Professional Documents
Culture Documents
W. Hugh Woodin
De Gruyter
Mathematics Subject Classification 2010: 03-02, 03E05, 03E15, 03E25, 03E35, 03E40,
03E57, 03E60.
ISBN 978-3-11-019702-0
e-ISBN 978-3-11-021317-1
ISSN 1438-1893
Library of Congress Cataloging-in-Publication Data
Woodin, W. H. (W. Hugh)
The axiom of determinacy, forcing axioms, and the nonstationary
ideal / by W. Hugh Woodin. 2nd rev. and updated ed.
p. cm. (De Gruyter series in logic and its applications ; 1)
Includes bibliographical references and index.
ISBN 978-3-11-019702-0 (alk. paper)
1. Forcing (Model theory) I. Title.
QA9.7.W66 2010
511.3dc22
2010011786
Contents
1 Introduction
1.1 The nonstationary ideal on !1 . . . . . . . . . . .
1.2 The partial order Pmax . . . . . . . . . . . . . . . .
1.3 Pmax variations . . . . . . . . . . . . . . . . . . .
1.4 Extensions of inner models beyond L.R/ . . . . .
1.5 Concluding remarks the view from Berlin in 1999
1.6 The view from Heidelberg in 2010 . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
2
6
10
13
15
18
2 Preliminaries
2.1 Weakly homogeneous trees and scales
2.2 Generic absoluteness . . . . . . . . .
2.3 The stationary tower . . . . . . . . .
2.4 Forcing Axioms . . . . . . . . . . . .
2.5 Reection Principles . . . . . . . . .
2.6 Generic ideals . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
21
21
31
34
36
41
43
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
184
184
187
192
199
221
232
238
274
6 Pmax variations
6.1 2 Pmax . . . . . . . . . . . . . . . . . .
6.2 Variations for obtaining !1 -dense ideals
6.2.1 Qmax . . . . . . . . . . . . . .
6.2.2 Qmax . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
287
288
306
306
334
vi
Contents
6.3
6.2.3 2 Qmax . . . . . . . . . . . . . . . .
6.2.4 Weak Kurepa trees and Qmax . . . .
6.2.5 KT Qmax . . . . . . . . . . . . . . .
6.2.6 Null sets and the nonstationary ideal
Nonregular ultralters on !1 . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
370
377
383
403
421
7 Conditional variations
426
7.1 Suslin trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
7.2 The Borel Conjecture . . . . . . . . . . . . . . . . . . . . . . . . . . 441
8 | principles for !1
493
8.1 Condensation Principles . . . . . . . . . . . . . . . . . . . . . . . . 496
|NS
8.2 Pmax
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501
C
CC
8.3 The principles, |NS and |NS . . . . . . . . . . . . . . . . . . . . . . 577
9 Extensions of L.; R/
9.1 ADC . . . . . . . . . . . . . . . . . .
9.2 The Pmax -extension of L.; R/ . . . .
9.2.1 The basic analysis . . . . . .
9.2.2 Martins Maximum CC .c/ . .
9.3 The Qmax -extension of L.; R/ . . . .
9.4 Changs Conjecture . . . . . . . . . .
9.5 Weak and Strong Reection Principles
9.6 Strong Changs Conjecture . . . . . .
9.7 Ideals on !2 . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
609
610
617
618
622
633
637
651
667
683
10 Further results
10.1 Forcing notions and large cardinals . . . . .
10.2 Coding into L.P .!1 // . . . . . . . . . . .
10.2.1 Coding by sets, SQ . . . . . . . . . .
10.2.2 Q.X/
max . . . . . . . . . . . . . . . .
.;/
10.2.3 Pmax . . . . . . . . . . . . . . . . .
.;;B/
. . . . . . . . . . . . . . . .
10.2.4 Pmax
10.3 Bounded forms of Martins Maximum . . .
10.4 -logic . . . . . . . . . . . . . . . . . . .
10.5 -logic and the Continuum Hypothesis . .
10.6 The Axiom ./C . . . . . . . . . . . . . .
10.7 The Effective Singular Cardinals Hypothesis
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
694
694
701
703
708
739
768
784
807
813
827
835
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
11 Questions
840
Bibliography
845
Index
849
Chapter 1
Introduction
As always I suppose, when contemplating a new edition one must decide whether
to rewrite the introduction or simply write an addendum to the original introduction.
I have chosen the latter course and so after this paragraph the current edition begins
with the original introduction and summary from the rst edition (with comments
inserted in italics and some other minor changes) and then continues beginning on
page 18 with comments regarding this edition.
The main result of this book is the identication of a canonical model in which
the Continuum Hypothesis (CH) is false. This model is canonical in the sense that
Godels constructible universe L and its relativization to the reals, L.R/, are canonical
models though of course the assertion that L.R/ is a canonical model is made in the
context of large cardinals. Our claim is vague, nevertheless the model we identify can
be characterized by its absoluteness properties. This model can also be characterized
by certain homogeneity properties. From the point of view of forcing axioms it is
the ultimate model at least as far as the subsets of !1 are concerned. It is arguably a
completion of P .!1 /, the powerset of !1 .
This model is a forcing extension of L.R/ and the method can be varied to produce a wide class of similar models each of which can be viewed as a reduction
of this model. The methodology for producing these models is quite different than
that behind the usual forcing constructions. For example the corresponding partial
orders are countably closed and they are not constructed as forcing iterations. We
provide evidence that this is a useful method for achieving consistency results, obtaining a number of results which seem out of reach of the current technology of iterated
forcing.
The analysis of these models arises from an interesting interplay between ideas
from descriptive set theory and from combinatorial set theory. More precisely it is
the existence of denable scales which is ultimately the driving force behind the arguments. Boundedness arguments also play a key role. These results contribute to a
curious circle of relationships between large cardinals, determinacy, and forcing axioms. Another interesting feature of these models is that although these models are
generic extensions of specic inner models (L.R/ in most cases), these models can be
characterized without reference to this. For example, as we have indicated above, our
canonical model is a generic extension of L.R/. The corresponding partial order we
denote by Pmax . In Chapter 5 we
give a characterization for this model isolating an
axiom . The formulation of does not involve Pmax , nor does it obviously refer to
L.R/. Instead it species properties of denable subsets of P .!1 /.
1 Introduction
The original motivation for the denition of these models resulted from the discovery that it is possible, in the presence of the appropriate large cardinals, to force
(quite by accident) the effective failure of CH. This and related results are the subject
of Chapter 3. We discuss effective versions of CH below.
Gdel was the rst to propose that large cardinal axioms could be used to settle
questions that were otherwise unsolvable. This has been remarkably successful particularly in the area of descriptive set theory where most of the classical questions have
now been answered. However after the results of Cohen it became apparent that large
cardinals could not be used to settle the Continuum Hypothesis. This was rst argued
by Levy and Solovay .1967/.
Nevertheless large cardinals do provide some insight to the Continuum Hypothesis.
One example of this is the absoluteness theorem of Woodin .1985/. Roughly this
theorem states that in the presence of suitable large cardinals CH settles all questions
with the logical complexity of CH.
More precisely if there exists a proper class of measurable Woodin cardinals then
21 sentences are absolute between all set generic extensions of V which satisfy CH.
The results of this book can be viewed collectively as a version of this absoluteness
theorem for the negation of the Continuum Hypothesis (:CH).
1.1
1 Introduction
more general problem of computing the effective size of the continuum. This problem
has a variety of formulations, two natural versions are combined in the following:
Is there a (consistent) large cardinal whose existence implies that the length of
any prewellordering arising in either of the following fashions, is less than the
least weakly inaccessible cardinal?
The prewellordering exists in a transitive inner model of AD containing all
the reals.
The prewellordering is universally Baire.
The second of these formulations involves the notion of a universally Baire set
of reals which originates in .Feng, Magidor, and Woodin 1992/. Universally Baire
sets are discussed briey in Section 10.3. We note here that if there exists a proper
class of Woodin cardinals then a set A R is universally Baire if and only if it is
1
-weakly homogeneously Suslin which in turn is if and only if it is 1 -homogeneously
Suslin. Another relevant point is that if there exist innitely many Woodin cardinals
with measurable above and if A R is universally Baire, then
L.A; R/ AD
and so A belongs to an inner model of AD. The converse can fail.
More generally one can ask for any bound provided of course that the bound is a
specic ! which can be dened without reference to 2@0 .
1
For example every
2 prewellordering has length less than !2 and if there is a
measurable cardinal then every
13 prewellordering has length less than !3 . A much
deeper theorem of .Jackson 1988/ is that if every projective set is determined then
every projective prewellordering has length less than !! . This combined with the
theorem of Martin and Steel on projective determinacy yields that if there are innitely
many Woodin cardinals then every projective prewellordering has length less than !! .
The point here of course is that these bounds are valid independent of the size of 2@0 .
The current methods do not readily generalize to even produce a forcing extension
of L.R/ (without adding reals) in which ZFC holds and !3 < L.R/ . Thus at this point
it is entirely possible that !3 is the bound and that this is provable in ZFC. If a large
cardinal admits an inner model theory satisfying fairly general conditions then most
likely the only (nontrivial) bounds provable from the existence of the large cardinal are
those provable in ZFC; i. e. large cardinal combinatorics are irrelevant unless the large
cardinal is beyond a reasonable inner model theory.
For example suppose that there is a partial order P 2 L.R/ such that for all transitive models M of ADC containing R, if G P is M -generic then
.R/M G D .R/M ,
.13 /M G D .!3 /M G ,
L.R/G ZFC,
where
13 is the supremum of the lengths of
13 prewellorderings of R. The axiom ADC
is a technical variant of AD which is actually implied by AD in many instances. Assuming DC it is implied, for example, by AD R . It is also implied by AD if V D L.R/.
By the results of .Woodin 2010b/ if inner model theory can be extended to the
level of one supercompact cardinal then the existence of essentially all large cardinals
is consistent with
13 D !3 .
It follows from the results of .Steel and VanWesep 1982/ and .Woodin 1983/ that
such a partial order P exists in the case of 12 , more precisely, assuming
L.R/ AD;
there is a partial order P 2 L.R/ such that for all transitive models M of ADC containing R, if G P is M -generic then
.R/M G D .R/M ,
.12 /M G D .!2 /M G ,
L.R/G ZFC.
Thus if a large cardinal admits a suitable inner model theory then the existence of the
large cardinal is consistent with 12 D !2 . We shall prove a much stronger result in
Chapter 3, showing that if is a Woodin cardinal and if there is a measurable cardinal
above then there is a semiproper partial order P of cardinality such that
V P 12 D !2 :
This result which is a corollary of Theorem 1.1, stated below, and Theorem 2.64, due
to Shelah, shows that this particular instance of the Effective Continuum Hypothesis is
as intractable as the Continuum Hypothesis.
Foreman and Magidor initiated a program of proving that
12 < !2 from various
combinatorial hypotheses with the goal of evolving these into large cardinal hypotheses, .Foreman and Magidor 1995/. By the (initial) remarks above their program if
successful would have identied a critical step in the large cardinal hierarchy.
Foreman and Magidor proved among other things that if there exists a (normal)
12 < !2 . In
!3 -saturated ideal on !2 concentrating on a specic stationary set then
Chapter 9 we improve this result slightly showing that this restriction is unnecessary;
if there is a measurable cardinal and if there is an !3 -saturated (uniform) ideal on !2
then
12 < !2 .
An early conjecture of Martin is that
1n D @n for all n follows from reasonable
1
1n prewellorderings.
hypotheses. n is the supremum of the lengths of
The following theorem proves the Martin conjecture in the case of n D 2.
Theorem 1.1. Assume that the nonstationary ideal on !1 is !2 -saturated and that
there is a measurable cardinal. Then
12 D !2 and further every club in !1 contains a
club constructible from a real.
t
u
1 Introduction
As a corollary we obtain,
Theorem 1.2. Assume Martins Maximum. Then
12 D !2 and every club in !1 contains a club constructible from a real.
t
u
Another immediate corollary is a renement of the upper bound for the consistency
strength of
ZFC C For every real x; x # exists. C !2 is the second uniform indiscernible.
Assuming in addition that larger cardinals exist then one obtains more information.
For example,
Theorem 1.3. Assume the nonstationary ideal on !1 is !2 -saturated and that there
exist ! many Woodin cardinals with a measurable cardinal above them all.
(1) Suppose that A R, A 2 L.R/, and that there is a sequence hB W < !1 i of
borel sets such that
A D [B j < !1 :
1
Then A is
2 .
(2) Suppose that X is a bounded subset of L.R/ of cardinality !1 . Then there exists
t
u
a set Y 2 L.R/ of cardinality !1 in L.R/ such that X Y .
We note that assuming for every x 2 R, x # exists, the statement (1) of Theorem 1.3
1
implies that
12 D !2 ; if 12 < !2 then every
3 set is an !1 union of borel sets.
1.2
Theorem 1.3 suggests that if the nonstationary ideal is saturated (and if modest large
cardinals exist) then one might reasonably expect that the inner model L.P .!1 // may
be close to the inner model L.R/. However if the nonstationary ideal is saturated one
can, by passing to a ccc generic extension, arrange that
P .R/ L.P .!1 //
and preserve the saturation of the nonstationary ideal. Nevertheless this intuition was
the primary motivation for the denition of Pmax .
The canonical model for :CH is obtained by the construction of this specic partial
order, Pmax . The basic properties of Pmax are given in the following theorem.
Theorem 1.4. Assume ADL.R/ and that there exists a Woodin cardinal with a measurable cardinal above it. Then there is a partial order Pmax in L.R/ such that;
(1) Pmax is !-closed and homogeneous (in L.R/),
(2) L.R/Pmax ZFC.
:
t
u
The partial order Pmax is denable and thus, since granting large cardinals
Th.L.R// is canonical, it follows that Th.L.R/Pmax / is canonical.
Many of the open combinatorial questions at !1 are expressible as 2 statements
in the structure
hH.!2 /; 2; INS i
and so assuming the existence of large cardinals these questions are either false, or
they are true in L.R/Pmax .
In some sense the spirit of Martins Axiom and its generalizations is to maximize
the collection of 2 sentences true in the structure
hH.!2 /; 2i
Indeed MA!1 is easily reformulated as a 2 sentence for hH.!2 /; 2i.
By the remarks above, assuming fairly weak large cardinal hypotheses, any such
sentence which is true in some set generic extension of V is true in a canonical generic
extension of L.R/.
The situation is analogous to the situation of 12 sentences and L. By Shoenelds
absoluteness theorem if a 12 sentence holds in V then it holds in L.
The difference here is that the model analogous to L is not an inner model but
rather it is a canonical generic extension of an inner model. This is not completely
unprecedented. Manselds theorem on 12 wellorderings can be reformulated as follows.
Theorem 1.5 (Manseld). Suppose that is a 13 sentence which is true in V and
there is a nonconstructible real. Then is true in LP where P is Sacks forcing .dened
in L/.
t
u
Of course the 13 sentence also holds in L so this is not completely analogous to
our situation. :CH is a (consistent) 2 sentence for hH.!2 /; 2i which is false in any
of the standard inner models.
Nevertheless the analogy with Sacks forcing is accurate. The forcing notion Pmax
is a generalization of Sacks forcing to !1 .
The following theorem, slightly awkward in formulation, shows that any attempt to realize in H.!2 / all suitably consistent 2 sentences, requires at least 12 Determinacy.
1 Introduction
Theorem 1.6. Suppose that there exists a model, hM; Ei, such that
hM; Ei ZFC
and such that for each 2 sentence if there exists a partial order P such that
hH.!2 /; 2iV
then
;
hH.!2 /; 2ihM;E i :
t
u
Pmax
:
t
u
Pmax
then
hH.!2 /; 2; INS ; X I X 2 L.R/; X Ri :
Then
L.P .!1 // D L.R/G
for some G Pmax which is L.R/-generic.
t
u
If one assumes in addition that R# exists then Theorem 1.8 can be reformulated as
follows. For each n 2 ! let Un be a set which is 1 denable in the structure
hL.R/; hi W i < ni; 2i
where hi W i < ni is an increasing sequence of Silver indiscernibles of L.R/, and
such that Un is universal.
