean algebra with cardinality < λ has the (λ, ω)-fn, which implies the (λ, κ)-fn for all regular κ [ω, λ). The (λ, κ)-fn is preserved by coproducts. 5 / 22

Properties of the (λ, κ)-fn Every boolean algebra with cardinality < λ has the (λ, ω)-fn, which implies the (λ, κ)-fn for all regular κ [ω, λ). The (λ, κ)-fn is preserved by coproducts. The κ-fn is equivalent to the (κ +, κ)-fn. In particular, the FN and WFN are equivalent to the (ω 1, ω)-fn and (ω 2, ω 1 )-FN. The algebra of (< κ)-supported, constructible subsets of λ 2 has the (λ, κ)-fn, but lacks the (λ, µ)-fn for all regular µ < κ. Hence, the implications κ-fn (κ +, ω)-fn FN are all strict if κ > ω. (E.g., WFN (ω 2, ω)-fn FN.) 5 / 22

Properties of the (λ, κ)-fn Every boolean algebra with cardinality < λ has the (λ, ω)-fn, which implies the (λ, κ)-fn for all regular κ [ω, λ). The (λ, κ)-fn is preserved by coproducts. The κ-fn is equivalent to the (κ +, κ)-fn. In particular, the FN and WFN are equivalent to the (ω 1, ω)-fn and (ω 2, ω 1 )-FN. The algebra of (< κ)-supported, constructible subsets of λ 2 has the (λ, κ)-fn, but lacks the (λ, µ)-fn for all regular µ < κ. Hence, the implications κ-fn (κ +, ω)-fn FN are all strict if κ > ω. (E.g., WFN (ω 2, ω)-fn FN.) If λ 2 κ, then P(κ) and P(κ)/[κ] <κ lack the (λ, κ)-fn, so 2 ω = ω 2 does not decide whether P(ω) and P(ω)/[ω] <ω have the WFN, but it does imply they lack the (ω 2, ω)-fn. 5 / 22

Elementary quotients Definition Given a compactum X and X M H θ, define the quotient map πm X : X X /M by p/m = q/m iff f (p) = f (q) for all f C(X, R) M. Basic facts and examples The sets of the form U/M where U is open F σ and U M form a base of X /M. (Also, (U/M) = U for such U.) If X is a 0-dimensional compactum, then X /M = Ult(Clop(X ) M). If X is a compactum, X M H θ, and M is countable, then X /M is metrizable. If X is a compact metric space and X M H θ, then X /M = X. If X = ω 1 + 1 and M H θ is countable, then X /M = δ + 1 where δ = ω 1 M. 6 / 22

The Stone dual of the FN A map between spaces is open if open sets have open images. Small-model definition (Ščepin, Bandlow). A compactum X is said to be openly generated iff πm X is an open map for a club of M in [H θ ] ω. Big-model definition. A compactum X is openly generated iff π X M is an open map for all M H θ with X M. A 0-dimensional compactum X is openly generated iff Clop(X ) has the FN. Example: If X = ω 1 + 1, M H θ is countable, and S is the set of countable successor ordinals, then S is open but S/M is not open, so πm X is not open. 7 / 22

Which compacta are openly generated? Compare being openly generated to being a continuous image of a product of metrizable compacta (i.e., dyadic): Theorem (Ščepin) The class of openly generated compacta includes all metrizable compacta and is closed with respect to products and hyperspaces (with the Vietoris topology). Theorem (Šapiro) The hyperspace exp( ω 2 2) is not a continuous image of a product of metrizable compacta. Theorem (Engelking, Ščepin) The quotient of ω 1 2 formed by identifying 0 i<ω1 and 1 i<ω1 is not openly generated. 8 / 22

Stone dual of the (λ, κ)-fn A map f : X Y is said to be κ-open if, for all open O X, f [O] is the intersection of (< κ)-many open sets. Hence, the ω-open maps are exactly the open maps. Small-model definition. A compactum X is (λ, κ)-openly generated iff π X M is κ-open for a club of M in [H θ] <λ. Big-model definition. A compactum X is (λ, κ)-openly generated iff π X M is κ-open for all M H θ with X M and either λ M or M λ λ > M. A 0-dimensional compactum X is (λ, κ)-openly generated iff Clop(X ) has the (λ, κ)-fn. 9 / 22

Which compacta are (λ, κ)-openly generated? Compare being (λ, κ)-openly generated to being a continuous image of a product of small factors: Theorem The class of (λ, κ)-openly generated compacta includes all compacta with weight < λ and is closed with respect to products and hyperspaces. Theorem (Šapiro) ( ) The hyperspace exp λ + 2 is not a continuous image of a product whose factors are all compacta with weight < λ. Theorem The quotient of λ 2 formed by identifying 0 i<λ and 1 i<λ is not (λ, κ)-openly generated. 10 / 22

Bases A family of open sets is called a base iff every open set is a union of sets from the family. Hence, a family of clopen sets in a 0-dimensional compactum is a base iff every clopen set is a finite union of sets from the family. A subset of a boolean algebra is called a base iff every element of the algebra is a finite join of elements of the subset. The weight w(x ) of a infinite T 0 space X is min{ E : E is a base}. The weight of an infinite boolean algebra B is just B. 11 / 22

