ITERATIONS WITH MIXED SUPPORT

Similar documents
The Bounded Axiom A Forcing Axiom

GREGORY TREES, THE CONTINUUM, AND MARTIN S AXIOM

THE PRECIPITOUSNESS OF TAIL CLUB GUESSING IDEALS

Annals of Pure and Applied Logic

BPFA and BAAFA. Their equiconsistency and their nonequivalence. Thilo Weinert. February 4, 2009

SOME REMARKS ON NON-SPECIAL COHERENT ARONSZAJN TREES

THE STRONG TREE PROPERTY AND THE FAILURE OF SCH

Cardinal characteristics, projective wellorders and large continuum

FORCING WITH SEQUENCES OF MODELS OF TWO TYPES

ON THE COFINALITY OF THE SPLITTING NUMBER

INSEPARABLE SEQUENCES AND FORCING

FORCING (LECTURE NOTES FOR MATH 223S, SPRING 2011)

Successive cardinals with the tree property

SEPARATION IN CLASS FORCING EXTENSIONS. Contents

A NOTE ON THE EIGHTFOLD WAY

FORCING AXIOMS AND THE CONTINUUM HYPOTHESIS, PART II: TRANSCENDING ω 1 -SEQUENCES OF REAL NUMBERS

More forcing notions imply diamond

A variant of Namba Forcing

NAMBA FORCING, WEAK APPROXIMATION, AND GUESSING

The axiom of choice and two-point sets in the plane

THE TREE PROPERTY AND THE FAILURE OF THE SINGULAR CARDINAL HYPOTHESIS AT ℵ ω 2

A NEW LINDELOF SPACE WITH POINTS G δ

PSEUDO P-POINTS AND SPLITTING NUMBER

Forcing with matrices of countable elementary submodels

Bounding by canonical functions, with CH

A DIRECT PROOF OF THE FIVE ELEMENT BASIS THEOREM

20 Ordinals. Definition A set α is an ordinal iff: (i) α is transitive; and. (ii) α is linearly ordered by. Example 20.2.

WHY Y-C.C. DAVID CHODOUNSKÝ AND JINDŘICH ZAPLETAL

Souslin s Hypothesis

Cardinal invariants above the continuum

VARIATIONS FOR SEPARATING CLUB GUESSING PRINCIPLES

THE TREE PROPERTY AT ℵ ω+1 DIMA SINAPOVA

ITERATING ALONG A PRIKRY SEQUENCE

SOME COMBINATORIAL PRINCIPLES DEFINED IN TERMS OF ELEMENTARY SUBMODELS. 1. Introduction

TOPOLOGY PROCEEDINGS Volume 28, No. 1, 2004 Pages

GROUPS WITH UNBOUNDED POTENTIAL AUTOMORPHISM TOWER HEIGHTS


Locally Connected HS/HL Compacta

Independence of Boolean algebras and forcing

On the open-open game

An Iterated Forcing Extension In Which All Aleph-1 Dense Sets of Reals Are Isomorphic

Singular Failures of GCH and Level by Level Equivalence

Forcing Axioms and Inner Models of Set Theory

COMPACT C-CLOSED SPACES NEED NOT BE SEQUENTIAL

An introduction to OCA

arxiv: v1 [math.lo] 1 Feb 2016

Increasing δ 1 2 and Namba-style forcing

SELF-DUAL UNIFORM MATROIDS ON INFINITE SETS

CH AND THE MOORE-MROWKA PROBLEM

M ath. Res. Lett. 15 (2008), no. 00, NN c International Press 2008 A UNIVERSAL ARONSZAJN LINE. Justin Tatch Moore. 1.

