on Rather new topi: Many op en questions Continuum funtion - - PowerPoint PPT Presentation

on
SMART_READER_LITE
LIVE PREVIEW

on Rather new topi: Many op en questions Continuum funtion - - PowerPoint PPT Presentation

F o ring, Combinato ris and Denabilit y Summa ry: T o da y: Denable W ello rders La rge a rdinals F o ring axioms Ca rdinal ha rateristis (new!) Other ontexts T uesda y: Ca rdinal Cha rateristis on


slide-1
SLIDE 1 F
  • r ing,
Combinato ri s and Denabilit y Summa ry: T
  • da
y: Denable W ello rders La rge a rdinals F
  • r ing
axioms Ca rdinal ha ra teristi s (new!) Other
  • ntexts
T uesda y: Ca rdinal Cha ra teristi s
  • n κ
Rather new topi : Many
  • p
en questions Continuum fun tion 2κ Dominating, b
  • unding
numb ers Conalit y
  • f
the symmetri group Almost disjointness, splitt i ng numb ers W ednesda y: Mo dels
  • f
PF A , BPF A
slide-2
SLIDE 2 Denable W ello rders In ZF , A C is equivalent to: H(κ+) an b e w ello rdered fo r every κ When an w e
  • btain
a denab le w ello rder
  • f
H(κ+) ?

Σ

n denab le w ello rder
  • f
H(κ+) : W ello rder
  • f
H(κ+) whi h is Σ n denab le
  • ver
H(κ+) with κ as a pa rameter Rema rks: 1. If n is at least 3, then κ an b e elimi nated, as {κ} is Π 2 denab le 2. If λ is a limi t a rdinal and H(κ+) has a denab le w ello rder fo r
  • nally
many κ < λ, then H(λ) has a denab le w ello rder

Σ

n denab le w ello rder
  • f
H(κ+) with pa rameters : W ello rder
  • f
H(κ+) whi h is Σ n denab le
  • ver
H(κ+) with a rbitra ry elements
  • f
H(κ+) as pa rameters
slide-3
SLIDE 3 Denable W ello rders: La rge a rdinals and H(ω 1) The b est situation: V = L → Ea h H(κ+) = Lκ+ has a Σ 1 denab le w ello rder Denable w ello rders and La rge Ca rdinals H(ω 1)

Σ

n denab le w ello rder
  • f
H(ω 1) (with pa rameters) ∼

Σ

1 n+ 1 denab le w ello rder
  • f
the reals (with real pa rameters) Theo rem (Manseld) Σ 1 2 w ello rder
  • f
the reals → every real b elongs to L. (Ma rtin-Steel) A Σ 1 n+ 2 w ello rder
  • f
the reals is
  • nsistent
with n W
  • din
a rdinals but in onsistent with n W
  • din
a rdinals and a measurable a rdinal ab
  • ve
them.
slide-4
SLIDE 4 Denable W ello rders: La rge Ca rdinals and H(ω 2) H(ω 2) A fo r ing is small i it has size less than the least ina essible. Small fo r ings p reserve all la rge a rdinals. Theo rem (Asp er-F) There is a small fo r ing whi h fo r es CH and a denab le w ello rder
  • f
H(ω 2) . The ab
  • ve
w ello rder is not Σ 1 . In fa t: Theo rem (W
  • din)
Measurable W
  • din
a rdinal + CH → there is no w ello rder
  • f
the reals whi h is Σ 1
  • ver
H(ω 2) . Ho w ever:
slide-5
SLIDE 5 Denable W ello rders: La rge a rdinals and H(ω 2) Theo rem (A vraham-Shelah) There is a small fo r ing whi h fo r es ∼ CH and a w ello rder
  • f
the reals whi h is Σ 1
  • ver
H(ω 2) . Question 1. Is there a small fo r ing whi h fo r es a Σ 2 w ello rder
  • f
H(ω 2) ?
slide-6
SLIDE 6 Denable W ello rders: La rge a rdinals and H(ω 2) Ab
  • ut
the p ro
  • f
  • f:
Theo rem (Asp er-F) There is a small fo r ing whi h fo r es CH and a denab le w ello rder
  • f
H(ω 2) . T w
  • ingredients:
Canoni al fun tion
  • ding
Strongly t yp e-guessing
  • ding
(Asp er)
slide-7
SLIDE 7 Denable W ello rders: La rge a rdinals and H(ω 2) Canoni al funtion
  • ding
F
  • r
ea h α < ω 2 ho
  • se
fα : ω 1 → α
  • nto
and dene gα : ω 1 → ω 1 b y: gα(γ) =
  • rdert
yp e fα[γ]. gα is a anoni al fun tion fo r α . No w
  • de
A ⊆ ω 2 b y B ⊆ ω 1 as follo ws:

α ∈

A i gα(γ) ∈ B fo r a lub
  • f γ
Assuming GCH , the fo r ing to do this is ω
  • strategi all
y losed and

ω

2
  • .
slide-8
SLIDE 8 Denable W ello rders: La rge a rdinals and H(ω 2) Asp er
  • ding
A lub-sequen e in ω 1
  • f
height τ is a sequen e C = ( Cδ | δ ∈ S) where S ⊆ ω 1 is stationa ry and ea h Cδ is lub in δ
  • f
  • rdert
yp e τ .
  • C
is strongly t yp e-guessing i fo r every lub C ⊆ ω 1 there is a lub D ⊆ ω 1 su h that fo r all δ in D ∩ S ,
  • rdert
yp e( C ∩ C +

δ ) = τ

, where C +

δ

denotes the set
  • f
su esso r elements
  • f
Cδ . An
  • rdinal γ
is p erfe t i ωγ = γ. Lemma (Asp er) Assume GCH . Let B ⊆ ω 1 . Then there is an

ω

  • strategi all
y losed, ω 2
  • fo
r ing that fo r es: γ ∈ B i the γ
  • th
p erfe t
  • rdinal
is the height
  • f
a strongly t yp e-guessing lub sequen e.
slide-9
SLIDE 9 Denable W ello rders: La rge a rdinals and H(ω 2) T
  • p
rove: Theo rem (Asp er-F) There is a small fo r ing whi h fo r es CH and a denab le w ello rder
  • f
H(ω 2) . Assume GCH . W rite H(ω 2) as Lω 2[ A] , A ⊆ ω 2 . Use Canoni al fun tion
  • ding
to
  • de
A b y B ⊆ ω 1 . Use Asp er
  • ding
to
  • de
B denab ly
  • ver
H(ω 2) . Problem: B
  • nly
  • des
H(ω 2)
  • f
the ground mo del, not H(ω 2)
  • f
the extension! Solution: P erfo rm b
  • th
  • dings
simult aneously. The fo r ing is a hyb rid fo r ing: halfw a y b et w een iteration and p ro du t.
slide-10
SLIDE 10 Denable W ello rders: La rge a rdinals and H(κ) H(κ) Theo rem (Asp er-F) There is a lass fo r ing whi h fo r es GCH , p reserves all sup er ompa t a rdinals (as w ell as a p rop er lass
  • f
n
  • huge
a rdinals fo r ea h n ) and adds a denab le w ello rder
  • f
H(κ+) fo r all regula r κ ≥ ω 1 . Co rolla ry There is a lass fo r ing whi h fo r es GCH , p reserves all sup er ompa t a rdinals (as w ell as a p rop er lass
  • f
n
  • huge
a rdinals fo r ea h n ) and adds a pa rameter-free denab le w ello rder
  • f
H(δ) fo r all a rdinals δ ≥ ω 2 whi h a re not su esso rs
  • f
singula rs. Su esso rs
  • f
singula rs? Σ 1 denab le w ello rders?
slide-11
SLIDE 11 Denable W ello rders: La rge a rdinals and H(κ) Su esso rs
  • f
singula rs: Theo rem (Asp er-F) Supp
  • se
that there is a j : L( H(λ+)) → L( H(λ+)) xing λ , with riti al p
  • int < λ.
Then there is no denab le w ello rder
  • f
H(λ+) with pa rameters. Question 2. Is there a small fo r ing that adds a denab le w ello rder
  • f
H(ℵω+ 1) with pa rameters?