Theorem 1.9. Assume ADL.R/ and that R# exists. Suppose that for each 2 sentence
in the language for the structure
hH.!2 /; 2; INS ; Un I n < !i
if
Pmax
hH.!2 /; 2; INS ; Un I n < !iL.R/
then
hH.!2 /; 2; INS ; Un I n < !i :
Then
L.P .!1 // D L.R/G
for some G Pmax which is L.R/-generic.
t
u
Thus in the statement of Theorem 1.9 one only refers to a structure of countable
signature.
These theorems suggest that the axiom:
./ AD holds in L.R/ and L.P .!1 // is a Pmax -generic extension of L.R/;
is perhaps, arguably, the correct maximal generalization of Martins Axiom at least as
far as the structure of P .!1 / is concerned. However an important point is that we do
not know if this axiom can always be forced to hold assuming the existence of suitable
large cardinals.
Conjecture. Assume there are ! 2 many Woodin cardinals. Then the axiom ./ holds
in a generic extension of V .
t
u
Because of the intrinsics of the partial order Pmax , this axiom is frequently easier to
use than the usual forcing axioms. We give some applications for which it is not clear
that Martins Maximum sufces. Another key point is:
There is no need in the analysis of L.R/Pmax for any machinery of iterated forcing. This includes the proofs of the absoluteness theorems.
Further
The analysis of L.R/Pmax requires only ADL.R/ .
For the denition of Pmax that we shall work with the analysis will require some iterated
forcing but only for ccc forcing and only to produce a poset which forces MA!1 .
In Chapter 5 we give three other presentations of Pmax based on the stationary
tower forcing. The analysis of these (essentially equivalent) versions of Pmax require
no local forcing arguments whatsoever. This includes the proof of the absoluteness
theorems.
10
1 Introduction
Also in Chapter 5 we shall discuss methods for exploiting ./, giving a useful
reformulation of the axiom. This reformulation does not involve the denition of Pmax .
We shall also prove that, assuming ./,
L.P .!1 // AC:
This we accomplish by nding a 2 sentence which if true in the structure,
hH.!2 /; 2i;
implies (in ZF C DC) that there is a surjection
W !2 ! R
which is denable in the structure
hH.!2 /; 2i
from parameters. This sentence is a consequence of Martins Maximum and an analogous, but easier, argument shows that assuming ADL.R/ , it is true in L.R/Pmax . Thus
the axiom ./ implies 2@0 D @2 . Actually we shall discuss two such sentences, AC
and AC . These are dened in Section 5.1 and Section 5.3 respectively.
11
Again ADL.R/ sufces for the analysis of L.R/Qmax and there are absoluteness theorems which characterize the Qmax -extension.
Collectively these results suggest that the consistency of ADL.R/ is an upper bound
for the consistency strength of many propositions at !1 , over the base theory,
ZFC C For all x 2 R, x # exists C 12 D !2 :
However there are two classes of counterexamples to this.
Suppose that R# exists and that L.R# / AD. For each sentence such that
L.R/ ;
the following:
There exists a sequence, hB; W < < !1 i, of borel sets such that
[ \
R# D
B; ;
>
and
L.R/ AD C ,
can be expressed by a 2 sentence in hH.!2 /; 2i which can be realized by forcing
with a Pmax variation over L.R# /. There must exist a choice of such that this 2
sentence cannot be realized in the structure hH.!2 /; 2i of any set generic extension
of L.R/. This is trivial if the extension adds no reals (take to be any tautology),
otherwise it is subtle in that if
L.R/ AD
then we conjecture that there is a partial order P 2 L.R/ such that
L.R/P ZFC C R# exists:
The second class of counterexamples is a little more subtle, as the following example illustrates. If the nonstationary ideal on !1 is !1 -dense and if Changs Conjecture
holds then there exists a countable transitive set, M , such that
M ZFC C There exist ! C 1 many Woodin cardinals;
(and so M ADL.R/ and much more). The application of Changs Conjecture is only
necessary to produce
X 2 H.!2 /
such that X \ !2 has ordertype !1 . The subtle and interesting aspect of this example
is that
L.R/Qmax Changs Conjecture;
but by the remarks above, this can only be proved by invoking hypotheses stronger
than ADL.R/ .
In fact the assertion,
L.R/Qmax Changs Conjecture,
is equivalent to a strong form of the consistency of AD. This is the subject of Section 9.4.
12
1 Introduction
13
obtainable with the usual methods. The rst example of Chapter 7 is the variation,
Smax , which conditions the model on a sentence which implies the existence of a Suslin
tree. The sentence asserts:
Every subset of !1 belongs to a transitive model M in which holds and such
that every Suslin tree in M is a Suslin tree in V .
If AD holds in L.R/ and if G Smax is L.R/-generic then in L.R/G the following
strengthening of the sentence holds:
For every A !1 there exists B !1 such that A 2 LB and such that if
T 2 LB is a Suslin tree in LB, then T is a Suslin tree.
In L.R/G every subset of !1 belongs to an inner model with a measurable cardinal
(and more) and under these conditions this strengthening is not even obviously consistent.
The second example of Chapter 7 is motivated by the Borel Conjecture. The rst
consistency proof for the Borel Conjecture is presented in .Laver 1976/. The Borel
Conjecture can be forced a variety of different ways. One can iterate Laver forcing or
Mathias forcing, etc. In Section 7.2, we dene a variation of Pmax which forces the
Borel Conjecture. The denition of this forcing notion does not involve Laver forcing, Mathias forcing or any variation of these forcing notions. In the model obtained,
a version of Martins Maximum holds. Curiously, to prove that the Borel Conjecture
holds in the resulting model we do use a form of Laver forcing. An interesting technical question is whether this can be avoided. It seems quite likely that it can, which
could lead to the identication of other variations yielding models in which the Borel
Conjecture holds and in which additional interesting combinatorial facts also hold.
1.4
In Chapter 9 we again focus primarily on the Pmax -extension but now consider extensions of inner models strictly larger than L.R/. These yield models of ./ with rich
structure for H.!3 /; i. e. with many subsets of !2 .
The ground models that we shall consider are of the form L.; R/ where
P .R/ is a pointclass closed under borel preimages, or more generally inner
models of the form L.S; ; R/ where P .R/ and S Ord. We shall require that
a particular form of AD hold in the inner model, the axiom is ADC which is discussed
in Section 9.1. It is by exploiting more subtle aspects of the consequences of ADC that
we can establish a number of combinatorially interesting facts about the corresponding
extensions.
Applications include obtaining extensions in which Martins Maximum holds for
partial orders of cardinality c, this is Martins Maximum.c/, and in which !2 exhibits
some interesting combinatorial features.
14
1 Introduction
Actually in the models obtained, Martins MaximumCC .c/ holds. This is the assertion that Martins MaximumCC holds for partial orders of cardinality c where Martins
MaximumCC is a slight strengthening of Martins Maximum. These forcing axioms,
rst formulated in .Foreman, Magidor, and Shelah 1988/, are dened in Section 2.5.
Recasting the Pmax variation for the Borel Conjecture in this context we obtain, in
the spirit of Martins Maximum, a model in which the Borel Conjecture holds together
with the largest fragment of Martins Maximum.c/ which is possibly consistent with
the Borel Conjecture.
Another reason for considering extensions of inner models larger than L.R/ is that
one obtains more information about extensions of L.R/. For example the proof that
L.R/Qmax Changs Conjecture;
requires considering the .Qmax /N -extension of inner models N such that
.R \ N /# 2 N
and much more.
Finally any systematic study of the possible features of the structure
hH.!2 /; INS ; 2i
in the context of
ZFC C ADL.R/ C 12 D !2
requires considering extensions of inner models beyond L.R/; as we have indicated,
there are (2 ) sentences which can be realized in the structure, hH.!2 /; INS ; 2i, of
these extensions but which cannot be realized in any such structure dened in an extension of L.R/.
The results of Chapter 9 suggest a strengthening of the axiom ./:
Axiom ./C : For each set X !2 there exists a set A R and a lter G Pmax
such that
(1) L.A; R/ ADC ,
(2) G is L.A; R/-generic and X 2 L.A; R/G.
This is discussed briey in Chapter 10 which explores the possible relationships between Martins Maximum and the axiom ./. One of the theorems we shall prove
Chapter 10 shows that in Theorem 1.8, it is essential that the predicate, INS , for the
nonstationary sets be added to the structure. We shall show that
Martins Maximum CC .c/ C Strong Changs Conjecture
together with all the 2 consequences of ./ for the structure
hH.!2 /; Y; 2 W Y R; Y 2 L.R/i
does not imply ./. We shall also prove an analogous theorem which shows that conally many sets from P .R/ \ L.R/ must be added; for each set Y0 2 P .R/ \ L.R/,
Martins Maximum CC .c/ C Strong Changs Conjecture
15
1.5
t
u
One of the main open problems of Shelahs pcf theory is whether there can exist a
set, A, of regular cardinals such that jAj < jpcf.A/j (satisfying the usual requirement
that jAj < min.A/).
Common to all Pmax variations is that Theorem 1.3(2) holds in the resulting models
and so the conclusions of Theorem 1.10 applies to these models as well. Though,
16
1 Introduction
recently, a more general class of variations has been identied for which Theorem 1.3(2) fails in the models obtained. These latter examples are variations only in
the sense that they also yield canonical models in which CH fails, cf. Theorem 10.185.
I end with a confession. This book was written intermittently over a 7 year period
beginning in early 1992 when the initial results were obtained. During this time the
exposition evolved considerably though the basic material did not. Except that the
material in Chapter 8, the material in the last three sections of Chapter 9 and much
of Chapter 10, is more recent. Earlier versions contained sections which, because of
length considerations, we have been compelled to remove.
This account represents in form and substance the evolutionary process which actually took place. Further a number of proofs are omitted or simply sketched, especially
in Chapter 10. Generally it seemed better to state a theorem without proof than not to
state it at all. In some cases the proofs are simply beyond the scope of this book and in
other cases the proofs are a routine adaptation of earlier arguments. Of course in both
cases this can be quite frustrating to the reader. Nevertheless it is my hope that this
book does represent a useful introduction to this material with few relics from earlier
versions buried in its text.
By the time (May, 1999) of this writing a number of papers have appeared, or are
in press, which deal with Pmax or variations thereof. P. Larson and D. Seabold have
each obtained a number of results which are included in their respective Ph. D. theses,
some of these results are discussed in this book.
Shelah and Zapletal consider several variations, recasting the absoluteness theorems in terms of 2 -compactness but restricting to the case of extensions of L.R/,
.Shelah and Zapletal 1999/.
More recently Ketchersid, Larson, and Zapletal .2007/ isolate a family of explicit
Namba-like forcing notions which can, under suitable circumstances, change the value
of
12 even in situations where CH holds. These examples are really the rst to be
isolated which can work in the context of CH. Other examples have been discovered
and are given in .Doebler and Schindler 2009/.
Finally there are some very recent developments (as of 1999) which involve a generalization of !-logic which we denote -logic. Arguably -logic is the natural limit
of the lineage of generalizations of classical rst order logic which begins with !-logic
and continues with -logic etc.
We (very briey) discuss -logic (updated to 2010) in Section 10.4 and Section 10.5. In some sense the entire discussion of Pmax and its variations should take
place in the context of -logic and were we to rewrite the book this is how we would
proceed. In particular, the absoluteness theorems associated to Pmax and its variations
are more naturally stated by appealing to this logic. For example Theorem 1.4 can be
reformulated as follows.
Theorem 1.11. Suppose that there exists a proper class of Woodin cardinals.
Suppose that is a 2 sentence in the language for the structure hH.!2 /; 2; INS i
and that
ZFC C hH.!2 /; 2; INS i
is -consistent, then
Pmax
:
17
t
u
In fact, using -logic one can give a reformulation of ./ which does not involve
forcing at all, this is discussed briey in Section 10.4.
Another feature of the forcing extensions given by the (homogeneous) Pmax variations, this holds for all the variations which we discuss in this book, is that each
provides a nite axiomatization, over ZFC, of the theory of H.!2 / (in -logic). For
Pmax , the axiom is ./ and the theorem is the following.
Theorem 1.12. Suppose that there exists a proper class of Woodin cardinals.
Then for each sentence , either
(1) ZFC C ./ ` H.!2 / , or
(2) ZFC C ./ ` H.!2 / :.
t
u
This particular feature underscores the fundamental difference between the method
of Pmax variations and that of iterated forcing. We note that it is possible to identify
nite axiomatizations over ZFC of the theory of hH.!2 /; 2i which cannot be realized
by any Pmax variation. Theorem 10.185 indicates such an example, the essential feature
is that 12 < !2 but still there is an effective failure of CH. Nevertheless it is at best
difcult through an iterated forcing construction to realize in hH.!2 /; 2iV G a theory
which is nitely axiomatized over ZFC in -logic. The reason is simply that generally
the choice of the ground model will inuence, in possibly very subtle ways, the theory
of the structure hH.!2 /; 2iV G . There is at present no known example which works,
say from some large cardinal assumption, independent of the choice of the ground
model.
-logic provides the natural setting for posing questions concerning the possibility
of such generalizations of Pmax , to for example !2 , i. e. for the structure H.!3 /, and
beyond. The rst singular case, H.!!C /, seems particularly interesting.
There is also the case of !1 but in the context of CH. One interesting result (but
as of 2010, this is contingent on the ADC Conjecture), with, we believe, potential
implications for CH, is that there are limits to any possible generalization of the Pmax
variations to the context of CH; more precisely, if CH holds then the theory of H.!2 /
cannot be nitely axiomatized over ZFC in -logic.
Acknowledgments to the rst edition. Many of the results of the rst half of this
book were presented in the Set Theory Seminar at UC Berkeley. The (ever patient)
participants in this seminar offered numerous helpful suggestions for which I remain
quite grateful.
I am similarly indebted to all those willingly to actually read preliminary versions
of this book and then relate to me their discoveries of mistakes, misprints and relics.
I only wish that the nal product better represented their efforts.
I owe a special debt of thanks to Ted Slaman. Without his encouragement, advice
and insight, this book would not exist.
18
1 Introduction
The research, the results of which are the subject of this book, was supported in part
by the National Science Foundation through a succession of summer research grants,
and during the academic year, 19971998, by the Miller Institute in Berkeley.
Finally I would like to acknowledge the (generous) support of the Alexander von
Humboldt Foundation. It is this support which enabled me to actually nish this book.
Berlin, May 1999
1.6
W. Hugh Woodin
In the 10 years since what was written above as the introduction to the rst edition of
this book there have been quite a number of mathematical developments relevant to
this book and I nd myself again in Germany on sabbatical from Berkeley working on
this book. This edition contains revisions that reect these developments including the
deletion of some theorems now not relevant because of these developments or simply
because the proofs, sketched or otherwise, were simply not correct. Finally I stress that
I make no claim that this revision is either extensive or thorough and I regret to say that
it is not I feel that the entire subject is at a critical crossroads and as always in such a
situation one cannot be completely condent in which direction the future lies. But it
is this future that dictates which aspects of this account should be stressed.
First and most straightforward, the theorems related to ! .!2 /, such as the theorem that Martins Maximum implies ! .!2 /, have all been rendered irrelevant by a
remarkable theorem of .Shelah 2008/ which shows that ! .!2 / is a consequence of
2!1 D !2 . Shelahs result shows that assuming Martins Maximum.c/, or simply assuming that 2!1 D !2 , then the nonstationary ideal at !2 cannot be semi-saturated on
the ordinals of countable conality. It does not rule out the possibility that there exists
a uniform semi-saturated at !2 on the ordinals of countable conality. On the other
hand, the primary motivation for obtaining such consistency results for ideals at !2 in
the rst edition was the search for evidence that the consistency strength of the theory
ZF C ADR C is regular
was beyond that of the existence of a superstrong cardinals. Dramatic recent results
.Sargsyan 2009/ have shown that this theory is not that strong, proving that the consistency of this theory follows from simply the existence of a Woodin cardinal which is
a limit of Woodin cardinals. Therefore in this edition the consistency results for semisaturated ideals at !2 are simply stated without proof. The proofs of these theorems
are sketched at length in the rst edition but based upon an analysis in the context of
ADC of HOD which is open without requiring that one work relative to the minimum
model of
ZF C ADR C is Mahlo
but of course the sketch in the case of obtaining the consistency that JNS is semisaturated is not correct that error was due to a careless misconception regarding
19
iterations of forcing with uncountable support. As indicated in the rst edition the
analysis of HOD in the context of ADC is not actually necessary for the proofs, it was
used only to provide a simpler framework for the constructions.