Applications of higher-dimensional forcing

DENSELY k-separable COMPACTA ARE DENSELY SEPARABLE

Recursive definitions on surreal numbers

MONOTONICALLY COMPACT AND MONOTONICALLY

SMALL SUBSETS OF THE REALS AND TREE FORCING NOTIONS

MORE ABOUT SPACES WITH A SMALL DIAGONAL

Non-trivial automorphisms from variants of small d

Covering a bounded set of functions by an increasing chain of slaloms

Template iterations with non-definable ccc forcing notions

A model with no magic set

COMPLETE NORMALITY AND COUNTABLE COMPACTNESS

CCC Forcing and Splitting Reals

P-FILTERS AND COHEN, RANDOM, AND LAVER FORCING

Strong theories, burden, and weight

HEREDITARILY NORMAL MANIFOLDS OF DIMENSION > 1 MAY ALL BE METRIZABLE

Selective Ultrafilters on FIN

Jónsson posets and unary Jónsson algebras

Projective well-orderings of the reals and forcing axioms

Topology Proceedings. COPYRIGHT c by Topology Proceedings. All rights reserved.

HYBRID PRIKRY FORCING

FORCING AXIOMS AND THE CONTINUUM HYPOTHESIS

COUNTABLE COMPACTNESS, HEREDITARY π CHARACTER, AND THE CONTINUUM HYPOTHESIS

A product of γ-sets which is not Menger.

Appalachian Set Theory Workshop on Coherent Sequences Lectures by Stevo Todorcevic Notes taken by Roberto Pichardo Mendoza

Generic Absoluteness. Joan Bagaria and Sy D. Friedman. Abstract. However, this is not true for 1 3 predicates. Indeed, if we add a Cohen real

ISOLATING CARDINAL INVARIANTS

arxiv: v1 [math.lo] 19 Feb 2019

The Arkhangel skiĭ Tall problem under Martin s Axiom

Maharam Algebras. Equipe de Logique, Université de Paris 7, 2 Place Jussieu, Paris, France

FRAGILITY AND INDESTRUCTIBILITY II

DO FIVE OUT OF SIX ON EACH SET PROBLEM SET

Compact subsets of the Baire space

THE LENGTH OF THE FULL HIERARCHY OF NORMS

THE EXTENDER ALGEBRA AND Σ 2 1-ABSOLUTENESS

Aronszajn Trees and the SCH

SEPARATING CLUB GUESSING PRINCIPLES IN THE PRESENCE OF FAT FORCING AXIOMS

For a xed value of, we will prove that there are some simple constraints on the triple of cardinals (b(); d(); 2 ). We will also prove that any triple

The Proper Forcing Axiom: a tutorial

PFA AND IDEALS ON ω 2 WHOSE ASSOCIATED FORCINGS ARE PROPER

Lusin sequences under CH and under Martin s Axiom

Forcing notions in inner models

Closure properties of parametric subcompleteness

A FIVE ELEMENT BASIS FOR THE UNCOUNTABLE LINEAR ORDERS

More Forcing Constructions

PROPER FORCING REMASTERED

The choice of ω 1 and combinatorics at ℵ ω

THE FAILURE OF DIAMOND ON A REFLECTING STATIONARY SET

HOW DO ULTRAFILTERS ACT ON THEORIES? THE CUT SPECTRUM AND TREETOPS

SOME RESULTS IN POLYCHROMATIC RAMSEY THEORY

TEMPLATE ITERATIONS AND MAXIMAL COFINITARY GROUPS

Transcription:

ITERATIONS WITH MIXED SUPPORT VERA FISCHER Abstract. In this talk we will consider three properties of iterations with mixed (finite/countable) supports: iterations of arbitrary length preserve ω 1, iterations of length ω 2 over a model of CH have the ℵ 2 -chain condition and iterations of length < ω 2 over a model of CH do not increase the size of the continuum. Definition 1. Let P κ be an iterated forcing construction of length κ, with iterands Q α : α < κ such that for every α < κ α Q α is σ-centered or α Q α is countably closed. Then P κ is finite/countable iteration if and only if for every p P κ, support(p) = {α < κ : p(α) 1 α } is countable and Fsupport(p) = {α : Q α is σ-centered, p(α) 1 α } is finite. Remark 1. In the context of the above definition, whenever α Q α is σ-centered we will say that α is a σ-centered stage and correspondingly, whenever Q α is countably closed we will say that α is a countably closed stage. From now on P κ is a finite/countable iteration of length κ. Definition 2. Let p, q P κ. We say than p D q if and only if p q and for every σ-centered stage α < κ, p α p(α) = q(α). Similarly p C q if and only if for every countably closed stage α, p α p(α) = q(α). Claim. Both D and C are transitive relations. Lemma 1. Let p n n ω be a sequence in P κ such that for every n ω, p n+1 D p n. Then there is a condition p P κ such that for every n ω, p D p n. Date: August 2, 2007. 1