Σ

1 denab le w ello rders: Theo rem There is a lass fo r ing whi h fo r es GCH , p reserves all sup er ompa t a rdinals (as w ell as a p rop er lass
  • f
n
  • huge
a rdinals fo r ea h n ) and adds a Σ 1 denab le w ello rder
  • f
H(κ+) with pa rameters fo r all regula r κ ≥ ω 1 .
slide-12
SLIDE 12 Denable W ello rders and F
  • r ing
Axioms Question 3. Is there a small fo r ing that adds a Σ 1 denab le w ello rder
  • f
H(ω 3) ? Denable w ello rders and F
  • r ing
Axioms H(ω 1) Theo rem MA is
  • nsistent
with a Σ 1 3 w ello rder
  • f
the reals. (Cai edo-F) BPF A + ω 1 = ω L 1 (whi h is
  • nsistent
relative to a ree ting a rdinal) implies that there is a Σ 1 3 w ello rder
  • f
the reals. Theo rem (Hjo rth) Assume ∼ CH and every real has a # . Then there is no

Σ

1 3 w ello rder
  • f
the reals.
slide-13
SLIDE 13 Denable W ello rders and F
  • r ing
Axioms Question 4. Do es BPF A + 0# do es not exist imply that there is a

Σ

1 3 w ello rder
  • f
the reals? Question 5. Is BMM
  • nsistent
with a p roje tive w ello rder
  • f
the reals? PF A is not. Question 6. Is MA
  • nsistent
with the nonexisten e
  • f
a p roje tive w ello rder
  • f
the reals? F
  • r
H(ω 2) : Theo rem (Cai edo-V eli k
  • vi )
BPF A + ω 1 = ω L 1 implies that there is a Σ 1 denab le w ello rder
  • f
H(ω 2) . Theo rem (La rson) Relative to enough sup er ompa ts, there is a mo del
  • f
MM with a denab le w ello rder
  • f
H(ω 2) .
slide-14
SLIDE 14 Denable W ello rders and F
  • r ing
Axioms F
  • r
la rger H(κ) : Theo rem MA is
  • nsistent
with a denab le w ello rder
  • f
H(κ+) fo r all κ . (Ree ting a rdinal) BSPF A is
  • nsistent
with a denab le w ello rder
  • f
H(κ+) fo r all κ . (Enough sup er ompa ts) MM is
  • nsistent
with a denab le w ello rder
  • f
H(κ+) fo r all regula r κ ≥ ω 1 .
slide-15
SLIDE 15 Denable W ello rders and Ca rdinal Cha ra teris ti s New
  • ntext
fo r denab le w ello rders: Ca rdinal Cha ra teristi s T emplate iteration T : A
  • untable
supp
  • rt, ω
2
  • iteration
whi h adds a Σ 1 3 w ello rder
  • f
the reals (and a Σ 1 w ello rder
  • f
H(ω 2) ). It is not p rop er, but is S
  • p
rop er fo r ertain stationa ry S ⊆ ω 1 . Broad p roje t: Mix the template iteration with a va riet y
  • f
iterations fo r
  • ntrolling
a rdinal ha ra teristi s. Theo rem (V.Fis her
  • F)
Ea h
  • f
the follo wing is
  • nsistent
with a Σ 1 3 w ello rder
  • f
the reals: d < c , b < a = s, b < g.