Ultimately of far more signicance for this book is that recent results concerning
the inner model program undermine the philosophical framework for this entire work.
The fundamental result of this book is the identication of a canonical axiom for :CH
which is characterized in terms of a logical completion of the theory of H.!2 / (in logic of course). But the validation of this axiom requires a synthesis with axioms for
V itself for otherwise it simply stands as an isolated axiom. This view is reinforced
by the use of the Conjecture to argue against the generic-multiverse view of truth
.Woodin 2009/. I remain convinced that if CH is false then the axiom ./ holds and
certainly there are now many results conrming that if the axiom ./ does hold then
there is a rich structure theory for H.!2 / in which many pathologies are eliminated.
But nevertheless for all the reasons discussed at length in .Woodin 2010b/, I think the
evidence now favors CH.
The picture that is emerging now based on .Woodin 2010b/ and .Woodin 2010a/
is as follows. The solution to the inner model problem for one supercompact cardinal
yields the ultimate enlargement of L. This enlargement of L is compatible with all
stronger large cardinal axioms and strong forms of covering hold relative to this inner
model. At present there seem to be two possibilities for this enlargement, as an extender model or as strategic extender model. There is a key distinction however between
these two versions. An extender model in which there is a Woodin cardinal is a (nontrivial) generic extension of an inner model which is also an extender model whereas
a strategic extender model in which there is a proper class of Woodin cardinals is not
a generic extension of any inner model. The most optimistic generalizations of the
structure theory of L.R/ in the context of AD to a structure theory of L.VC1 / in the
context of an elementary embedding,
j W L.VC1 / ! L.VC1 /
with critical point below
require that V not be a generic extension of any inner model
which is not countably closed within V . Therefore these generalizations cannot hold
in the extender models and this leave the strategic extender models as essentially the
only option. Thus there could be a compelling argument that V is a strategic extender
model based on natural structural principles. This of course would rule out that the
axiom ./ holds though if V is a strategic extender model (with a Woodin cardinal)
then the axiom ./ holds in a homogeneous forcing extension of V and so the axiom
./ has a special connection to V as an axiom which holds in a canonical companion
to V mediated by an intervening model of ADC which is the manifestation of -logic.
An appealing aspect to this scenario is that the relevant axiom for V can be explicitly
stated now and in a form which claries the previous claims without knowing
the detailed level by level inductive denition of a strategic extender model .Woodin
2010b/: in its weakest form the axiom is simply the conjunction of:
20
1 Introduction
W. Hugh Woodin
Chapter 2
Preliminaries
We briey review, without giving all of the proofs, some of the basic concepts which
we shall require, .Foreman and Kanamori (Eds.) 2010/ covers most of what we need
and obviously quite a bit more. In the course of this we shall x some notation. As
is the custom in Descriptive Set Theory, R denotes the innite product space, ! ! .
Though sometimes it is convenient to work with the Cantor space, 2! , or even with
the standard Euclidean space, .
1; 1/. If at some point the discussion is particularly
sensitive to the manifestation of R then we may be more careful with our notation. For
example L.R/ is relatively immune to such considerations, but Wadge reducibility is
not.
We shall require at several points some coding of sets by reals or by sets of reals.
There is a natural coding of sets in H.!1 / (the hereditarily countable sets) by reals.
For example if a 2 H.!1 / then the set a can be coded by coding the structure
hb [ !; a; 2i
where b is the transitive closure of a.
A real x codes a if x decodes sets A ! and E ! ! such that
hb [ !; a; 2i h!; A; Ei;
where again b is the transitive closure of a.
Suppose that M 2 H.c C / and let N be the transitive closure of M . Fix a reasonable decoding of a set X R to produce an element of
P .R/ P .R R/ P .R R/:
A set X R codes M if X decodes sets A R, E R R and R R such
that is an equivalence relation on R, A R, E R R, A and E are invariant
relative to , and such that
hN; M; 2i hR=; A=; E=i:
We shall be interested in sets M which are coded in this fashion by sets X R such
that X belongs to a transitive inner model in which the Axiom of Choice fails.
2.1
For any set X , X <! is the set of nite sequences of elements of X . If s 2 X <! then
`.s/ denotes the length of s, which formally is simply the domain of s. A tree T on
a set X is a set of nite sequences from X which is closed under initial segments. So
T X <! .
We abuse this convention slightly and say that a tree T is a tree on ! where
is an ordinal if T is a set of pairs .s; t / such that
22
2 Preliminaries
Tx D [Txjk j k 2 !:
23
t
u
24
2 Preliminaries
25
We come to the key notions of homogeneous trees and weakly homogeneous trees.
These denitions are due independently to Kunen and Martin.
Denition 2.4. Suppose that is an ordinal and 0. Suppose that T is a tree on
! .
(1) The tree T is -weakly homogeneous if there is a partial function
W ! <! ! <! ! m. <! /
such that
a) if .s; t / 2 dom./ then .s; t /.Ts / D 1 and .s; t / is a -compete measure,
b) for all x 2 ! ! , x 2 pT if and only if there exists y 2 ! ! such that
.xjk; yjk/ j k < ! dom./,
h.xjk; yjk/ W k 2 !i is a countably complete tower.
(2) The tree T is <-weakly homogeneous if T is -weakly homogeneous for all
< .
(3) The tree T is weakly homogeneous if T is -weakly homogeneous for some . u
t
Denition 2.5. Suppose that is an ordinal and 0. Suppose that T is a tree on
! .
(1) The tree T is -homogeneous if there is a partial function
W ! <! ! m. <! /
such that
a) if s 2 dom./ then .s/.Ts / D 1 and .s/ is a -compete measure,
b) for all x 2 ! ! , x 2 pT if and only if
xjk j k 2 ! dom./,
h.xjk/ W k 2 !i is a countably complete tower.
(2) The tree T is <-homogeneous if T is -homogeneous for all < .
(3) The tree T is homogeneous if T is -homogeneous for some .
t
u
Any tree on ! ! is -weakly homogeneous for all and similarly any tree on
! 1 is -homogeneous for all . In each case the associated measures are principal.
The denition of a weakly homogeneous tree has a simple reformulation which
is frequently more relevant to the process of actually verifying that specic trees are
weakly homogeneous. This reformulation is given in the following lemma which we
leave as an exercise.
26
2 Preliminaries
t
u
27
Denition 2.10. (1) WH is the set of all A R such that A is -weakly homoWH
is the set of all A R such that A is < -weakly
geneously Suslin. <
WH
homogeneously Suslin. 1
is the set of all A R such that A is -weakly
homogeneously Suslin for all .
H
(2) H is the set of all A R such that A is -homogeneously Suslin. <
is the
H
set of all A R such that A is <-homogeneously Suslin. 1 is the set of all
A R such that A is -homogeneously Suslin for all .
t
u
The next lemma gives the elementary closure properties for these pointclasses.
Lemma 2.11.
tions.
(2) WH is closed under continuous preimages, continuous images, countable intersections and countable unions.
(3) H WH .
(4) If 1 < 2 then
WH
WH
2
1
and
H2 H1 :
t
u
t
u
We shall need the following theorem, .Koellner and Woodin 2010/. This theorem
can be used in place of the Martin-Steel theorem on scales in L.R/, Theorem 2.3, in
the analysis of L.R/Pmax .
Theorem 2.13. Suppose that is a limit of Woodin cardinals and that there a measurable cardinal above . Suppose that A R and that A 2 L.R/. Then A is < weakly
homogeneously Suslin.
t
u
The basic machinery for establishing that sets are weakly homogeneously Suslin
is developed in Larson .2004/. An important application is given in the following
theorem of Steel.
28
2 Preliminaries
Theorem 2.14 (Steel). Suppose that 0 < 1 are Woodin cardinals and
A 2 WH
C:
1
t
u
WH
Then A has a scale in <
.
0
The fundamental theorem of .Martin and Steel 1989/ implies that if is a Woodin
cardinal then
H
WH
C < :
An immediate corollary to this is the following theorem which is extremely useful
in developing the elementary theory of these pointclasses.
Theorem 2.15. Suppose that is a limit of Woodin cardinals. Then
WH
H
<
D <
:
t
u
Remark 2.17. (1) Suppose that 0 1 are pointclasses which are closed under
continuous preimages. Suppose that 1 is a -algebra and is closed under continuous images. Suppose every set in 1 is the continuous image of a set in 0 .
Suppose every set in 1 is determined. Then 0 D 1 . Therefore if is a limit
WH
H
and <
follows abstractly from
of Woodin cardinals, the equivalence of <
WH
.
the determinacy of the sets in <
(2) Suppose that is strongly compact. Then by the results of .Larson 2004/ the
pointclass WH is a -algebra with the property that every set in WH admits
a scale in WH . Further WH has very strong closure properties. For example,
2
assuming CH, then if A 2 WH then every set which is
1 denable from A is in
WH
. An interesting open question is the following.
Suppose that is strongly compact. Must WH D H ?
This is equivalent to the question of whether every -weakly homogeneously
Suslin set is determined (given that is strongly compact).
t
u
It is convenient in many situations to associate with a pointclass P .! ! / a
transitive set M . Roughly M is simply the set of all sets X which are coded by a set
in . For technical reasons we actually dene M to be a possibly smaller set, though
in practice this distinction will never really be important to us. It does however raise
an interesting question.
29
t
u
Clearly M and N are each transitive. With our coding conventions N is simply
the set of all sets X which are coded by a set in .
Remark 2.19. Suppose that P .R/ is a pointclass as in Denition 2.18.
(1) Suppose that Y 2 N is transitive and that
hY; 2i hR=; E=i
where
a) is an equivalence relation on R,
b) E is a binary relation on R,
c) ; E are each in .
Let W R ! Y be the associated surjection. Then 2 N .
(2) M is a transitive set which is closed under the Godel operations. Even with
t
u
determinacy assumptions on we do not know if this is true of N .
Remark 2.20.
30
2 Preliminaries
t
u
Remark 2.23. (1) We do not know if one can prove that M D N , assuming
either every set in is determined or even assuming
L.; R/ AD:
(2) Note that if 0 1 are each (boolean) pointclasses closed under continuous
images and preimages then
N0 N1 :
However the relationship between M0 and M1 is less clear, even with determinacy assumptions.
31
2.2
Generic absoluteness
Suppose A R is -weakly homogeneously Suslin. Then the set A has an unambiguous interpretation in V G where G is V -generic for a partial order in V . The
interpretation is independent of the choice of the representation of A as the projection
of a tree which is -weakly homogeneous. This is an immediate consequence of the
next two lemmas.
Lemma 2.26. Suppose T is a tree on ! and T is -weakly homogeneous. Then
there is a tree S on ! .2 /C such that if P 2 V is a partial order and G P is
V -generic then
t
u
.pT /V G D RV G n .pS /V G :
32
2 Preliminaries
t
u
Thus B 2 .
Suppose P 2 V is a partial order and that G P is V -generic. Let AG and BG
be the interpretations of A and B in V G. Then
BG D t 2 R j hH.!1 /V G ; AG ; 2i t; a :
This is an easy consequence of Lemma 2.26 and Lemma 2.27. Alternate formulations are given in the next two lemmas.
Lemma 2.28. Suppose A R and let B R be the set of reals which code elements
of the rst order diagram of the structure
hH.!1 /; 2; Ai:
Suppose S and T are trees on ! such that
(1) S and T are -weakly homogeneous,
(2) A D pS and B D pT .
Suppose P 2 V and G P is V -generic. Let AG D pS and let BG D pT , each
computed in V G. Then in V G, BG is the set of reals which code elements of the
rst order diagram of the structure
hH.!1 /V G ; 2; AG i:
t
u
Lemma 2.29. Suppose A R and suppose that each set B R which is projective
in A, is -weakly homogeneously Suslin.
Suppose Z V is a countable elementary substructure such that C ! < ,
2 Z and such that A 2 Z.
Let MZ be the transitive collapse of Z and let Z be the image of under the
collapsing map. Suppose P 2 .MZ /Z is a partial order and that g P is MZ generic.
33
Then
(1) A \ MZ g 2 MZ g,
(2) hV!C1 \ MZ g; A \ MZ g; 2i hV!C1 ; A; 2i.
Suppose further that
A 2 .WH /V :
Then
A \ MZ g 2 .WH
/MZ g :
Z
t
u
Suppose that A R and that every set B 2 P .R/ \ L.A; R/ is -weakly homogeneously Suslin. Then .A; R/# is -weakly homogeneously Suslin. This is easily
veried by noting that .A; R/# is a countable union of sets in L.A; R/.
This observation yields the following generic absoluteness theorem.
Theorem 2.30. Suppose that A R and that every set in P .R/\L.A; R/ is -weakly
homogeneously Suslin. Suppose that T is a -weakly homogeneous tree such that
A D pT
and that P 2 V is a partial order.
Suppose that G P is V -generic. Then there is a generic elementary embedding
jG W L.A; R/ ! L.AG ; RG /
such that
(1) jG .A/ D AG D pT V G ,
(2) RG D RV G ,
(3) L.AG ; RG / D jG .f /.a/ j a 2 RG ; f W R ! L.A; R/ and f 2 L.A; R/.
Further the properties (1)(3) uniquely specify jG .
t
u
t
u
34
2 Preliminaries
The next theorem shows, in essence, that the key property of weakly homogeneous
trees given in Lemma 2.26 is equivalent in the presence of large cardinals to weak
homogeneity.
Theorem 2.32. Suppose is a Woodin cardinal. Suppose that S and T are trees on
! such that if G Coll.!; / is V -generic then,
.pT /V G D RV G n .pS /V G :
Then S and T are each <-weakly homogeneous.
2.3
t
u
We briey review some of the basic facts concerning the stationary tower forcing.
Denition 2.33.
t
u
35
Denition 2.34 (Stationary Tower). Suppose a and b are stationary sets. Then a b
if [b [a and
\ .[b/ j 2 a b:
(1) For each ordinal , P< , is the partial order given by,
P< D a 2 V j a is stationary:
(2) For each ordinal , Q< , is the partial order given by,
Q< D a 2 V j a is stationary and a P!1 .[a/:
t
u
Remark 2.35. This generalization of the notion of a stationary set appears in Woodin
.1985/, where it is exploited in this generality and where the stationary tower is introduced. The idea for generalizing the notion of a stationary set in this fashion originates
in work of Shelah. The motivation for some of the key denitions relating to the stationary tower is from consideration of the results of Foreman, Magidor, and Shelah
.1988/. An expanded treatment can be found in .Larson 2004/.
t
u
There are numerous variations of P< . The partial order Q< is one such example.
The others are dened in a similar fashion as suborders of P< .
Except in the proof of Theorem 9.68, we shall need to use only Q< . Suppose
G Q<
is V -generic. For each a 2 G, G denes in V G an ultralter Ua on V \ P .a/. The
ultralter is simply
Ua D G \ b a j b 2 G and [ b D [a:
This in turn yields an elementary embedding
ja W V ! .Ma ; Ea /
where Ma D Ult .V; Ua /. If a < b and a 2 G then there is a natural embedding
jb;a W Mb ! Ma
and this denes a directed system. The verication relies on (i).
Let .M; E/ be the limit and let
j W V ! .M; E/
be the resulting embedding. It is straightforward to verify the following, each of which
is a consequence of (ii).