2 VERA FISCHER Proof. Define p inductively. It is sufficient to define p for successor stages α. Suppose we have defined p α so that for every n ω, p α D p n α. If α is σ-centered then p α p 0 (α) = p 1 (α) =... and so we can define p(α) = p 0 (α). If α is countably closed stage, then p α p 0 (α) p 1 (α)... and since α Q α is countably closed there is a P α -name p(α) such that p α p(α) p n (α) for every n ω. Lemma 2. Let p, q P κ be such that p q. Then there is a condition r P κ such that p C r D q. Proof. Again we will define r inductively. Suppose we have defined r α so that p α C r α D q α. Then if α is a σ-centered stage, let r(α) = q(α). If α is a countably closed stage, define r(α) to be a P α -term such that: if r p α then r α r(α) = p(α), if r is incompatible with p α then r α r(α) = q(α). To verify p C r note that if α is a σ-centered stage then p α p(α) q(α) = r(α). If α is a countably closed stage, then p α p(α) = r(α). To see that r D q note that if α is a σ-centered stage then by definition r α r(α) = q(α). If α is countably closed stage, it is sufficient to show that 1 α r(α) q(α). Let r P α. If r p α fix a common extension t. Then t p(α) = r(α) p(α) q(α) and so t r(α) q(α). If r is incompatible with p α then by definition r r(α) = q(α). Definition 3. Let α be a σ-centered stage and let ṡ α be a P α -name such that α (ṡ α : Qα ω) [ p, q Q α (ṡ α (p) = ṡ α (q) p q)]. Condition p P κ is determined if and only if for every α Fsupport(p) there is n ω such that p α ṡ α (p(α)) = ň. Claim. The set of determined conditions in P κ is dense. Proof. Proceed by induction on the length of the iteration κ. It is sufficient to consider successor stages. Let α = β + 1 and p P α. We can assume that β is a σ-centered stage. By inductive hypothesis there is a determined condition r p β such that r ṡ β (p(β)) = ň for some n ω. If r P α is such that r β = r and r(β) = p(β), then r is determined and r p. Lemma 3. Let q 1, q 2 be (determined) conditions in P κ such that Fsupport(q 1 ) = Fsupport(q 2 ) = F