b =

the b
  • unding
numb er, a = the almost disjointness numb er,

s =

the splitt i ng numb er, g = the group wise densit y numb er
slide-16
SLIDE 16 Denable W ello rders The template iteration T is gentle (ωω b
  • unding)
but also exible (it an b e mixed with any
  • untable
supp
  • rt
p rop er iteration
  • f
p
  • sets
  • f
size ω 1 ) One an also ask fo r ni ely denab le witnesses to a rdinal ha ra teristi s. A sample result: Theo rem (F-Zdomskyy) It is
  • nsistent
that a = ω 2 and there is a Π 1 2 innite maximal almost disjoint family . Question 7. Is it
  • nsistent
with a = ω 2 that there is a Σ 1 2 innite maximal almost disjoint family?
slide-17
SLIDE 17 Denable W ello rders in
  • ther
Contexts Questions. 8. Is it
  • nsistent
that fo r all innite regula r κ , GCH fails at κ and there is a denab le w ello rder
  • f
H(κ+) ? 9. Is the tree p rop ert y at ω 2
  • nsistent
with a p roje tive w ello rder
  • f
the reals? 10. Is it
  • nsistent
that the nonstationa ry ideal
  • n ω
1 is saturated and there is a Σ 1 4 w ello rder
  • f
the reals? 11. Is it
  • nsistent
that GCH fails at a measurable a rdinal κ and there is a denab le w ello rder
  • f
H(κ+) ?
slide-18
SLIDE 18 Ca rdinal Cha ra teristi s at κ Ca rdinal ha ra teristi s
  • n ω
is a vast subje t. Examples from Blass' survey:

a, b, d, e, g, h, i, m, p, r, s, t, u

These a re all at most c , the a rdinalit y
  • f
the
  • ntinuum.

κ

regula r, un oun tab le. W e
  • nsider
analogues
  • f
some
  • f
the ab
  • ve
fo r κ

a(κ), b(κ), d(κ) . . .

Why?
slide-19
SLIDE 19 Ca rdinal Cha ra teristi s at κ F
  • ur
reasons: 1. Higher iterated fo r ing Ca rd Cha rs ω / Countable supp
  • rt
iterations ≡ Ca rd Cha rs κ / Higher supp
  • rt
iterations 2. La rge a rdinal
  • ntext:
Ca rd Cha rs at a measurable 3. Global b ehaviour as κ va ries, Internal
  • nsisten y
4. Solve p roblems at κ that a re unsolved at ω Illustr at e with some examples
slide-20
SLIDE 20 Ca rdinal Cha ra teristi s at κ The Ca rd Cha r 2κ Global b ehaviour Theo rem (Co rolla ry to Easton's Theo rem) It is
  • nsistent
that 2α = α++ fo r all regula r α . F
  • r ing
used: Easton p ro du t
  • f α
  • Cohen
fo r ings Cohen(α, α++) . Internal
  • nsisten y
Theo rem (F-Ondrejovi) Assuming that 0# exists, there is an inner mo del in whi h 2α = α++ fo r all regula r α . F
  • r ing
used: Reverse Easton iteration
  • f α
  • Cohen
fo r ings.
slide-21
SLIDE 21 Ca rdinal Cha ra teristi s at κ La rge a rdinal
  • ntext
Theo rem (W
  • din)
Assume that κ is hyp ermeasurable. Then in a fo r ing extension, κ is measurable and 2κ = κ++ . F
  • r ing
used: Reverse Easton iteration
  • f α
  • Cohen
fo r ings, α ≤ κ ,

α

ina essible, follo w ed b y Cohen(κ+, κ++) .
slide-22
SLIDE 22 Ca rdinal Cha ra teristi s at κ No w lo
  • k
at The Ca rd Cha r d(κ) Global Behaviour Theo rem (Cummings-Shelah) It is
  • nsistent
that d(α) = α+ < 2α fo r all regula r α . F
  • r ings
used: α
  • Cohen
p ro du t and α
  • He hler
iteration.
slide-23
SLIDE 23 Ca rdinal Cha ra teristi s at κ La rge a rdinal
  • ntext
Theo rem (F-Thompson) Assume that κ is hyp ermeasurable. Then in a generi extension, κ is measurable and d(κ) = κ+ < 2κ . F
  • r ing
used: Reverse Easton iteration
  • f α
  • Sa ks
p ro du ts, α ≤ κ ,