(1) For all x 2 V , there exists y 2 M such that
t 2 M j t E y D j.a/ j a 2 x:
(2) For all a 2 Q< , a 2 G if and only if there exists y 2 M such that y E j.a/
and such that
t 2 M j t E y D j.b/ j b 2 [a:
36
2 Preliminaries
Suppose .M; E/ is wellfounded and let N be the transitive collapse of .M; E/. In
this case (1) asserts that for each x 2 V , j x 2 N . Therefore for each < ,
j jV 2 N and so by (2), G \ V 2 N .
If .M; E/ is not wellfounded these conclusions still hold. (1) implies that for each
< , V belongs to the wellfounded part of .M; E/ and so by (2), G \ V also
belongs to the wellfounded part of .M; E/.
The next theorem indicates a key inuence of large cardinals.
Theorem 2.36. Suppose is a Woodin cardinal and that G Q< is V -generic. Let
j W V ! .M; E/
be the induced generic elementary embedding. Then .M; E/ is wellfounded and further
N < N
in V G where N is the transitive collapse of .M; E/.
t
u
Remark 2.37. (1) Theorem 2.36 holds for P< and this leads to a variety of unusual
forcing effects. For example if is a Woodin cardinal and if is a measurable
cardinal below , then in a forcing extension of V which adds no new bounded
subsets to , it is possible to collapse C to and preserve the measurability
of .
(2) Theorem 2.36 can be proved from a variety of large cardinal assumptions. For
example it follows from the assumption that is strongly compact.
t
u
2.4
Forcing Axioms
We briey survey some of the forcing axioms which we shall be interested in.
Suppose that P is a partial order, 2 V P is a term, and that G P is a V -generic
lter. Then IG . / denotes the interpretation of in V G given by G.
Denition 2.38 (Shelah). Suppose that P is a partial order.
(1) P is proper if for all sufciently large ; if
X H. C /
is a countable elementary substructure with P 2 X , then for each p0 2 P \ X
there exists p1 2 P such that p1 p0 and such that for each term
2 V P \ X;
if G P is V -generic with p1 2 G then either IG ./ Ord or IG ./ 2 X .
37
t
u
t
u
There are several equivalent denitions of proper partial orders. One elegant version is given in the next lemma.
Lemma 2.40 (Shelah). Suppose that P is a partial order. The following are equivalent.
(1) P is proper.
(2) For all stationary sets a such that a P!1 .[a/,
V P a is stationary:
t
u
Denition 2.41. (1) (Baumgartner, Shelah) Proper Forcing Axiom .PFA/: Suppose
that P is a proper partial order and that D P .P / is a collection of dense
subsets of P with
jDj !1 :
Then there exists a lter F P such that
F \D ;
for all D 2 D.
38
2 Preliminaries
t
u
\ V:
t
u
t
u
(2) SPFA.
There are several variations of these forcing axioms which we shall be interested
in. We restrict our attention to variations of Martins Maximum.
Denition 2.45 (ForemanMagidorShelah). (1) Martins MaximumC : Suppose
that P is a partial order which is stationary set preserving.
Suppose that D P .P / is a collection of dense subsets of P with
jDj !1 ;
P
and that 2 V is a term for a stationary subset of !1 . Then there exists a lter
F P such that:
a) for all D 2 D, F \ D ;;
b) < !1 j for some p 2 F ; p 2 is stationary in !1 .
39
(2) Martins MaximumCC : Suppose that P is a partial order which is stationary set
preserving.
Suppose that D P .P / is a collection of dense subsets of P with
jDj !1 ;
and that h W < !1 i is a sequence of terms for stationary subsets of !1 . Then
there exists a lter F P such that:
a) For all D 2 D, F \ D ;;
b) For each < !1 ,
< !1 j for some p 2 F ; p 2
t
u
is stationary in !1 .
The following lemma notes useful consequences of these axioms which are quite
relevant to the themes of this book. These consequences of Martins Maximum and
of Martins MaximumCC are not equivalences; however they are equivalences for
bounded versions of these forcing axioms, see Lemma 10.93 and Lemma 10.94 of
Section 10.3.
Lemma 2.46. Suppose that P is a partial order which is stationary set preserving.
(1) (Martins Maximum) Then
P
t
u
Denition 2.47 (ForemanMagidorShelah). (1) Martins MaximumC .c/: Martins MaximumC holds for partial orders P with jP j c.
(2) Martins MaximumCC .c/: Martins MaximumCC holds for partial orders P with
jP j c.
t
u
Remark 2.48. One can naturally dene SPFA.c/. One subtle aspect of the equivalence of Martins Maximum and SPFA is that Martins Maximum.c/ is not equivalent
to SPFA.c/; Martins Maximum.c/ implies that INS (the nonstationary ideal on !1 )
is !2 -saturated whereas SPFA.c/ does not. One strong indication of the difference
follows from the results of Section 9.5:
Assume Martins Maximum.c/. Then Projective Determinacy holds.
The consistency of SPFA.c/ can be obtained from that of the existence of a strong
t
u
cardinal and so SPFA.c/ does not imply even
12 -Determinacy.
40
2 Preliminaries
We end this section with the denition of a somewhat technical variation of Martins Maximum.c/. For many applications where Martins Maximum.c/ is used, this
variation sufces. For example, it implies that INS is !2 -saturated. However we shall
see in Section 9.2.2 that this forcing axiom is (probably) signicantly weaker than
Martins Maximum.c/. We require the following denition.
Denition 2.49. Suppose P D .R; <P / is a partial order of cardinality c. The partial
order P is absolutely stationary set preserving if the following holds. Suppose
X hH.!2 /; 2; <P i
is a countable elementary substructure and let MX be the transitive collapse of X . Let
PX be the partial order dened by the image of <P and suppose that N is a countabletransitive model such that
(1) N ZFC,
(2) H.!2 /N D H.!2 /MX and PX 2 N .
Then
N PX is stationary set preserving:
t
u
Remark 2.50. As we have suggested, many (but not all) of the partial orders to which
one applies Martins MaximumCC .c/ are in fact absolutely stationary set preserving.
These include the partial orders for sealing antichains in .P .!1 / n INS ; /, which are
dened immediately before Denition 2.56.
t
u
Denition 2.51. Martins Maximum ZF .c/: Martins Maximum holds for partial orders
P such that:
(1) jP j c.
(2) P is absolutely stationary set preserving.
t
u
We shall prove, in Section 9.5, that Martins Maximum.c/ implies that for every
A !2 , A# exists. The following lemma is an immediate corollary of this.
Lemma 2.52 (Martins Maximum .c/). Suppose that A !2 is such that
H.!2 / LA:
Then
LA Martins MaximumZF .c/:
t
u
2.5
41
Reection Principles
Forcing axioms generalizing MA!1 to various classes of partial orders are inherently
reection principles in the spirit of supercompactness but for !2 . In the presence of
large cardinals these forcing axioms can be viewed as assertions that !2 is generically
supercompact. Suppose be a collection of partial orders. MA!1 ./ holds if for every
partial order P 2 and for every set X of dense subsets of P if X has cardinality !1
then there exists a lter F P which is X -generic.
Theorem 2.53. Assume there is a proper class of Woodin cardinals. Let be a collection of partial orders. Then the following are equivalent.
(1) MA!1 ./.
(2) For every poset P 2 and for every
there exists a generic elementary embedding
j W V ! M V P Q
such that cp.j / D !2 and such that M M in V P Q .
Proof. We rst show that (1) implies (2). This is a straightforward consequence of
the existence of the generic elementary embeddings associated to the stationary tower.
See Theorem 2.36 and Remark 2.37. We shall use the version of Theorem 2.36 which
concerns P< .
Fix P 2 and
2 Ord. Let be a Woodin cardinal such that
< and such that
P 2 V .
Let D be the set of d P such that d is dense in P .
Let < be a limit ordinal such that P 2 V and let a be the set of
X V
such that
(1.1) !1 X ,
(1.2) jX j D !1 ,
(1.3) there is a lter F X such that F is X -generic.
We claim that a is stationary in P!2 .V /.
Fix a function
H W V<! ! V :
Let be a term such that if g P is generic then the interpretation of by g is a
function
h W !1 ! V
such that if X D h!1 then
X V ;
42
2 Preliminaries
such that:
a) X Y and jY j D !1 ;
b) Z \ P!1 .Y / is stationary in P!1 .Y /.
43
such that:
a) X Y and jY j D !1 ;
b) Z \ P!1 .Y / contains a set which is closed and unbounded in P!1 .Y /. u
t
Remark 2.55. (1) The principle WRP was introduced in .Foreman, Magidor, and
Shelah 1988/ as Strong Reection. It implies the (weaker) assertion that for any
partial order P , P is semiproper if and only if forcing with P preserves stationary
subsets of !1 , see .Foreman, Magidor, and Shelah 1988/. Interestingly, Todorcevic had previously proved that a special case of WRP implies that c @2 . The
results of .Foreman, Magidor, and Shelah 1988/ show that WRP is consistent
with CH.
(2) The principle SRP was formulated in .Todorcevic 1984/ and is based on Shelahs
proof that Martins Maximum is equivalent to SPFA. The precise formulation
given in Denition 2.54(2) is the principle of Projective Stationary Reection of
Feng and Jech .1998/. Feng and Jech proved that Projective Stationary Reection is actually equivalent to Todorcevics principle.
(3) SRP implies WRP and many of the consequences of Martins Maximum follow
from it. For example, SRP implies the nonstationary ideal on !1 is saturated and
that 2@1 @2 see .Todorcevic 1984/.
It will follow from the principal results of Chapter 3, that SRP implies that
12 D !2 and so SRP implies that c D @2 . Theorem 9.79 shows that a fairly
weak fragment of SRP sufces.
(4) Both WRP and SRP follow from SPFA.
(5) One can show that SRP is consistent with the existence of a Suslin tree on !1
and so SRP does not imply Martins Maximum.
u
t
2.6
Generic ideals
One of the main results of Chapter 3 is that if the nonstationary ideal on !1 is !2 saturated and if there is a measurable cardinal, then there is an effective failure of CH.
The force of this result is greatly amplied by the results of .Foreman, Magidor,
and Shelah 1988/ and Shelah .1987/ which show that if suitable large cardinals exist
then there is a semiproper partial order P such that in V P , the nonstationary ideal on
!1 is !2 -saturated.
44
2 Preliminaries
Combining these results yields that the effective version of the Continuum Hypothesis is as intractable a problem as the Continuum Hypothesis itself.
We review briey the results of .Foreman, Magidor, and Shelah 1988/ and .Shelah
and Woodin 1990/.
We begin with the key denition. Suppose that
A P .!1 / n INS
is nonempty. Let PA denote the following partial order. Conditions are pairs .f; c/
such that
(1) for some < !1 , f W ! A,
(2) c !1 is a countable closed subset such that for each 2 c, if 2 dom.f /
then
2 f ./
for some < , and such that c ;.
The ordering on PA is by extension. Suppose that
.f1 ; c1 / 2 PA
and that .f2 ; c2 / 2 PA . Then
.f2 ; c2 / .f1 ; c1 /
if f1 f2 and c1 D c2 \ .max.c1 / C 1/.
We note that if .f; c/ 2 PA then necessarily sup.c/ 2 c. This is because c is closed
in !1 and not conal.
One of the key theorems of .Foreman, Magidor, and Shelah 1988/ is that if
A P .!1 / n INS
is predense in .P .!1 / n INS ; / then forcing with PA preserves stationary subsets of
!1 .
It is not difcult to show that PA is proper if and only if there exists a sequence
hA W < !1 i of elements of A and a closed conal set C !1 such that for all
2 C,
2 A
for some < .
The question of when the partial order PA is semiproper is more interesting. This
isolates a fundamental combinatorial condition on the predense set A which we dene
below. This condition is implicit in .Foreman, Magidor, and Shelah 1988/.
Denition 2.56. Suppose that
A P .!1 / n INS :
Then A is semiproper if for any transitive set M such that
M P .H.!2 // M;
if
X M
is a countable elementary substructure such that A 2 X , then there exists a countable
elementary substructure
Y M
45
such that
(1) X Y ,
(2) X \ !1 D Y \ !1 ,
(3) Y \ !1 2 S for some S 2 Y \ A.
t
u
t
u
The nonstationary ideal on !1 is presaturated if for any A 2 P .!1 / n INS and for
any sequence hAi W i < !i of maximal antichains in P .!1 / n INS there exists B A
such that B INS and such that for each i < !, X 2 Ai j X \ B INS has
cardinality at most !1 .
Theorem 2.58 (ForemanMagidorShelah). Suppose that for each predense set
A P .!1 / n INS ;
A is semiproper. Then the nonstationary ideal on !1 is precipitous.
t
u
t
u
The large cardinal hypothesis of Theorem 2.59 can be reduced, this yields the following theorem.
Theorem 2.60. Suppose that is a Woodin cardinal and that
G Coll.!1 ; </
is V -generic. Suppose that
hA W < i 2 V G
is a sequence such that in V G, for each < ,
A P .!1 / n INS
and A is predense.
46
2 Preliminaries
Then there exists a < such that is strongly inaccessible in V , such that
hA W < i 2 V Gj ;
and such that in V Gj , for each < ,
A P .!1 / n INS ;
A is predense, and A is semiproper.
t
u
The conclusion of Theorem 2.60 is weaker than that of Theorem 2.59, nevertheless
it is sufcient to prove INS is presaturated in V G.
Theorem 2.61. Suppose that is a Woodin cardinal and that
G Coll.!1 ; </
is V -generic. Then in V G, INS is presaturated.
t
u
Suppose that
A P .!1 / n INS
and that A is predense and not semiproper. Let TA be the set of countable
X P .H.!2 //
such that there does not exist
Y P .H.!2 //
such that X Y , X \ !1 D Y \ !1 , and such that
Y \ !1 2 S
for some S 2 Y \ A. Since A is not semiproper, the set
TA P!1 .P .H.!2 ///
is stationary in P!1 .P .H.!2 ///.
Shelah has generalized Theorem 2.60 obtaining the following theorem. For the
statement of this theorem we require a denition. Suppose N M are transitive
models of ZFC such that
!1N D !1M :
Then M is a good extension of N if for each set A 2 N such that in N ,
A P .!1 / n INS ;
A is predense and not semiproper; the set
.TA /N
is a stationary set in M .
Theorem 2.62 (Shelah). Suppose that is a Woodin cardinal and that
P V
is a -cc partial order such that:
(1) There is a conal set S such that if 2 S then is a strongly inaccessible
cardinal such that if G P is V -generic then G \ V is V -generic for P \ V ,
and V G is a semiproper extension of V G \ V .
47
(2) There exists a closed unbounded set C such that for all 2 C , if is
strongly inaccessible and if G P is V -generic then
a) !1V D !1V G ,
b) G \ V is V -generic for P \ V ,
c) D !2 in V G \ V ,
d) V G is a good extension of V G \ V .
Suppose that G P is V -generic and that
hA W < i 2 V G
is a sequence such that in V G, for each < ,
A P .!1 / n INS
and A is predense.
Then there exists < such that is strongly inaccessible in V , such that
(1) G \ V is V -generic for P \ V ,
(2) hA W < i 2 V G \ V ,
(3) in V G \ V , for each < ,
A P .!1 / n INS ;
A is predense, and A is semiproper.
t
u
PA
be the product with countable support of all the partial orders PA such that
A P .!1 / n INS
and such that A is semiproper. Then the partial order P is semiproper.
Suppose that G P is V -generic. Then V G is a good extension of V .
Proof. Let M be a transitive set such that
M H./ M
where is a regular cardinal such that
jP .P .!1 //j < :
Suppose that
A0 P .!1 / n INS
and that A0 is predense and not semiproper. Since A0 is not semiproper, the set TA0
is stationary in P!1 .P .H.!2 ///. Therefore there exists
X0 M
48
2 Preliminaries
such that X0 2 TA0 . The key point is the following. Suppose that
X M
is a countable elementary substructure such that X0 X and such that
X \ !1 D X0 \ !1 :
Then X 2 TA0 .
By constructing an elementary chain, there exists
X M
such that
(1.1) X0 X ,
(1.2) X \ !1 D X0 \ !1 ,
(1.3) for each predense set
A P .!1 / n INS
such that A 2 X and such that A is semiproper, there exists
S 2X \A
with X \ !1 2 S .