Local π-bases A family of nonempty open subsets of a space is called a local π-base at a point iff every neighborhood of that point contains an element of the family. Hence, a family of nonempty clopen subsets of a 0-dimensional compactum is a local π-base iff every clopen neighborhood of the point contains an element of the family.. A subset of a boolean algebra is called a local π-base at an ultrafilter iff every element of the subset is 0, and everything in the ultrafilter is something in the subset. The π-character πχ(p, X ) of a point p is min{ E : E is a local π-base at p}. The π-character πχ(u, B) of an ultrafilter is min{ S : S is a local π-base at U}. 12 / 22

Continuous images and subalgebras Theorem (Ščepin) Assuming that: Y is an infinite compactum and a continuous image of an openly generated compactum X, and B is an infinite subalgebra of a boolean algebra A where A has the FN, it follows that: w(y ) = sup p Y πχ(p, Y ) and B = sup U Ult(B) πχ(u, B), and every regular uncountable cardinal is a caliber of Y and a precaliber of B. Definition A regular cardinal ν is a caliber (precaliber) of a space (boolean algebra) if every ν-sized open family (subset) has a ν-sized subset that contains a common point (extends to a proper filter). 13 / 22

Continuous images and subalgebras, part II Theorem Assuming that: Y is a compactum and a continuous image of a (λ, ω)-openly generated compactum X, B is a subalgebra of a boolean algebra A where A has the (λ, ω)-fn, and w(y ) λ and B λ, it follows that: w(y ) = sup p Y πχ(p, Y ) and B = sup U Ult(B) πχ(u, B), and every regular µ λ is a caliber of Y and a precaliber of B. 14 / 22

Continuous images and subalgebras, part II Theorem Assuming that: Y is a compactum and a continuous image of a (λ, ω)-openly generated compactum X, B is a subalgebra of a boolean algebra A where A has the (λ, ω)-fn, and w(y ) λ and B λ, it follows that: w(y ) = sup p Y πχ(p, Y ) and B = sup U Ult(B) πχ(u, B), and every regular µ λ is a caliber of Y and a precaliber of B. Question. If κ = κ <κ, can we replace λ and ω in the theorem with κ + and κ? Remark. The real interval algebra has size 2 ω and the (ω 2, ω 1 )-FN, yet all its ultrafilters have π-character ω. 14 / 22

Flat vs. top-heavy Definition A cone in a poset P is a set of the form p = {q P : q p} where p P. A poset is κ-top-heavy if some cone has size κ. A poset is κ op -like if every cone has size < κ. Example. {1/(n + 1) : n ω} is ω op -like. Example. ω is ω op 1 -like and ω-top-heavy. 15 / 22

Flat vs. top-heavy Definition A cone in a poset P is a set of the form p = {q P : q p} where p P. A poset is κ-top-heavy if some cone has size κ. A poset is κ op -like if every cone has size < κ. Example. {1/(n + 1) : n ω} is ω op -like. Example. ω is ω op 1 -like and ω-top-heavy. Convention. A family of subsets of a topological space is ordered by. Theorem (joint with Spadaro) If κ is regular and uncountable, X is a compactum with weight κ, and X has a dense set of points p with πχ(p, X ) < κ, then X does not have a κ op -like base. 15 / 22

Flatness from the (µ +, cf µ)-fn Theorem Assuming that: A has the (µ +, cf µ)-fn, S is a subset of A, B is a subalgebra of A, and E is a base of B, it follows that: S has a dense subset that is µ op -like, and, if also πχ(u, B) = B for all U Ult(B), then E includes a µ op -like base F of B. Remark If µ is regular, then the (µ +, cf µ)-fn is equivalent to the µ-fn. 16 / 22

Flatness from (µ +, cf µ)-open generation Theorem. Assuming that: X is a (µ +, cf µ)-openly generated compactum, Y is a compactum and a continuous image of X, E is a base of Y, and S P(Y ) is such that the interior of every U S includes the closure of some V S, it follows that: S has a dense subset that is µ op -like, and, if also πχ(p, Y ) = w(y ) for all p Y, then E includes a µ op -like base F of Y. Corollary. A base of a compact group always includes an ω op -like base of the group. Question. Are all homogeneous compacta ((2 ω ) +, ω)-openly generated? 17 / 22

Comments on proving the last two theorems All those big-model definitions were (re?)invented in order to prove the last two theorems. The construction of the µ op -like dense sets uses continuous elementary chains of submodels and induction on the weight of the space or boolean algebra. On the other hand, each successor stage of the construction of µ op -like bases uses special properties of countable boolean algebras and metrizable compacta to build a countable piece of the base. To perform constructions of length ω 2 one countable piece at a time, I used a generalization of Jackson-Mauldin trees of elementary submodels (which they used to build a Steinhaus set without assuming CH). 18 / 22