ITERATIONS WITH MIXED SUPPORT 3 and for every α F there is n ω such that q 1 α ṡ α (q 1 (α)) = ň and q 2 α ṡ α (q 2 (α)) = ň. Furthermore, let p P κ such that q 1 C p and q 2 p. Then q 1 and q 2 are compatible. Proof. The common extension r of q 1 and q 2 will be defined inductively. Suppose we have defined r α for some α < κ such that r α q 1 α and r α q 2 α. If α is a σ-centered stage, then there is n ω such that r α ṡ α (q 1 (α)) = ṡ α (q 2 (α)) = ň and so there is a P α -name r(α) for a condition in Q α such that r α r(α) q 1 (α) r(α) q 2 (α). If α is countably closed then r α q 1 (α) = p(α) q 2 (α) p(α). Thus we can define r(α) = q 2 (α). Definition 4. An antichain q ξ : ξ < η of determined conditions is concentrated with witnesses p ξ : ξ < η if and only if ξ < η(q ξ C p ξ ) and ζ < ξ(p ξ D p ζ ). Lemma 4. There are no uncountable concentrated antichains. Proof. Suppose to the contrary that q ξ : ξ < ω 1 is a concentrated antichain with witnesses p ξ : ξ < ω 1. For every ξ < ω 1 let F ξ = Fsupport(q ξ ). Since a subset of a concentrated antichain is a concentrated antichain, we can assume that F ξ : ξ < ω 1 form a -system with root F such that for some α < κ, F α < min F ξ \F for every ξ < ω 1. Claim. q ξ α : ξ < ω 1 is a concentrated antichain in P α. Proof. Suppose to the contrary that there are ζ < ξ such that for some r P α, r q ζ α and r q ξ α. Then for every γ α, define r(γ) as follows: if γ F ζ let r(γ) = q ζ (γ), otherwise let r(γ) = q ξ (γ). Inductively we will show that r is a common extension of q ζ and q ξ. It is sufficient to show that for all countably closed stages γ if r γ q ξ γ and r γ q ζ γ, then r γ r(γ) q ζ (γ) r(γ) q ξ (γ). Note that r γ (q ζ (γ) = p ζ (γ)) (p ξ (γ) p ζ (γ)) (q ξ (γ) = p ξ (γ)) and so r γ r(γ) = q ξ (γ) q ζ (γ). For every ξ < ω 1 let f ξ : F ω be such that f ξ (γ) = n if and only if q ξ γ ṡ γ (q ξ (γ)) = ň. Since there are only countably many such functions, there are ζ < ξ such that f ζ = f ξ. Then q ζ α, q ξ α and p ζ α satisfy the hypothesis of Lemma 3 and so q ζ α and q ξ α are compatible, which is a contradiction.

4 VERA FISCHER Lemma 5. Let p P κ and let ẋ be a P κ -name such that p ẋ V. Then there is q D p and a ground model countable set X such that q ẋ ˇX. Proof. Inductively construct concentrated antichain q ξ : ξ < η < ω 1 with witnesses p ξ : ξ < η < ω 1 such that for all ξ < η, q ξ p, p ξ D p and x ξ V, q ξ ẋ = ˇx ξ. Furthermore we will have that if ξ ζ, then x ξ x ζ. Suppose q ξ : ξ < η and p ξ : ξ < η have been defined. Since η is countable, by Lemma 1 there is condition p such that ξ < η(p D p ξ ). Case 1. If p ẋ {x ξ : ξ < η}, then let q = p and X = {x ξ : ξ < η}. Case 2. Otherwise, there is q η p, x η V such that x η / {x ξ : ξ < η} and q η ẋ = ˇx η. By Lemma 2 there is p η such that q η C p η D p. This extends the concentrated antichain and so completes the inductive step. Since there are no uncountable concentrated antichains at some countable stage of the construction Case 1 must occur. Corollary 1. Let p be a condition in P κ and let f be a P κ -name such that p f : ω V. Then there is q D p and X V, X countable such that q f ω ˇX. Therefore P κ preserves ω 1. Proof. Inductively define a sequence of conditions p n n ω in P κ, where p 1 = p and a sequence of countable sets {X n } n ω V such that for every n, p n+1 D p n and p n f(n) ˇX n. Let q P κ be such that q D p n for every n and let X = n ω X n. Then X is a countable ground model set, q D p and q f ω ˇX. Theorem 1 (CH). Let P ω2 be a finite/countable support iteration with iterands Q α : α < ω 2 such that α < ω 2, α Q α = ℵ 1. Then P ω2 is ℵ 2 -c.c. Proof. It is sufficient to show that for every α < ω 2 there is a dense subset D α in P α of cardinality ℵ 1. Suppose q ξ : ξ < ω 2 is an antichain in P ω2 of size ℵ 2. We can assume that F ξ : ξ < ω 2 where F ξ = Fsupport(q ξ ) form a -system with root F such that for some α < ω 2, F α < min F ξ \F for every ξ < ω 2. Then q ξ α : ξ < ω 2 is an antichain in P α of size ℵ 2 which is not possible. As an additional requirement we will have that for every p P α there is d D α such that d D p. Proceed by induction. Suppose α = β + 1 and we have defined D β D -dense in P β of size ℵ 1. Let { d ξ : ξ < ω 1 } be a set of P β -terms such that β Q β = { d ξ : ξ < ω 1 }. For every countable antichain A D β and function f : A ω 1 let q(f) be a P β -term such that for