Now suppose that g X \ P is a lter which is X -generic. By (1.3) it follows
that there is a condition p 2 P such that
p<q
for all q 2 g. This veries that P is semiproper. Suppose that
GP
is V -generic and that p 2 G.
Thus there exists an elementary substructure
Y M G
such that Y \ M D X .
Since
M H./ M
and since
X 2 TA0 ;
V
it follows that .TA0 / is a stationary set in V G. This veries that V G is a good
extension of V .
t
u
As a corollary to Theorem 2.62 and Lemma 2.63 one obtains the following theorem of Shelah. The only additional ingredients required are the iteration theorems for
semiproper forcing.
Theorem 2.64 (Shelah). Suppose is a Woodin cardinal. Then there is a semiproper
partial order P such that;
(1) P is homogeneous and -cc,
(2) V P INS is saturated.
t
u
49
50
2 Preliminaries
t
u
Chapter 3
3.1
R M <!1
52
a) f 2 M n R,
b) for all < ,
f j 2 R;
c) if D C 1 then for all g 2 R, if
f j g
then f j D g.
We let ZFC be the corresponding fragment of ZFC.
Remark 3.1. (1) The second condition is a form of !1 -DC which is stronger than
!1 -replacement .
(2) At rst glance, (2c) might seem strange in its formulation. Suppose though that
M is simply a transitive set closed under the Godel operations and that R 2 M .
Suppose that
hW!!M
is an element of R. Then there exists
f W!C1!M
such that f extends h and such that f R.
(3) Assuming ZFC, if is an ordinal of conality > !1 then V ZFC . Also,
t
u
assuming ZFC, L.P .!1 // ZFC as does the transitive set H.!2 /.
The following lemma is a standard variation of os theorem.
Lemma 3.2. Suppose M is a transitive model of ZFC and that U is an ultralter on
P .!1M / \ M . Let hN; Ei be the model obtained from the M -ultrapower,
.M !1 /M =U
where
.M !1 /M D f W !1M ! M j f 2 M :
t
u
Let S be the set of stationary subsets of !1 . The partial order .S; / is not separative. It is easily veried that
RO.S; / D RO.P .!1 /=INS /:
Denition 3.3. Suppose M is a model of ZFC .
(1) .P .!1 / n INS /M denotes the partial order .S; / computed in M .
(2) A lter G .P .!1 / n INS /M is M -generic if G \ D ; for all predense sets
D 2 M.
53
(3)
M The nonstationary ideal on !1 is !2 saturated
if in M every predense subset of .P .!1 / n INS /M contains a predense subset of
t
u
cardinality !1M in M .
Remark 3.4. (1) The nonstationary ideal is saturated has several possible formulations within ZFC and they are not in general equivalent.
(2) H.!2 / The nonstationary ideal on !1 is !2 saturated .
(3) Suppose the nonstationary ideal on !1 is !2 -saturated, M is a transitive set,
M ZFC , and P .!1 / M . Then
M The nonstationary ideal on !1 is !2 saturated:
(4) Suppose that M is a transitive model of ZFC ,
.P .!1 //M 2 M;
and that G .P .!1 / n INS /M is a lter such that G \ D ; for all dense sets
D 2 M . Then G is M -generic.
t
u
Denition 3.5. Suppose that M is a countable model of ZFC . A sequence
hM ; G ; j; W < < i
is an iteration of M if the following hold.
(1) M0 D M .
(2) j; W M ! M is a commuting family of elementary embeddings.
(3) For each C 1 < , G is M -generic for .P .!1 / n INS /M , M C1 is the M ultrapower of M by G and j ; C1 W M ! M C1 is the induced elementary
embedding.
(4) For each < if is a (nonzero) limit ordinal then M is the direct limit of
M j < and for all < , j; is the induced elementary embedding.
If is a limit ordinal then is the length of the iteration, otherwise the length of
the iteration is where C 1 D .
A model N is an iterate of M if it occurs in an iteration of M . The model M is
iterable if every iterate of M is wellfounded.
t
u
Remark 3.6. (1) In many instances a slightly weaker notion sufces. A model M
is weakly iterable if for any iterate N of M , !1N is wellfounded. For elementary
substructures of H.!2 / weak iterability is equivalent to iterability.
(2) Suppose M is a countable iterable model of ZFC. Then:
M The nonstationary ideal is precipitous :
54
55
(1) !1
M
D !1 ;
t
u
Remark 3.9. The Lemma 3.8 has an obvious interpretation for arbitrary models. We
shall for the most part only use it for wellfounded models.
t
u
For the second lemma we need to use a stronger fragment of ZFC. There are
obvious generalizations of this lemma, see Remark 3.11.
Lemma 3.10. Suppose M is a countable transitive model of
ZFC C Powerset C AC C 1 -Replacement
in which the nonstationary ideal on !1 is !2 -saturated. Suppose
hM ; G ; j; W < < i
is an iteration of M such that M \ Ord. Then M is wellfounded for all < .
Proof. Let .0 ; 0 ; 0 / be the least triple of ordinals in M such that:
(1.1) M cof. 0 / > !1 ;
(1.2) 0 < 0 ;
56
1
such that:
(2.1) M cof. 0 / > !1 ;
(2.2) 0 < 0 ;
(2.3) there exist an ordinal 2 M , an M -generic lter G Coll.!; /, and an
iteration,
W < < 0 C 1i 2 M G;
hN ; G ; j;
of V0 \ M of length 0 such that j0;0 .0 / not wellfounded.
Further since M is wellfounded the same considerations apply to M and so
.j0; .0 /; j0; . 0 /; j0; .0 // must be the triple as dened in V for M . However
the tail of the iteration
hN ; G ; j; W < < 0 C 1i
starting at is an iteration of j0; .V0 \ M / of length at most 0 and
0 C 1 j0; .0 / C 1:
Further the image of by this iteration is not wellfounded. This is a contradiction
t
u
since < j0; .0 /.
Remark 3.11. Lemma 3.10 can be easily generalized to any iteration of generic elementary embeddings. A generic elementary embedding is an elementary embedding
j WV !M VP
where M is the transitive collapse of the ultrapower,
Ult.V; E/
of V by E where E is a V -extender in V P . As usual, this ultrapower is computed
using only functions in V .
t
u
57
Lemma 3.12. Let M be a transitive set such that M ZFC and such that
P .!1 / M . Suppose the nonstationary ideal on !1 is !2 -saturated in M , X M
and that X is countable.
Let D X \ !1 and let
Y D f ./ j f 2 X :
Let NX D collapse.X /, let NY D collapse.Y /, and let j W NX ! NY be the induced
58
N D collapse.X /
t
u
59
60
61
rank.< j / <
!2 D suprank M.z # ; !1 C 1/ j z 2 R:
t
u
Theorem 3.17. Suppose that the nonstationary ideal on !1 is !2 -saturated and that
12 D !2 .
P .!1 /# exists. Then
Proof. P .!1 /# exists and so H.!2 /# exists. By Lemma 3.14, there exists a countable
elementary substructure X H.!2 / whose transitive collapse is iterable. The theorem
follows by Theorem 3.16.
t
u
There is a version of Theorem 3.16 which does not require the hypothesis that the
nonstationary ideal is saturated.
Remark 3.18. The proof that (2) follows from (4) in Theorem 3.19 plays a fundamental role in the analysis of the Pmax -extension and its generalizations. This analysis is of
course the main subject of this book.
t
u
Theorem 3.19. The following are equivalent.
(1) There exists a countable elementary substructure X H.!2 / whose transitive
collapse is iterable.
(2) For every countable X H.!2 /, the transitive collapse of X is iterable.
62
(3) For all reals x, x # exists and if C !1 is closed and unbounded, then C
contains a closed unbounded subset which is constructible from a real.
(4) If C !1 is closed and unbounded, then there exists x 2 R such that
< !1 j L x is admissible C:
Proof. This is similar to the proof of Theorem 3.16.
(1 ) 3). As in the proof of Theorem 3.16 it follows that if C !1 is closed and
unbounded, then C contains a closed unbounded subset which is constructible from a
real. It remains to show that for every z 2 R, z # exists. Since X H.!2 / we need
only show this for z 2 X . Fix z 2 X \ R. Let M be the transitive collapse of X .
We prove that every uncountable cardinal of V is a regular cardinal in Lz. From
this it follows that z # exists by Jensens Covering Lemma. In fact we prove the following claim.
Claim: Suppose N is a countable transitive model of ZFC and that N is iterable.
Suppose that t is a real in N . Then !1N is a regular cardinal in Lt .
The proof of the claim is straightforward.
Let S be the set of < !1N such that is singular in L! N t . Assume toward a
1
contradiction that S is stationary in N . Let
j W N ! N
be an iteration of N of length !1 . Thus j.!1N / D !1 . Let G be V -generic for
Coll(!; !1 ). In V G let U .P .!1 / n INS /N be N -generic with
j.S / 2 U:
Let N be the generic ultrapower of N by U and let k W N ! N be the corresponding elementary embedding. Thus !1V 2 k.j.S // and so !1V is singular in
L! N t a contradiction.
1
Thus there exists club C !1N such that C 2 N and such that for all 2 C , is
a regular cardinal in L! N t . Finally suppose that !1N is not a regular cardinal in Lt .
1
63
64
. Fix a
.x; !1N /
2 N . Thus by absoluteness,
j .!1N / D Ord \ N
65
t
u
66
We shall need the following theorem due independently to Martin and Welch.
Theorem 3.22 (Martin, Welch). Suppose that
12 D !2 and that for every x 2 R, x #
exists. Then for every x 2 R, x exists.
u
t
Theorem 3.22 can be improved, obtaining much more than for every x 2 R, x
exists. It should be the case that the hypothesis implies
12 -Determinacy but this is still
an open question.
For each t 2 R let L; t denote the smallest transitive inner model N of ZFC
containing the ordinals and t such that N is closed under and such that \ N 2 N .
Lemma 3.23. Suppose that for all x 2 R, x # exists and that u2 D !2 . Suppose x 2 R.
(1) For each n < !,
P .un / \ L; x Z:
(2) There is an elementary embedding
j W L; x ! N
such that for all X 2 Vu! \ L; x, j.X / D .X /, and such that
j j L ; x 2 L; x
for all 2 Ord.
Proof. We rst prove (1). Suppose F is a function and t 2 R. For each ordinal
dene J F; t by induction on . JC1 F; t is the closure of
J F; t [ J F; t [ J F; t \ F [ F .X / j X 2 dom.F / \ J F; t
under the Godel operations.
By the denitions, we prove (1) if we prove the following claim.
Claim: Suppose t 2 R, is an ordinal and for all n < !,
P .un / \ J ; t Z:
Suppose m < !, y 2 R, B um and that B 2 JC1 ; t . Then every set A !1
such that A 2 LB; y is constructible from a real.
The argument for this is similar to the proof of Theorem 2.3 in .Hjorth 1993/. We
sketch the argument. Since 12 D !2 (and for all x 2 R, x # exists), by Theorem 3.22,
for all x 2 R, x exists.
Suppose the claim fails. Fix t 2 R and 2 Ord for which the claim fails. Fix a
ordinal such that V ZFC and < . Let X V be an elementary substructure
such that X has cardinality !1 , !1 X and such that X \ !2 has conality !. The
latter condition is the key condition. Let M be the transitive collapse of X . Choose an
! sequence, hzk W k < !i, of reals in M such that !2M D sup k j k < ! where for
each k < !, k is the least indiscernible of Lzk above !1 . The point of course is that
!2M D uM
2 and so this sequence exists.
67
Let z 2 R code the pair .hzk W k < !i; t /. z exists and so F \ Lz; F is an
ultralter in Lz; F where F is the club lter on !1 .
Let M be the image of under the transitive collapse of X . Let .J ; t /M be
the image of J ; t under the collapsing map and let M be the image of .
Thus
.J ; t /M D JM M ; t :
The key point is that JM M ; t 2 Lz; F and that
M jJM M ; t 2 Lz; F :
The verication that M jJM M ; t 2 Lz; F follows from the fact that for all
m < !,
P .um / \ J ; t Z
together with the observation that there is a map e 2 Lz; F such that for all
B 2 Z \ X , e.BM / D M .BM /. Here BM is the image of B under the collapsing map. The latter observation is easily veried as follows. By the choice of X ,
!1 X and so M is uniquely determined by the map
W Z \ X \ P .!1 / ! V
M
where .B/ D .B/ \ uM
2 and u2 is the image of u2 under the collapsing map. The
M
map is computed from exactly as and Z are computed from the set of B !1
such that B is constructible from a real. For this one uses the sequence hzi W i < !i.
It is straightforward to verify that for B 2 Z \ X \ P .!1 /, M .B/ D j.B/ \ uM
2
where
j W Lz; F ! Lz; j.F /;
68
jx W L; x ! Mx
be the induced elementary embedding. It follows that for all X 2 Z \ L; x,
jx .X / D .X /.
For each x 2 R let Ex be the .u1 ; u! / extender derived from jx . Thus
Ex 2 L; x. Let
Nx D Ult .L; x; Ex /
and let
jx0 W L; x ! Nx
be the corresponding embedding. The ultrapower Ult .L; x; Ex / is wellfounded
since it embeds into Mx . Since Ex 2 L; x it follows that
jx0 j L ; x 2 L; x
for all 2 Ord. Further by the denition of Ex it follows that jx D jx0 when restricted
t
u
to Vu! \ L; x.
The proof of part 2 of Lemma 3.23 shows that assuming that u2 D !2 , the map
is obtained from a restricted ultrapower.
Theorem 3.24. Assume the nonstationary ideal on !1 is !2 -saturated. Then the following are equivalent.
(1) 12 D !2 .
(2) There is an inner model N of ZFC containing the ordinals and an elementary
embedding
k W N ! N
such that if G .P .!1 / n INS ; / is V -generic and if
j WV !M
is the associated generic elementary embedding then
a) kjN 2 N for all ;
b) j jN! D kjN! .
where ! D sup n j n < ! and h n j n < !i is the critical sequence of k.
Proof. (1) implies (2) by the previous lemma.
We now prove that (2) implies (1).
Fix a cardinal such that jV j D and such that cof./ > !1 . Thus V ZFC and
.N! ; kjN! / 2 V :
Let X V be a countable elementary substructure such that N! ; kjN! 2 X . We
show that X \ H.!2 / is iterable. The relevant point is that .N! ; kjN! / is naturally a structure that can be iterated and further all of its iterates are wellfounded. Let
k! D kjN! . The fact that .N! ; k! / is iterable is a standard fact. k N and
69
N contains the ordinals, therefore .N; k/ is iterable; i. e. any iteration of set length
is wellfounded. The image of .N! ; k! / under an iteration of .N; k/ of length is
simply the th iterate of .N! ; k! /.
Let MX be the transitive collapse of X . Let NX! be the image of N! under the
collapsing map and let kX! be the image of k! .
We claim that .NX! ; kX! / is iterable. This too is a standard fact. Any iterate of .NX! ; kX! / embeds into an iterate of .N! ; k! / which is wellfounded since
.N! ; k! / is iterable.
The image of .NX! ; kX! / under any iteration of MX is an iterate of .NX! ; kX! /.
This is an immediate consequence of the denitions and the hypothesis, (2), of the
lemma. Therefore the image of !2 under any iteration of MX is wellfounded and so by
Lemma 3.8, the transitive collapse of X \H.!2 / is iterable. But then by Theorem 3.16,
12 D !2 .
t
u
Combining Shelahs theorem with Theorem 3.17 yields a new upper bound for the
consistency strength of
ZFC C For every real x, x # exists C 12 D !2 :
With an additional argument the upper bound can be further rened to give the following theorem.
One corollary is that one cannot prove signicantly more than
12 -Determinacy
from the hypothesis of Theorem 3.22. It is proved in .Koellner and Woodin 2010/ that
1
3 -Determinacy implies that there exists an inner model with two Woodin cardinals.
Therefore Theorem 3.22 cannot be improved to obtain
13 -Determinacy.
Theorem 3.25. Suppose is a Mahlo cardinal and that there exists < such that:
(i) is a Woodin cardinal in L.V /;
(ii) V V .