α

ina essible. T w
  • interesting
p
  • ints:
i. If y
  • u
try this with κ
  • Cohen
and κ
  • He hler
then y
  • u
need some sup er ompa tness ii. The p ro
  • f
is easier than W
  • din's
p ro
  • f,
whi h
  • nly
gives

κ+ <

slide-24
SLIDE 24 Ca rdinal Cha ra teristi s at κ La rge Ca rdinal
  • ntext
together with Global Behaviour Theo rem (F-Thompson) Assume that κ is hyp ermeasurable. Then in a generi extension, κ is measurable and (α) = α+ < 2α fo r all regula r α . F
  • r ings
used: (Reverse Easton iteration
  • f
) α
  • Sa ks
at ina essible

α ≤ κ

, α
  • Cohen
p ro du t follo w ed b y α
  • He hler
iteration at su esso rs
  • f
non-ina essibles, something new at α+ , α ina essible (α+
  • Cohen
p ro du t follo w ed b y a mixture
  • f α
  • Sa ks
p ro du t and

α+

  • He hler
iteration). Con lusion: Understanding d(κ) in the La rge a rdinal setting requires a a reful hoi e
  • f
fo r ings; mixing it with the Global Behaviour
  • f d(α)
requires the invention
  • f
new fo r ings
slide-25
SLIDE 25 Ca rdinal Cha ra teristi s at κ Rema rk. (F-Honzik) Easton's Theo rem fo r 2α has b een w
  • rk
ed
  • ut
in the la rge a rdinal setting. But: Question 12. What Global Behaviours fo r d(α) a re p
  • ssible
when there is a measurable a rdinal?
slide-26
SLIDE 26 Ca rdinal Cha ra teristi s at κ The Ca rd Cha r CofSym(α) Sym(α) = group
  • f
p ermutations
  • f α
under
  • mp
  • sition.
CofSym(α) = least λ su h that Sym(α) is the union
  • f
a stri tl y in reasing λ - hain
  • f
subgroups. Sha rp and Thomas: CofSym(α) an b e anything reasonable. But its Global Behaviour is nontrivial! Theo rem (Sha rp-Thomas) (a) Supp
  • se
that α < β a re regula r and GCH holds. Then in a
  • nalit
y-p reserving fo r ing extension, CofSym(α) = β . (b) If CofSym(α) > α+ then CofSym(α+) ≤ CofSym(α) . Question 13. Is it
  • nsistent
that CofSym(ω) = CofSym(ω 1) = ω 3 ?
slide-27
SLIDE 27 Ca rdinal Cha ra teristi s at κ CofSym has b een studied in the La rge Ca rdinal setting: Theo rem (F-Zdomskyy) Supp
  • se
that κ is hyp ermeasurable. Then in a fo r ing extension, κ is measurable and CofSym(κ) = κ++ . F
  • r ings
used: Iteration
  • f
Mill er(κ) (with
  • ntinuous
lub-splitti ng) and a generalisation
  • f
Sa ks(κ) . The p ro
  • f
also uses g l(κ) (group wise densit y numb er fo r
  • ntinuous
pa rtitions).
slide-28
SLIDE 28 Ca rdinal Cha ra teristi s at κ

a(κ)

and d(κ)

a(κ) =

minimum size
  • f
a (size at least κ ) maximal almost disjoint family
  • f
subsets
  • f κ
An
  • ld
  • p
en p roblem: Question 14. Do es d(ω) = ω 1 imply a(ω) = ω 1 ? But this is solved at un oun tab le a rdinals! Theo rem (Blass-Hyttinen-Zhang) F
  • r
un oun tab le α , d(α) = α+ implies