ITERATIONS WITH MIXED SUPPORT 5 every p P β the following holds: if there is a A such that p a then p q(f) = d f(a) ; if p is incompatible with every element of A then p q(f) = 1 β. The collection T of all such names is of size ℵ 1 and so D α = {p P α : p β D β and p(β) T } is of size ℵ 1 as well. We will show that D α is D -dense in P α. Consider arbitrary p P α and let A be a maximal antichain of conditions in D β such that for every a A there is ξ < ω 1 such that a p(β) = d ξ. Claim. There is q D p β s. t. A = {a A : a q} is countable. Proof. Fix an enumeration a ξ : ξ < ω 1 of A and let ẋ = { ˇξ, a ξ : ξ < ω 1 }. Then ẋ is a P β -name for an ordinal and so repeating the proof of Lemma 4 we can obtain X V [ω 1 ] ω and q D p such that q ẋ ˇX. Then q ẋ sup ˇX and so ξ > sup X(q a ξ ). Let f : A ω 1 be such that f(a) = ξ if and only if a p(β) = d ξ. Then q q(f) = p(β). By inductive hypothesis, we can assume that q D β and so if r P α is such that r β = q and r(β) = q(f) then r D p and r D α. Suppose α is a limit and for every β < α we have defined a D - dense subset D β of P β of size ℵ 1. Let D β be the image of D β under the canonical embedding of P β into P α. Then D = β<α Dβ is of size ℵ 1 and furthermore there is a set D P α of size ℵ 1 which contains D and such that for every sequence p n n ω D for which n(p n+1 D p n ) there is p D such that n ω(p D p n ). We will show that D is D -dense in P α. Let p P α. If sup(support(p)) = β < α then by inductive hypothesis there is d D β such that d D p. Otherwise fix an increasing and cofinal sequence α n n ω in α such that Fsupport(p) α 0. Inductively define a sequence d n n ω such that for all n, d n P αn and d n+1 D d n p α n+1. If d D is such that n(d D d n ), then n(d D p α n ) and so d D p. Lemma 6 (CH). A forcing notion which preserves ω 1 and has a dense subset of size ℵ 1 does not increase the size of the continuum. Proof. Suppose P is a forcing notion which preserves ω 1 and has a dense subset D of size ℵ 1. Let T be the collection of all pairs p, ẏ where p D, ẏ is a P-name for a subset of ω and for every n ω, there is a countable antichain of conditions in D, deciding ň ẏ which is maximal below p. Then T 2 ℵ 0 = ℵ 1. We will show that V P 2 ℵ 0 T.

6 VERA FISCHER Consider any p P and ẏ a P-name for a subset of ω. Then for every n ω let r n,ξ : ξ < ω 1 be a maximal antichian of conditions in D deciding ň ẏ. Let f be a P-term such that f(n) = ξ iff r n,ξ Ġ. Then p f : ω ω 1 and since P preserves ω 1 and D is dense, there is q D such that q p and q f ω ˇβ for some β < ω 1. Then q, ẏ is a pair in T. Corollary 2 (CH). Let P ω2 be a finite/countable iteration of length ω 2 with iterands Q α : α < ω 2 such that α Q α 2 ℵ 0. Then for every α < ω 2, V Pα CH. Proof. Proceed by induction on α, repeating the proof of Theorem 1 and using Lemma 6. References [1] J. Baumgartner Iterated forcing, Surveys in set theory (A.R.D. Mathias, editor), London Mathematical Society Lecture Notes Series, no. 87, Cambridge University Press, Cambridge, 1983, pp. 1-59. [2] P. Dordal, A Model in which the base-matrix tree cannot have cofinal branches, The Journal of Symbolic Logic, Vol. 52, No. 3, 1987, pp. 651-664. [3] K. Kunen Set Theory, North Holland, Amsterdam, 1980.