Then there is a semiproper partial order P such that
V P ZFC C For every real x, x # exists C 12 D !2 :
If in addition is a Woodin cardinal then
V P INS is !2 -saturated:
Proof. The partial order is simply the partial order P dened by Shelah in his proof
of Theorem 3.20. We shall need a little more information from this proof which we
sketched in Section 2.4. The partial order P is obtained as an iteration of length ,
hP W < i, such that:
(1.1) hP W < i V for all < such that jV j D ;
(1.2) hP W < i is denable in V for all < such that jV j D ;
(1.3) For each < , if is strongly inaccessible then
P D [P j < :
70
71
Note that this upper bound is strictly stronger than the assumption that is a
Woodin cardinal in L.V / and V# exists. This is because (with notation as in the
statement of the theorem) is a Woodin cardinal in L.V / and V# exists.
Remark 3.26. We do not know if the hypothesis needed to obtain
1 D !2
2
can be weakened below that of Theorem 3.25. A natural question is whether the assumption that is Mahlo can be reduced to the assumption that is inaccessible.
To obtain both
1 D !2
2
and that INS is !2 -saturated, the hypothesis indicated in Theorem 3.25 is plausibly
optimal.
t
u
The next theorem has been considerably improved by G. Hjorth. One of the main
theorems of .Hjorth 1993/ shows that the hypothesis that the nonstationary ideal is
saturated in unnecessary and that the conclusion can be strengthened to include all of
the usual regularity properties. In fact Hjorths theorem is an immediate consequence
of Lemma 3.23. The point is that the Martin-Solovay tree T2 is easily seen to be in
L. Lemma 3.23 is simply a mild strengthening of Hjorths result. We include our
original proof.
Theorem 3.27. Assume that the nonstationary ideal on !1 is !2 -saturated and that
!2 D
12 :
1
Then every uncountable
3 set contains a perfect subset.
Proof. The key point is that if M is a countable transitive model of ZFC which is
iterable then 13 statements with parameters from M which are true in M are true in
V.
Suppose that A R is an uncountable
13 set. Choose X H.!2 / such that X is
countable and such that X contains the parameters for the
13 denition of A. Let M
be the transitive collapse of X . By Theorem 3.16, M is iterable.
Suppose G is M -generic for .P .!1 / n INS /M . Let N be the generic ultrapower
of M by G. Let AM be A as computed in M . M is iterable hence AM D A \ M .
Similarly let AN be A as computed in N . M is iterable and so N is iterable. Hence
AN D A \ N .
Since A is uncountable it follows that there exists an injective function
f W !1 ! A
such that f 2 X . Let fM be the image of f under the collapsing map. Thus
fM D f j where D !1M .
Let hSk W k < !i be an enumeration of .P .!1 / n INS /M . For each x 2 2! let
Gx D Sk j x.k/ D 1. By a routine construction there is a perfect set Z 2! such
that:
(1.1) Gx is M -generic for each x 2 Z;
(1.2) tx ty for all x 2 Z, y 2 Z such that x y.
72
2
73
Remark 3.29. Theorem 3.28 does not really seem optimal. For example consider the
following question.
Suppose the nonstationary ideal on !1 is !2 -saturated and that AD holds in
L.R/. Is
12 D !2 ?
The proof of Theorem 3.28 does not generalize to answer this question. One really
seems to need a deeper analysis of when the forcing iteration to make the nonstationary
12 D !2 . A typical question is the following.
ideal on !1 saturated necessarily forces
Suppose V is a core model and that is the least ordinal satisfying that is a
Woodin cardinal in L.V /. Let G be L.V /-generic for the iteration which forces
the nonstationary ideal saturated. Is
12 D !2 in L.V /G?
At this point there is not even a good understanding of how the forcing iteration for
making the nonstationary ideal saturated adds any reals. For example (for the specic
iteration indicated in Section 2.4) does the iteration add any reals which are V -generic
for proper forcing?
t
u
We shall need a slight variation on the notion of iterability.
Denition 3.30. Suppose A R and M is a countable transitive iterable model of
ZFC . The model M is A-iterable if for all iterations
j WM !N
j.A \ M / D A \ N where
j.A \ M / D [j. / j 2 M and A:
t
u
Remark 3.31. (1) This is really a strengthening of the notion of weak iterability.
Suppose A is the complete 11 set. Suppose M is a countable transitive model
of ZFC and that for any iteration j W M ! N , j.A \ M / D A \ N . Then M
is weakly iterable.
(2) In most (but not all) cases when we are considering M which are A-iterable we
shall also have that A \ M 2 M .
t
u
One easy consequence of the denition of A-iterability is the following theorem
which in some sense generalizes the theorem of AD which states that every !1 union
1
of borel sets is
2 .
We require the following denition. Suppose that A R. A set B R is
11 .A/
if it is 1 denable in the structure
hV!C1 ; A; 2i
11 .A/
74
Theorem 3.32. Suppose that A R and that there exists a countable elementary
substructure X H.!2 / such that
hX; A \ X; 2i hH.!2 /; A; 2i
and such that the transitive collapse of X is B-iterable for each set B which is 11 .A/.
Suppose that there exists a sequence hB W < !1 i of borel sets such that
A D [B j < !1 :
1
Then A is
2 .
Proof. Fix a 11 set
universal for all
1
Similarly x a
U RR
sets. For each x 2 R let
11 .A/
Ux D y 2 R j .x; y/ 2 U :
set
UA R R
1
x 2 R let
universal for all
1 .A/ sets. For each
A
Ux D y 2 R j .x; y/ 2 U A :
Fix a countable elementary substructure X H.!2 / such that
hX; A \ X; 2i hH.!2 /; A; 2i
and such that the transitive collapse of X is U A -iterable.
Thus there exists a set R of cardinality @1 such that 2 X and such that
A D [Ux j x 2 :
Let M be the transitive collapse of X and let M be the image of under the
collapsing map. Thus M D \ M .
Let be the union of all sets j.M / such that
j W M ! M
is a countable iteration of M . Thus since, by Corollary 3.13, there exists an
iteration
j W M ! M
75
76
t
u
77
78
79
80
X \ S
is comeager in S ,
(1.2) for all x 2 X , .x; f .x// 2 Z,
1
(1.3) f is
1 in the codes.
It is straightforward to show that AD implies !1 -**uniformization.
Assume !1 -**uniformization and that V D L.A; R/ where A R. Then !1 **AC holds; i. e. if
F W H.!1 / ! V n ;
81
X \ S
is comeager in S .
In fact, both assertions (1) and (2) in the statement of theorem follow simply from
the assumption that !1 -**uniformization holds in L.A; R/ and we shall prove (1) and
(2) from only this weaker assumption.
Let S be the following partial order. Conditions are triples .N; g; / such that
(2.1) N !1 ,
(2.2) < !1 ,
(2.3) g 2 S .
Suppose .N1 ; g1 ; 1 / 2 S and .N2 ; g2 ; 2 / 2 S. Then
.N2 ; g2 ; 2 / < .N1 ; g1 ; 1 /
if N1 2 LN2 , 1 < 2 , g1 g2 and g2 \ Q1 ;2 is LN1 -generic.
We will need the following consequences of !1 -**AC and **uniformization. Suppose X H.!1 / and that < !1 . Suppose that for each < !1 , X \ S; is
comeager in S; . Then there exists N !1 such that for all < !1 , if < then
g Q; j g is LN -generic X:
By **uniformization, every set of reals has the property of Baire and so for every set
N !1 and for every < !1 , if < then
g Q; j g is LN -generic
is comeager in S; .
We rst prove the following. Suppose .N; g; / 2 S and D0 S is a set which is
dense below .N; g; /. Let D be the set of p 2 S such that for some q 2 D0 ,
p q 2 G
where G is the term for the generic lter.
We claim there exist N !1 and < !1 such that < and such that
.N ; g h; / 2 D
for all h Q; which are LN -generic.
By !1 -**AC, there exists X H.!1 / and there exists a function
F W X ! P .!1 /
such that for all < !1 , if < then
(3.1) X \ S; is comeager in S; ,
(3.2) if h 2 S; \ X and if there exists M !1 such that .M; g h; / 2 D then
.F .h/; g h; / 2 D.
82
.N ; g h; / 2 D:
83
Let
E D p j .; p/ 2 S for some < !1 :
We claim that E is predense in Coll.!; <!1 /. Fix p0 2 Coll.!; <!1 /. Fix 0 < !1
such that p0 2 Q;0 . Let h0 Q;0 be LN -generic with p0 2 h0 . Let
.M1 ; g1 ; 1 / 2 D be such that
.M1 ; g1 ; 1 / < .N ; g h0 ; 0 /:
Thus g1 D g h1 where h1 Q;1 , h0 h1 and h1 is LN -generic. Therefore
.N ; g h1 ; 1 / 2 D
and further h1 2 Y1 . Therefore there exists p1 2 h1 such that .1 ; p1 / 2 S . This
proves that E is predense. Therefore there exists < !1 such that
p 2 Coll.!; </ j . ; p/ 2 S for some <
is predense in Coll.!; </.
Finally suppose h Q; is LN -generic. Then there exists p 2 h and there
exists < such that . ; p/ 2 S . Therefore
.N ; g h ; / 2 D
where h D h \ Q; , and so
.N ; g h; / 2 D:
This proves our claim about .N; g; / and D.
We prove (1). Fix G. Let
jG W L.A; R/ ! L.AG ; RG /
be the elementary embedding given by (3).
For each < !1 let
G D G \ Q
and let
j W L.A; R/ ! L.AG ; RG /
be the associated embedding. Similarly for each < < !1 let
G; D G \ Q;
and let
j; W L.A; R/ ! L.AG; ; RG; /
be the associated embedding.
Let
FG D .N; g; / 2 jG .S/ j g G and G \ Q;!1 is LN -generic:
We claim that FG is L.AG ; RG /-generic.
Suppose that
p0 2 S
and that p0 D .N0 ; g0 ; 0 /. Let FGp0 be the set of .N; g; / 2 jG .S/ such that
(6.1) .N; g; / < p0 ; g Gp0 ,
(6.2) Gp0 \ Q;!1 is LN -generic.
84
where
Gp0 D g0 .G \ Q0 ;!1 /
is the perturbation of G to extend g0 .
We prove that FGp0 \ jG .D/ ; for all D S such that D is dense.
Suppose D S is dense. We may assume that for all p 2 S, if
p q 2 G
for some q 2 D then p 2 D.
Since every set of reals has the property of Baire and since !1 -**AC holds, it
follows by the remarks above, that there exists < !1 such that 0 < and there
exists N1 !1 such that .N1 ; g0 h; / 2 D for all h Q0 ; such that h is
LN1 -generic.
Thus .N1 ; g0 G0 ; ; / 2 jG .D/ and .N1 ; g0 G0 ; ; / < jG .p0 /. Clearly
G \ Q;!1 is LN1 -generic. Thus
.N1 ; g0 G0 ; ; / 2 jG .D/ \ FGp0 :
This proves that FGp0 \ jG .D/ ; for all D S such that D is dense.
It now follows that FG \ D ; for all D 2 L.AG ; RG / such that D jG .S/
and such that D is dense. The point is that any set in L.AG ; RG / is the image of a
set in L.AG ; RG / for some and so genericity follows by relativizing the previous
argument, with a suitable choice of p0 , to
L.AG ; RG / V G :
Finally we prove (2). It sufces to show that if
F S
is L.A; R/-generic then !1 -AC holds in L.A; R/F . We work in L.A; R/.
Suppose is a term, .N; g; / 2 S and
.N; g; / ;:
We prove that there exists N
85
Let N !1 be a set such that N 2 LN and such that if h Q; is LN generic then h 2 X . F denes a term and
.N ; g; / 2 :
Now suppose 2 L.A; R/S is a term, .N; g; / 2 S, and
.N; g; / W !1 ! V n ;:
Let h W < < !1 i be a sequence of terms such that for all < !1 ,
.N; g; / ./ D :
where C 1 C D .
By !1 -**AC and by **uniformization and by the result proved above, there exists
X H.!1 / and two functions,
F0 W X ! P .!1 /
and
F1 W X ! L.A; R/
with the following properties. For all < !1 , if < then X \ S; is comeager in
S; and for all h 2 X \ S; , F1 .h/ is a term and
.F0 .h/; g h; / F1 .h/ 2 :
As we did above we extract the term dened by F0 . Let T be the set of triples
.; q; / such that
(7.1) < < !1 ,
(7.2) q 2 Q; ,
(7.3) < !1 ,
(7.4) h 2 S; j 2 F0 .h/ is comeager in the open subset of S; given by q.
For each < !1 such that < let Y be the set of h 2 X \ S; such that
F0 .h/ D < !1 j .; q; / 2 T for some q 2 h:
Thus Y is comeager in S; . Let
Y D [Y j < < !1 :
Finally let N !1 be such that .N; T / 2 LN and such that for all < !1 , if
< then
h 2 S; j h is LN -generic X \ Y:
Suppose h Q; and that h is LN -generic. Therefore h 2 X and so
.F0 .h/; g h; / F1 .h/ 2 :
The genericity of h relative to LN also implies that h 2 Y . Therefore
F0 .h/ 2 LT h:
86
However T 2 LN and so
.N ; g h; / .F0 .h/; g h; / 2 G
where G is the term for the generic lter.
Thus for all < !1 , if < and if h 2 S; is LN -generic then
.N ; g h; / F1 .h/ 2 :
The function F1 yields a term 2 L.A; R/S such that
.N ; g; / is a choice function for :
t
u
Remark 3.37. (1) As indicated in the proof, one does not need AD for this. For
example, (3) follows assuming only that **uniformization holds in L.A; R/.
See .Woodin 1983/.
(2) The partial order S, dened in the proof of Theorem 3.36, is equivalent to the
forcing notion of .Steel and VanWesep 1982/. Assuming AD,
.N; g; / 2 S j N !
is dense in S and the order on S can be rened to make the partial order !-closed;
i. e. S is !-strategically closed.
(3) With additional requirements on the inner model, L.A; R/, one in fact gets !1 DC in L.AG ; RG /G. The additional assumption is ADC , it is implied by AD
if V D L.R/. !1 -choice is sufcient for our purposes. A brief survey of ADC
is given in the rst section of Chapter 9.
t
u
As a corollary to Theorem 3.36 we obtain the following theorem.
Theorem 3.38. Suppose A R,
L.A; R/ AD
and that G Coll.!; <!1 / is V -generic. Let RG D RV G and let
AG D [X V G j X 2 V; X is borel and X A
where if X 2 V is a borel set then X V G denotes its interpretation in V G. Suppose
is an ordinal. Then:
(1) Suppose f W !1 ! is a function in L.AG ; RG /. Then there is a function
g 2 L.A; R/ such that f and g agree on a club in !1 .
(2) Suppose X is a set of size !1 in L.A; R/G. Then there is a set Y in
L.A; R/ of size !1 such that X Y .
(3) Suppose f W !1 ! is a function in L.AG ; RG /G. Then there exists a
sequence
hS W < !1 i 2 L.AG ; RG /G
87
5S j < !1
contains a club and such that for each < !1 , there is a club C on which
f j.S \ C / D g
for some g 2 L.AG ; RG /.
(4) Suppose X is a set of size !1 in L.AG ; RG /G. Then there is a set Y
in L.AG ; RG / of size !1 such that X Y .
Proof. (1) follows immediately from Theorem 3.36(3) noting that the club lter on !1
is an ultralter in L.A; R/ and that the ultrapower
Ord!1 =U
is wellfounded in L.A; R/ where U is the club lter on !1 .
(2) is an elementary consequence of the fact that the partial order,
Coll.!; <!1 /
is ccc and of cardinality !1 .
(4) follows from (2) and the elementary embedding given by Theorem 3.36(3).