a(α) = α+

slide-29
SLIDE 29 Ca rdinal Cha ra teristi s at κ Are there
  • ther
  • p
en questions fo r ω whi h an b e solved fo r un oun tab le a rdinals? Question 15. Can p(κ) b e less than t(κ) ? Ma yb e it will help to assume that κ is a la rge a rdinal. Question 16. Can s(κ) b e singula r? Mo re
  • p
en Questions. 17. (Without la rge a rdinals) Is b(κ) < a(κ)
  • nsistent
fo r an un oun tab le κ ? 18. Whi h Global Behaviours fo r b(α), d(α) a re internally
  • nsistent?
Cummings-Shelah answ ered this fo r
  • rdina
ry
  • nsisten y
. 19. (Without sup er ompa tness) Can s(κ) b e greater than κ+ ? Zapletal: Need (almost) a hyp ermeasurable. 20. Is it
  • nsistent
that CofSym(κ) = κ+++ fo r a measurable κ ?
slide-30
SLIDE 30 Some mo dels
  • f
PF A , BPF A Let C b e a lass
  • f
fo r ings F A(C) = F
  • r ing
Axiom fo r C F
  • r
P in C , an hit ω 1
  • many
p redense sets in P with a lter
  • n
P BF A(C) = Bounded F
  • r ing
Axiom fo r C F
  • r
P in C , an hit ω 1
  • many
p redense sets
  • f
size ≤ ω 1 in P with a lter
  • n
P PF A = F A( Prop er) = Prop er F
  • r ing
Axiom BPF A = BF A( Prop er) = Bounded Prop er F
  • r ing
Axiom Useful F a t. (Baga ria, Stavi-Vnnen ) BPF A is equivalent to the

Σ

1 elementa rit y
  • f
H(ω 2) V in H(ω 2) V [ G] fo r p rop er P and P
  • generi
G
slide-31
SLIDE 31 Some mo dels
  • f
PF A , BPF A Theo rem (a) (Baumga rtner) If there is a sup er ompa t then PF A holds in a p rop er fo r ing extension. (b) (Goldstern-Shelah) If there is a ree ting a rdinal (i.e., a regula r κ su h that H(κ) ≺Σ 2 V ) then BPF A holds in a p rop er fo r ing extension.
slide-32
SLIDE 32 Some mo dels
  • f
PF A , BPF A Ca rdinal Minimal i t y V is a rdinal minimal i whenev e r M is an inner mo del with the
  • rre t
a rdinals (i.e., Ca rd M = Ca rd V ) then M = V . Lo al version: κ a a rdinal. V is κ
  • minimal
i whenev e r M is an inner mo del with the
  • rre t
a rdinals ≤ κ then H(κ) M = H(κ) . Examples L is triviall y a rdinal minimal . Let x b e κ
  • Sa ks, κ
  • Mill
er
  • r κ
  • Laver
  • ver
L . Then L[ x] is not a rdinal minimal . Let f : κ → κ+ b e a minimal
  • llapse
  • f κ+
to κ
  • ver
L . Then L[ f ] is a rdinal minimal . Mo re interesting examples: Co re mo dels
slide-33
SLIDE 33 Some mo dels
  • f
PF A , BPF A Theo rem Let K b e the
  • re
mo del fo r a measurable, hyp ermeasurable, strong
  • r
W
  • din
a rdinal. Then K is a rdinal minimal . In fa t, K is

κ

  • minimal
fo r all κ ≥ ω 2 .

ω

1
  • minimal
i t y fails fo r
  • re
mo dels, and in fa t whenev e r 0# exists: Theo rem Supp
  • se
that 0# exists. Then V is not ω 1
  • minimal
. In fa t, there is an inner mo del M with the
  • rre t ω
1 whi h is a fo r ing extension
  • f
L.
slide-34
SLIDE 34 Some mo dels
  • f
PF A , BPF A Another sour e
  • f
a rdinal minimal i t y: Mo dels
  • f
fo r ing axioms SPF A = F A( Semip rop er) = Semip rop er F
  • r ing
Axiom BSPF A = BF A( Semip rop er) = Bounded Semip rop er F
  • r ing
Axiom Theo rem (V eli k
  • vi )
Supp
  • se
that SPF A holds. Then V is ω 2
  • minimal
. There is a related result fo r BPF A: Theo rem (Cai edo-V eli k
  • vi )
Supp
  • se
that BPF A holds. Then V is