Long κ-approximation sequences Assume κ = cf κ > ω. Let Ω κ denote the tree of finite sequences of ordinals ξ i i<n which satisfy κ ξ i > ξ j for all {i < j} n. 19 / 22

Long κ-approximation sequences Assume κ = cf κ > ω. Let Ω κ denote the tree of finite sequences of ordinals ξ i i<n which satisfy κ ξ i > ξ j for all {i < j} n. The is a unique order isomorphism Υ κ from the ordinals to Ω κ ordered lexicographically. Thus, κ = Υ κ ( Ω κ ) is a {κ}-definable tree-ordering of the ordinals such that α κ β α < β and all branches are finite. 19 / 22

Long κ-approximation sequences Assume κ = cf κ > ω. Let Ω κ denote the tree of finite sequences of ordinals ξ i i<n which satisfy κ ξ i > ξ j for all {i < j} n. The is a unique order isomorphism Υ κ from the ordinals to Ω κ ordered lexicographically. Thus, κ = Υ κ ( Ω κ ) is a {κ}-definable tree-ordering of the ordinals such that α κ β α < β and all branches are finite. A long κ-approximation sequence is a sequence M α α<η or arbitrary ordinal length satisfying κ, M β β<α M α H θ and κ M α κ > M α. 19 / 22

Long κ-approximation sequences Assume κ = cf κ > ω. Let Ω κ denote the tree of finite sequences of ordinals ξ i i<n which satisfy κ ξ i > ξ j for all {i < j} n. The is a unique order isomorphism Υ κ from the ordinals to Ω κ ordered lexicographically. Thus, κ = Υ κ ( Ω κ ) is a {κ}-definable tree-ordering of the ordinals such that α κ β α < β and all branches are finite. A long κ-approximation sequence is a sequence M α α<η or arbitrary ordinal length satisfying κ, M β β<α M α H θ and κ M α κ > M α. If α η and {β : β κ α} = {β 0 κ κ β m }, then N i = {M γ : β i γ < β i+1 } satisfies N i N i H θ. 19 / 22

Long κ-approximation sequences Assume κ = cf κ > ω. Let Ω κ denote the tree of finite sequences of ordinals ξ i i<n which satisfy κ ξ i > ξ j for all {i < j} n. The is a unique order isomorphism Υ κ from the ordinals to Ω κ ordered lexicographically. Thus, κ = Υ κ ( Ω κ ) is a {κ}-definable tree-ordering of the ordinals such that α κ β α < β and all branches are finite. A long κ-approximation sequence is a sequence M α α<η or arbitrary ordinal length satisfying κ, M β β<α M α H θ and κ M α κ > M α. If α η and {β : β κ α} = {β 0 κ κ β m }, then N i = {M γ : β i γ < β i+1 } satisfies N i N i H θ. Thus, long ω 1 -approximation sequences allow one to construct things by adding one new countable piece per stage, each time collecting all the old pieces into a finite union of possibly uncountable elementary substructures. 19 / 22

Using long approximation sequences For constructing the µ op -like bases, a single long ω 1 -approximation sequence M does the job if µ = ω. If µ > ω, then the construction uses a long µ + -approximation sequence M built such that each M α is the union of an ω 1 -approximation sequence N α of length µ with M β β<α N α,0. The above can be arranged because for all κ, if τ is a cardinal, then {α : α κ τ} = {0, τ}, so {N α,γ : 0 γ < µ} H θ. 20 / 22

More questions, and partial answers Let κ be any infinite cardinal. If A and B are boolean algebras whose coproduct has a κ op -like base, must one of A and B have a κ op -like base? If X and Y are compacta and X Y has a κ op -like base, must one of X and Y have a κ op -like base? Theorem (joint with Spadaro). There are non-compact X, Y such that each lacks ω op -like bases, but X Y has one. 21 / 22

More questions, and partial answers Let κ be any infinite cardinal. If A and B are boolean algebras whose coproduct has a κ op -like base, must one of A and B have a κ op -like base? If X and Y are compacta and X Y has a κ op -like base, must one of X and Y have a κ op -like base? Theorem (joint with Spadaro). There are non-compact X, Y such that each lacks ω op -like bases, but X Y has one. If A is a boolean algebra with bases E, F and E is κ op -like, must F contain a κ op -like base of A? If X is a compactum with bases E, F and E is κ op -like, must F contain a κ op -like base of X? Theorem. ω ω has an ω op -like base and another base that contains no ω op -like base of ω ω. Theorem. (GCH) If X is a homogeneous compactum with a κ op -like base, then all its bases include κ op -like bases of X. 21 / 22

A question about ℵ ω Let A be the algebra of countably supported subsets of ℵω 2. Is it consistent, relative to large cardinals, that every dense subset of A \ { } is ω 1 -top-heavy? A natural place to look is a model of Chang s Conjecture at ℵ ω. (These exist, assuming (roughly) a huge cardinal.) Theorem (Kojman, Spadaro) A \ { } has a κ op -like dense subset where κ = cf([ℵ ω ] ω, ) (which is < ℵ ω4 ). If we assume ℵω, then A \ { } has an ω op 1 -like dense subset. 22 / 22