We prove (3). For each x 2 R, let
kx W Lx ! Lx
be the canonical elementary embedding where kx has critical point !1V and
kx .!1V / D .12 /V D .!2 /L.A;R/ :
These conditions together with the condition
88
Let
T D .p; ; / j p 2 Coll.!; <!1 /; p ./ D :
Since AD holds in L.A; R/, there exists x 2 R such that T 2 Lx. Let
TQ D kx .T /:
Thus TQ is a term for a function
fQ W
12 ! 12
in the forcing language for Coll.!; <12 /.
Let A Coll.!; <12 / be a maximal antichain such that A 2 Lx and such that
for all p 2 A there exists <
12 such that
.p; !1 ; / 2 TQ :
Thus A has cardinality at most !1 in L.AG ; RG /G. Let
D sup j .p; !1 ; / 2 A:
Thus
<
12 D .!2 /L.A;R/ :
Let y0 2 R be such that x 2 Ly0 and such that
< ..!1V /C /Ly0 :
Thus T 2 Ly0 G and so f 2 Ly0 G.
We work in Ly0 G. Let
A D p 2 A j pj.!1V !/ 2 G :
Let hp W < !1V i be an enumeration of A . For each < !1V let be such that
.p ; !1V ; / 2 TQ :
Let hg W < !1V i 2 Ly0 be a sequence of functions such that
g W !1V ! !1V
and such that
kQy0 .g /.!1V / D :
89
E \ S ;:
Therefore in L.AG ; RG /G, S contains a closed unbounded subset of !1 .
t
u
90
B D f 1 CG
Uz D y 2 R j .z; y/ 2 U :
If M is a transitive model of ZFC we let U M D U \ M . By absoluteness U M is
dened in M by the same 11 formula which denes U in V .
Suppose
X hH.!2 /; C; 2i
is a countable elementary substructure in V such that MX is D-iterable and MX is
E-iterable where
D D z j Uz C
and
E D z j Uz R n C :
Let
Y D X G
and let N be the transitive collapse of Y . Therefore
N D MX G \ Coll.!; <!1X /
where !1X D X \ !1 .
Since **AC holds in L.A; R/ (see the proof of Theorem 3.36) it follows that
Y hH.!2 /V G ; CG ; 2i:
Suppose in V G,
91
k W N ! N
N D k.MX /k.g/
92
Similarly
[Uzk.N / j z 2 EG \ k.MX / RG n CG :
This implies
k.N \ CG / CG
and
k.N \ .RG n CG // RG n CG :
Therefore
k.N \ CG / D k.N / \ CG
and
k.N \ .RG n CG // D k.N / \ .RG n CG /:
This proves that N is CG -iterable in V G.
Finally suppose
Z P!1 .H.!2 //
is stationary in P!1 .H.!2 //. Then
X G j X 2 Z
V G
.
is stationary in .P!1 .H.!2 ///
Therefore in V G, the set
t
u
To prove the rst of the two covering theorems we need the following theorem of
Steel which is a corollary of the results of .Steel 1981/.
Theorem 3.40 (Steel, (AD + DC)). Suppose that < and cof./ > !. Suppose
that Y R R is a prewellordering of length .
Then there exists a set X R and a surjection
WX !
such that
1
(1) for each
1 set Z X ,
sup.t / j t 2 Z < ;
Proof. Suppose that there exists .; X / which satises (1). Then by the Moschovakis
Coding Lemma there exists .; X / satisfying both (1) and (2).
We prove (1). We x some notation. Suppose A R. Let A be the least ordinal
such that
LA .A; R/ ZF n Powerset
and let
A D P .R/ \ LA .A; R/:
93
A ;
(1.2) ordertype A j A 2 D .
These conditions uniquely specify . Let
D sup A j A 2 :
By (1.2), cof. / D cof./ and so cof. / > !.
We shall assume the basic facts concerning Wadge reducibility in the context of
AD, see the discussion after Denition 9.25. One such fact is that there exists a set
B ! ! of minimum Wadge rank such that B . Fix B0 and let
D B ! ! j B is a continuous preimage of B0 :
Let
O D ! ! n A j A 2
94
1
then
sup.x0 ; x1 / j .x0 ; x1 / 2 Z < :
This is where we use (3.2). Since the range of has ordertype , this boundedness
property will sufce to prove (1).
Let
Z D .x0 ; x1 ; y; z0 ; z1 / j .x0 ; x1 / 2 Z; .x0 /.y/ D z0 ; and .x1 /.y/ D z1 :
1
Thus Z is
1 . Let
95
But
! ! n A D .x0 ; x1 ; y; z0 ; z1 / j .x0 ; x1 ; y; z0 ; z1 / Z or z0 A0
D .x0 ; x1 ; y; z0 ; z1 / j .x0 ; x1 ; y; z0 ; z1 / Z or z1 2 A0
D .! ! n Z / [ .x0 ; x1 ; y; z0 ; z1 / j z1 2 A0
and so since is closed under nite unions (and contains all
11 sets),
! ! n A 2 :
Therefore
A 2 \ O D :
1
A0 . Therefore
and so Z is bounded.
We begin with a technical lemma.
Lemma 3.41. Suppose that M is a transitive inner model such that
M ZF C DC C AD;
and such that
(i) R M ,
(ii) Ord M ,
(iii) for all A 2 M \ P .R/, the set
X hH.!2 /; 2i j MX is A-iterable and X is countable
is stationary where MX is the transitive collapse of X .
Suppose < M , S !1 is stationary and f W S ! . Suppose that
g W !1 !
is a function such that g 2 M and such that
f ./ g./
for all 2 S .
Then there exists a sequence h.T ; g / W < !1 i such that
S D 5T j < !1
and such that for all < !1 ,
(1) T is stationary,
(2) g 2 M ,
96
(3) either
f jT D gjT ;
or for all 2 T ,
f ./ g ./ < g./:
f ./ gi ./ < g./:
S1 D 2 S j cof.g.// D !;
S2 D 2 S j cof.g.// > !:
97
1
By Theorem 3.40, dom.P / is exactly the set of x 2 R such that x codes a countable
ordinal.
The key point is that !1 -choice holds in L.AG ; RG /G and so we can nd a sequence
h.x ; y / W < !1 i 2 L.AG ; RG /G
of elements of P such that for each < !1 , x codes a countable ordinal such that
h. / D g./.
Choose in V G an !1 sequence of reals hz W < !1 i such that for each < !1 ,
z 2 eld.y / and such that for each 2 S2 , f ./ is the rank of z relative to y .
Let S D h.x ; y / W < !1 i and let T D hz W < !1 i.
Choose a countable elementary substructure
X H.!2 /V G
containing the sequences S and T and such that MX is P -iterable where MX is the
transitive collapse of X . Let SX and TX be the images of S and T under the collapsing
98
map. Thus SX D Sj!1MX and similarly for TX . By Corollary 3.13, there is an iteration
j W MX ! N of length !1 such that
j.SX / D S and j.TX / D T :
Fix 2 !1 . Let Z be the set of all z 2 RG such that there is an iteration
k W MX ! N
of length C 1 such that k.SX /j. C 1/ D Sj. C 1/ and z D k.TX /./. Thus Z
1
is a
1 set and z 2 Z . Further since MX is P -iterable we have Z eld.y /.
Thus this set is bounded. The denition of Z is uniform in Sj. C 1/ and hence
hZ W < !1 i 2 L.AG ; RG /:
Therefore there is a function
g 2 L.AG ; RG /G
such that for all 2 S2 , f ./ g ./ < g./. This proves that the claim holds for
t
u
the triple .f; g; S2 /.
Lemma 3.41 yields the following theorem as an easy corollary.
Theorem 3.42. Suppose that M is a transitive inner model such that
M ZF C DC C AD;
and such that
(i) R M ,
(ii) Ord M ,
(iii) for all A 2 M \ P .R/, the set
X hH.!2 /; 2i j MX is A-iterable and X is countable
is stationary where MX is the transitive collapse of X .
Suppose < M and that f W !1 ! .
Then there exists a sequence h.S ; g / W < !1 i such that
!1 D 5S j < !1
and such that for all < !1 ,
(1) S is stationary,
(2) g 2 M ,
(3) f jS D g jS .
Proof. By Lemma 3.41 there exists a sequence
hFi W i < !i
of functions such that for each i < !:
99
Fi .S / W !1 ! ;
Fi .S /jS D f jS;
or for all 2 S ,
f ./ < Fi .S /./I
(1.5) for each T 2 dom.FiC1 / there exists S 2 dom.Fi / such that T S ;
(1.6) suppose that S 2 dom.Fi /, T 2 dom.FiC1 / and that T S , then for each
2 T,
FiC1 .T /./ < Fi .S /./I
(1.7) for each S 2 dom.Fi / if
Fi .S /jS f jS
100
and so jAj !1 .
We obtain as an immediate corollary the rst covering theorem.
!3 sup M
where M ranges over transitive inner models such that
(i) R M ,
(ii) Ord M ,
(iii) M ZF C DC C AD,
(iv) every set A 2 M \ P .R/ is weakly homogeneously Suslin in V .
Suppose G P .!1 / n INS is V -generic and that
j WV !M
is the induced generic elementary embedding.
Then j j 2 V for every ordinal .
Proof. By the last theorem j j!3 2 V . It follows on general grounds that j jOrd is a
denable class in V .
u
t
101
AG D [ X V G j X 2 V; X is borel and X A
where if X 2 V is a borel set then X V G denotes its interpretation in V G.
By Theorem 3.38, it sufces to nd Y such that:
(1.1) Y 2 L.AG ; RG /G;
(1.2) X Y ;
(1.3) Y has cardinality !1 in L.AG ; RG /G.
By Lemma 3.39 we may apply the rst covering theorem in V G at to obtain
functions which are in L.AG ; RG /.
Let D L.A;R/ D L.AG ;RG / . Thus L .AG ; RG / ZFnPowerset. Fix <
such that:
(2.1) L .AG ; RG / L .AG ; RG /;
(2.2) has conality !2 in L.AG ; RG /;
(2.3) < .
102
103
f .Z \ !1 / D g .Z \ !1 /
t
u
104
There is another formulation of Theorem 3.45. Recall P!1 .X / denotes the set of
all countable subsets of X .
Theorem 3.46. Suppose that M is a transitive inner model such that
M ZF C DC C AD;
and such that
(i) R M ,
(ii) Ord M ,
(iii) for all A 2 M \ P .R/, the set
X hH.!2 /; 2i j MX is A-iterable and X is countable
is stationary where MX is the transitive collapse of X .
Suppose < M and that
f W !1 ! :
Then there exists a function
g W !1 ! P!1 ./
such that g 2 M and such that for all < !1 , f ./ 2 g./.
Proof. Let
X D f ./ j < !1 :
By Theorem 3.45, there exists a set Y such that X Y and such that
jY jM D !1 :
By Theorem 3.19,
.!2 /M D !2
and so we may reduce to the case that D !1 .
Let
C D < !1 j f :
The set C is closed and unbounded in !1 . By Theorem 3.19, there exists a closed,
conal, set D C such that for some x 2 R,
D 2 Lx:
Therefore D 2 M . Dene
g W !1 ! P!1 .!1 /
by
g./ D min.D n /;
where D C 1. Thus g is as required.
t
u
105
106
107
Therefore
jO.A \ MX / D A \ MX :
t
u
108
3.2
We still do not know if CH implies that the nonstationary ideal on !1 is not saturated.
In light of the results in the previous section this seems likely.
Shelah, Shelah .1986/, has proved that assuming CH the nonstationary ideal is not
!1 -dense. We prove a generalization of this theorem.
It is a standard fact, which is easily veried, that the boolean algebra, P .!1 /=INS ,
is !2 -complete; i. e. if
X P .!1 /=INS
is a subset of cardinality at most @1 then _X exists in P .!1 /=INS .
Theorem 3.50. Suppose that the quotient algebra
P .!1 /=INS
is !1 -generated .equivalently !-generated/ as an !2 -complete boolean algebra. Then
2@0 D 2@1 :
t
u
b j !1 < !2 :
bA
Suppose
Z P .!1 /=INS
is of cardinality @1 . Then there exists a set A !1 such that
Z BA :
Thus Theorem 3.50 is an immediate corollary of the next theorm.
Theorem 3.51. Suppose that for some set A !1
BA D P .!1 /=INS
Then
2@0 D 2@1 :
109
2 N j G0 \ .ba /N ; :
Finally
110
111
112
is unbounded in !1 .
t
u
113
F <! :
Let 2 M be a bijection from !1 to . For each < !1 ,
2 M \ P!1 .!2 /
and
< !1 j is closed under F
contains a club in !1 . The existence of follows.
Therefore M \ P!1 .!2 / is stationary. This implies that the set
2 P!1 .!2 / j hS W < !1 i guesses A j 2
is stationary in P!1 .!2 / since it contains P!1 .!2 / \ M .
It follows
X j X P .!1 / is countable and hS j < !1 i guesses X
is stationary in !2 where X D sup j A 2 X .
Q
Therefore hS W < !1 i and hA W < !2 i together witness .
t
u
114
Fix satisfying (1.3). Let hX W < !1 i be the elementary chain where X0 D X
and for all < !1 ,
(2.1) XC1 D f .X \ !1 / j f 2 X ,
(2.2) if is a limit ordinal,
X D [X j < :
Fix < !1 such that
!1 \ X < !1 \ XC1 :
Note that for all < !1 , X \ !2 is conal in X \ !2 . Therefore
j 2 X and A \ 2 S
is conal in X \ !2 . Thus by replacing X by X if necessary we may assume that
D 0; i. e. that
!1 \ X < !1 \ Y
where
Y D f .X \ !1 / j f 2 X :
Let NX be the transitive collapse of X , let NY be the transitive collapse of Y and
let
j W NX ! NY
be the induced elementary embedding (the image of the inclusion map). However
the nonstationary ideal on !1 is !2 -saturated and P .!1 / H.!3 /. Therefore by
Lemma 3.12, NY is a generic ultrapower of NX and j is the induced embedding.
Transferring to V (or equivalently, working in NX ) there exists a stationary set
S !1 and ordinal 0 such that !1 0 < !2 and such that if
G P .!1 / n INS
is V -generic with S 2 G then
< !2 j j.A / \ 0 2 S0
is conal in !2 where
j W V ! N V G
t
u
As an immediate corollary to Theorem 3.58 and Theorem 3.59 we obtain the following.
Corollary 3.60. Assume that the nonstationary ideal on !1 is !2 -saturated. Then
u
t
fails in any transitive inner model which correctly computes !2 .
115
t
u
Remark 3.61. (1) For this question the fragment of ZFC is important. The answer
should be the same for all reasonably strong fragments. But note the answer is
yes for ZFC for trivial reasons.
(2) It is straightforward to show that the answer is no if the model M is iterable or
if M P .!1 /=INS is countably generated.
(3) Suppose the nonstationary ideal on !1 is saturated and CH holds. Suppose there
exists an inaccessible cardinal. Then the answer is yes.
t
u
One could ask this question for any iteration of generic embeddings.
Suppose V is the inner model for one Woodin cardinal. Suppose
G Coll.!1 ; </
is V -generic where is the Woodin cardinal. Then in V G there are saturated ideals
on !1 . Suppose < and that is inaccessible. Let X V G be a countable
elementary substructure of V G and let M be the transitive collapse of X . Thus
M 2 V . Suppose j W M ! N is an elementary embedding with N transitive and
!1 D !1N . Then it follows that R N and so M 2 N . Thus if there is any wellfounded iteration of M of length !1 then the answer to the more general form of the
question is yes.
An even more general class of iterations is obtained by mixing generic ultrapowers
with iteration trees. For this notion of iteration it is possible for a model to be an
element of an iterate of itself. We state without proof a theorem which illustrates the
possibilities.
Theorem 3.62. Suppose there are two Woodin cardinals with an inaccessible above
them both. Then there is a sequence hM0 ; M1 ; M2 i of countable transitive models of
ZFC such that:
(1) M1 is an iterate of M0 by an iteration tree on M0 ;
(2) M2 is a generic ultrapower of M1 . for the stationary tower/;
(3) M0 2 M2 .
t
u
Chapter 4
The results of Chapter 3 suggest that under suitable large cardinal hypotheses, if the
nonstationary ideal on !1 is !2 -saturated then the inner model L.P .!1 // may be close
to the inner model L.R/. Perhaps the most important clue is given by Corollary 3.13;
if the nonstationary ideal on !1 is saturated and there is a measurable cardinal, then
every subset of !1 appears in an iterate of a countable iterable model. Motivated by
these considerations we shall dene and analyze in Section 4.2 a partial order
Pmax 2 L.R/
for which the corresponding generic extension,
L.R/G;
is an optimal version of L.P .!1 // (assuming ADL.R/ ).