ω

2
  • minimal
with resp e t to inner mo dels satisfyi ng BPF A : If M is an inner mo del satisfyi ng BPF A with the
  • rre t ω
2 then H(ω 2) M = H(ω 2) .
slide-35
SLIDE 35 Some mo dels
  • f
PF A , BPF A Theo rem (a) Supp
  • se
that there is a sup er ompa t. Then in some fo r ing extension, PF A holds and the universe is not ω 2
  • minimal
. (b) Supp
  • se
that there is a ree ting a rdinal. Then in some fo r ing extension, BPF A holds and the universe is not ω 2
  • minimal
. The p ro
  • fs
a re based
  • n:
Lemma (Collapsing to ω 2 with nite
  • nditions)
Assume GCH . Supp
  • se
that κ is ina essible and S denotes [κ]ω
  • f
V . Then there is a fo r ing P su h that: (a) P fo r es κ = ω 2 . (b) P is S
  • p
rop er, and hen e p reserves ω 1 , in any extension
  • f
V in whi h S remains stationa ry .
slide-36
SLIDE 36 Some mo dels
  • f
PF A , BPF A W e sk et h the p ro
  • f
  • f
(a): (a) Supp
  • se
that there is a sup er ompa t. Then in some fo r ing extension, PF A holds and the universe is not ω 2
  • minimal
.

κ

sup er ompa t. Collapse κ to ω 2 with nite
  • nditions,
p ro du ing V [ F] . P erfo rm Baumga rtner's PF A iteration, but at stage α < ω 2 , ho
  • se
a fo r ing in V [ F ↾ α, Gα] whi h is S
  • p
rop er there; a rgue that it is also S
  • p
rop er in V [ F, Gα] . Imp
  • rtant:
Only use names from V [ F ↾ α, Gα] , to k eep the fo r ing small! Diagonal iteration V erify that PF A (indeed F A(S − Prop er) ) holds in V [ F, G] . As κ = ω 2 b
  • th
in V [ F] and in V [ F, G] , this sho ws that V [ F, G] is not ω 2
  • minimal
.
slide-37
SLIDE 37 Some mo dels
  • f
PF A , BPF A Ho w to
  • llapse
an ina essible κ to ω 2 with nite
  • nditions?
Let # : [κ]ω → κ b e inje tive. P
  • nsists
  • f
all pairs p = ( A, S) su h that: 1. A is a nite set
  • f
disjoint losed intervals [α, β] , α ≤ β < κ ,
  • f(α) ≤ ω
1 . 2. S is a nite subset
  • f [κ]ω
(side
  • nditions).
3. T e hni al. 4. Let F b e the set
  • f
un oun tab le
  • nalit
y α fo r [α, β] in A, together with κ . The height
  • f
x ∈ S is the least element
  • f
F greater than sup x . Then: i. (Closure under trun ation) x in S , α in F implies x ∩ α in S . ii. (Almost an ∈
  • hain)
If x, y ∈ S have the same height then

#( x) ∈

y , #( y) ∈ x
  • r
x = y .
slide-38
SLIDE 38 Some mo dels
  • f
PF A , BPF A The fo r ing is κ
  • and
adds a lub in κ
  • nsisting
  • nly
  • f
  • rdinals
  • f
  • nalit
y ≤ ω 1 . So κ b e omes ω 2 . Questions. 21. Supp
  • se
that BSPF A holds. Then is V ω 2
  • minimal
with resp e t to inner mo dels satisfyi ng BSPF A ? 22. Is there a fo r ing whi h
  • llapses
an ina essible to ω 3 with nite
  • nditions?