First we generalize the notion of iterability slightly to accommodate the denition.
4.1
Iterable structures
117
(2) If is a limit ordinal then is the length of the iteration, otherwise the length of
the iteration is where C 1 D .
(3) A pair .N ; J/ is an iterate of .M; I/ if it occurs in an iteration of .M; I/.
(4) .M; I/ is iterable if every iterate is wellfounded.
t
u
Remark 4.2. (1) This is the natural denition for iterability relative to a set of ideals. We shall only use it in the case that the set of ideals is nite.
(2) Suppose that M is a countable transitive model of ZFC such that
.P .!1 //M 2 M:
Then .M; .INS /M / is iterable if and only if M is iterable in the sense of Denition 3.5.
(3) We will often write .M; I / when referring to .M; I / in the case where only
one ideal is designated.
t
u
We dene the corresponding notion of X -iterability where X R.
Denition 4.3. Suppose M is a countable transitive model of ZFC . Suppose I 2 M
is a set of uniform normal ideals on !1M . Suppose .M; I/ is iterable, X R and that
X \ M 2 M. Then .M; I/ is X -iterable if for any iteration of .M; I/,
j W .M; I/ ! .M ; I /
j.X \ M/ D X \ M .
t
u
The next two lemmas are the generalizations of Lemma 3.8 and Lemma 3.10 respectively. The proofs are similar and we omit them.
Lemma 4.4. Suppose that M and M are countable models of ZFC such that
118
(1) !1
M
D !1 ;
t
u
t
u
119
t
u
120
(4) For each < if is a nonzero limit ordinal then for every k < !, Nk is
the direct limit of Nk j < and for all < , j; is the induced 0
elementary embedding.
If is a limit ordinal then is the length of the iteration, otherwise the length of
the iteration is where C 1 D .
A sequence h.Nk ; Jk / W k < !i is an iterate of h.Nk ; Jk / W k < !i if it occurs in
an iteration of h.Nk ; Jk / W k < !i.
t
u
The sequence h.Nk ; Jk / W k < !i is iterable if every iterate is wellfounded.
Condition (iii) in Denition 4.8 guarantees that nontrivial iterations exist.
Lemma 4.9. Suppose h.Nk ; Jk / W k < !i is an iterable sequence. Suppose
I 2 J0
Then there exist
G [.P .!1 //Nk j k < !
and a sequence hIk W k < !i such that I0 D I and such that for all k < !, G \ Nk is
Nk -generic for
.P .!1 / n Ik /Nk :
Proof. By condition (iii) in Denition 4.8 there exists a sequence hIk W k < !i such
that I D I0 and such that for all k < !,
(1.1) Ik 2 Jk ,
(1.2) IkC1 \ Nk D Ik ,
N
121
t
u
We now prove a lemma which we shall use to show that condition (iii) of Denition 4.8 is satised by the ! sequences of structures that we shall be interested in.
Ultimately we shall apply the lemma within models of only ZFC and so we prove the
lemma assuming only ZFC .
Suppose that J is a normal uniform ideal on !1 and that A P .!1 / n J has
cardinality at most !1 . Suppose that
hA W < !1 i
and
hA W < !1 i
are each enumerations of A possibly with repetition. Then the diagonal unions
5A j < !1 and
5A j < !1
are equal on a club in !1 and so they are equal modulo J . Thus modulo J the diagonal
union, 5A is unambiguously dened. The same considerations apply to diagonal
intersections. We let 4A denote the diagonal intersection of A.
Lemma 4.10 (ZFC ). Suppose M0 is a countable transitive model, I0 2 M0 is a set
of normal uniform ideals on !1M0 , and M0 ZFC . Suppose that
(i) for all I0 ; I1 2 I0 , if
I0 b 2 .P .!1 //M0 j b \ a 2 I1
for some a 2 .P .!1 //M0 such that
!1M0 n a I1 ;
then I0 D I1 .
Suppose J is a normal uniform ideal on !1 and that
j W .M0 ; I0 / ! .M0 ; I0 /
is a wellfounded iteration of length !1 such that J \ M0 D J for some J 2 I0 . Let
X be the set of A 2 M0 such that A P .!1 /M0 n J and A is a maximal antichain.
Let
A D 45A j A 2 X:
122
Then
(1) !1 n A 2 J ,
t
u
Remark 4.11. (1) The set A in Lemma 4.10 is analogous to a master condition.
As a condition in P .!1 / n J it forces that the generic lter is M0 -generic for
P .!1 /M0 =J .
(2) The requirement (i) in the statement of Lemma 4.10 can be weakened though
some assumption is necessary.
t
u
Lemma 4.12 is a version of Lemma 4.10 where the assumption
J \ M0 2 I0
is dropped and where no additional assumptions are made about I0 .
123
A D 45A j A 2 X:
and X is the set of A 2 M0 such that A P .!1 /M0 n J0 and such that A is a
maximal antichain.
Proof. Let
h.M ; I /; G ; j; W < !1 i
be the iteration such that j D j0;!1 .
For each < !1 , there is an ideal I 2 I such that
G M n I
and such that G is M -generic. The ideal I is not necessarily unique however distinct
candidates differ in a trivial manner.
For each < !1 let IO be the set of I 2 I such that
(1.1) G M n I ,
(1.2) G is M -generic for
.P .!1 / n I; /M
:
Since J is normal it follows that there exist S !1 , 0 < !1 , and I 2 IO0 such
that
(2.1) S J ,
(2.2) S .0 ; !1 /,
(2.3) for all 2 S ,
j0 ; .I / 2 IO :
t
u
124
t
u
and
!1 k D !1 kC1 :
125
Remark 4.16. Denition 4.15 is really just a slight generalization of Denition 3.5.
Suppose hNk W k < !i is an iterable sequence such that for all k < !,
jNk jNkC1 D .!1 /N0 :
Let
N D [Nk j k < !:
In general, N is not a model of ZFC , however N is a model of a fragment of ZFC
which is rich enough to make it possible to apply Denition 3.5. Iterations of the
sequence hNk W k < !i correspond to iterations of N .
In virtually every situation in which we consider iterations of hNk W k < !i it will
be the case that for all k < !,
jNk jNkC1 D .!1 /N0 :
t
u
j.!1N0 / D !1 0
126
t
u
127
However a 2 NiC1
aiC1 .
Let hai W i < !i be a sequence satisfying (1.1)(1.4) and let
128
t
u
Remark 4.21. The previous lemma is also true if condition (iv) is replaced by the
condition that for all
x 2 R \ .[Nk j k 2 !/;
x # 2 [Nk j k 2 !.
t
u
rank.Nk / <
12 ;
129
(1) M0 D M .
(2) j; W M ! M is a commuting family of elementary embeddings.
(3) For each C1 < , G is an M -normal ultralter, M C1 is the M -ultrapower
of M by G and j; C1 W M ! M C1 is the induced elementary embedding.
(4) For each < if is a limit ordinal then M is the direct limit of M j <
and for all < , j; is the induced elementary embedding.
A model N is a semi-iterate of M if it occurs in an semi-iteration of M . The model
M is strongly iterable if every semi-iterate of M is wellfounded.
t
u
Clearly if
M INS is saturated
then every semi-iteration of M is an iteration of M .
We recall the following notation. Suppose A R. Then
11 .A/ is the set of all
BR
such that B can be dened from real parameters by a 1 formula in the structure
hV!C1 ; A; 2i:
if both B and R n B are
11 .A/.
A set B R is
Let
11 .A/ be the supremum of the lengths of the prewellorderings of R that are
1
1 .A/.
1
1 .A/
Lemma 4.24. Suppose that A R and that there exists X H.!2 / such that
hX; A \ X; 2i hH.!2 /; A; 2i
and such that the transitive collapse of X is A-iterable.
Suppose that M is a transitive set, H.!2 / M ,
M ZFC ;
and that
M \ Ord <
11 .A/:
Then the set of
Y M j Y is countable and MY is strongly iterable
contains a club in P!1 .M /. Here MY is the transitive collapse of Y .
Proof. Let D rank.M / and let
WR!
be a surjection such that
1
.x; y/ j .x/ .y/ 2
1 .A/:
130
Let
B D .x; y/ j .x/ .y/:
Let N be a transitive set such that
N ZFC
and such that M; ; A [ H.!2 / N .
Let Y N be a countable elementary substructure such that ; M; A Y and
let NY be the transitive collapse of Y . Let Y be the image of under the transitive
collapse and let Y be the image of .
Let X D Y \ M and let MX be the transitive collapse of X .
Suppose
j W .MX ; 2/ ! .M ; E /
is an elementary embedding given by a countable semi-iteration.
Since
H.!2 /MX D H.!2 /NY ;
j lifts to dene a semi-iteration
k W .NY ; 2/ ! .N ; E /:
We identify the standard part of N with its transitive collapse.
Thus
kjH.!2 /MX W H.!2 /MX ! k.H.!2 /MX /
is a countable iteration.
By Theorem 3.34, H.!2 /MX is A-iterable. Therefore
k.A \ NY / D A \ N
1
and so since B is
1 .A/ in parameters from NY ,
k.B \ NY / D B \ N :
t
u
Denition 4.25. The nonstationary ideal on !1 is semi-saturated if for all generic extensions, V G, of V , if U 2 V G is a V -normal ultralter on !1V , then Ult.V; U / is
wellfounded.
t
u
131
132
W !1 !
133
t
u
134
We will encounter situations in which the nonstationary ideal on !1 is semisaturated and not saturated cf. Denition 6.11 and Theorem 6.13. Nevertheless the
assertion that INS is semi-saturated has many of the consequences proved in Section 3.1
for the assertion that INS is saturated.
For example it is routine to modify the proofs in Section 3.1 to obtain the following
variations of Lemma 3.14 and Theorem 3.17, together with the subsequent generalization of Theorem 3.47.
Clearly, if the nonstationary ideal is semi-saturated in V then it is semi-saturated in
L.P .!1 //.
Theorem 4.29. Suppose that the nonstationary ideal on !1 is semi-saturated and that
P .!1 /# exists. Suppose that
X H.!2 /
is a countable elementary substructure. Then the transitive collapse of X is iterable.
Proof. Clearly for all x 2 R, x # exists.
Let
Y L.P .!1 //
be a countable elementary substructure containing innitely many Silver indiscernibles
of L.P .!1 //.
Let X D Y \ H.!2 /, let N be the transitive collapse of Y and let M be the
transitive collapse of X .
Thus
M D .H.!2 //N
and
N D L .M /
where D N \ Ord.
Since Y contains innitely many indiscernibles of L.P .!1 //,
L .M / L.M /:
Finally INS is semi-saturated and so
L.P .!1 // INS is semi-saturated:
Therefore
N INS is semi-saturated
and so
L.M / INS is semi-saturated:
We claim that M is iterable. Suppose M is an iterate of M occurring in an
iteration of length .
Let < !1 be such that < and such that
L .M / L.M /:
135
By an absoluteness argument analogous to the proof of Lemma 3.10, any semiiterate of L .M / occurring in a semi-iteration of L .M / of length less than is wellfounded.
The iteration of M of length witnessing M is an iterate of M induces a semiiteration of L .M / of length producing a semi-iterate of L .M / into which M can
be embedded. Therefore M is wellfounded and so M is iterable.
Thus there exists a countable elementary substructure
X H.!2 /
whose transitive collapse is iterable. Thus by Theorem 3.19, if
X H.!2 /
is any countable elementary substructure, the transitive collapse of X is iterable.
t
u
Theorem 4.30. Suppose that the nonstationary ideal on !1 is semi-saturated and that
12 D !2 .
P .!1 /# exists. Then
Proof. By Theorem 3.19, the theorem is an immediate corollary of Theorem 4.29. u
t
The proof of Lemma 3.35 can similarly be adapted to prove the corresponding
generalization of Lemma 3.35.
Lemma 4.31. Suppose that the nonstationary ideal on !1 is semi-saturated. Suppose
A R and that B is weakly homogeneously Suslin for each set B which is projective
in A. Let M be a transitive set such that M ZFC , P .!1 / M , and such that M #
exists. Then
X M j X is countable and MX is A-iterable
t
u
contains a club in P!1 .M /. Here MX is the transitive collapse of X .
Finally we obtain the generalization of the second covering theorem, Theorem 3.47,
to the case when INS is simply assumed to be semi-saturated.
Theorem 4.32. Suppose that the nonstationary ideal on !1 is semi-saturated. Suppose
that M is a transitive inner model such that
M ZF C DC C AD;
and such that
(i) R M ,
(ii) Ord M ,
(iii) every set A 2 M \ P .R/ is weakly homogeneously Suslin in V .
Suppose < M , X and jX j D !1 . Then there exists Y 2 M such that
M jY j D !1
and such that X Y .
Proof. This is an immediate corollary of Lemma 4.31, applied to the set, H.!2 /, and
Theorem 3.45.
t
u
136
4.2
t
u
Remark 4.34. (1) Given the results of Section 3.1 it would be more natural to dene
Pmax as the set of pairs .M; a/ where M is an iterable model in which the nonstationary ideal is saturated. Assuming
12 -Determinacy this yields an equivalent
forcing notion. More precisely assuming
12 -Determinacy, the set of conditions
h.M; I /; ai 2 Pmax such that I is a saturated ideal in M and such that I is the
nonstationary ideal in M, is dense in Pmax .
(2) We shall prove that the nonstationary ideal is saturated in L.R/Pmax and that
ZFC holds there. Thus Pmax is in some sense converting the existence of models
with precipitous ideals (which are relatively easy to nd) into the existence of
models in which the nonstationary ideal on !1 is saturated. This is an aspect
we shall exploit when we modify Pmax to show the relative consistency that the
t
u
nonstationary ideal on !1 is !1 -dense.
There are equivalent versions of Pmax that do not require that the models which
appear in the conditions be models of MA!1 , this is a degree of freedom which is
essential for the variations that we shall dene. In Chapter 5 we shall give three other
(T)
0
, Pmax
and Pmax
. The rst of these will involve
presentations of Pmax , denoted by Pmax
137
using the generic elementary embeddings associated to the stationary tower in place of
embeddings associated to ideals on !1 . The second will be closer to Pmax , however
the stationary tower will be used to generate the necessary conditions and so certain
aspects of the analysis will differ. In fact there are strong arguments to support the
0
is actually the best presentation of Pmax . The third, Pmax
, is
claim that in the end, Pmax
(T)
a combination of Pmax and Pmax . In dening two of the variations of Pmax , we shall use
these alternate formulations as a template, see Denition 6.54 and Denition 8.30.
The following lemma indicates the utility of working with models of MA. We state
it in a more general form than is strictly necessary for the analysis of Pmax .
Lemma 4.35. Suppose M is a countable transitive model of ZFC C MA!1 . Suppose
a 2 M,
a !1M ;
and
Suppose
j1 W M ! M1
and
j2 W M ! M2
are semi-iterations of M such that
(i) M1 is transitive,
(ii) M2 is transitive,
(iii) j1 .a/ D j2 .a/,
(iv) j1 .!1M / D j2 .!1M /.
Then M1 D M2 and j1 D j2 .
Proof. This is a relatively standard fact. The key point, which we prove below, is that
since both
j1 .a/ D j2 .a/
and
j1 .!1M / D j.!2M /;
it follows that j1 .b/ D j2 .b/ for each set b 2 M such that b !1M . From this it
follows easily by induction that at every stage the generic lters are the same and so
j1 D j 2 .
Let hs W < !1M i be the sequence of almost disjoint subsets of ! where for each
< !1M , s is the rst subset of ! constructed in L.a; x/ which is almost disjoint
from s for each < . Thus
hs W < !1M i 2 L.a; x/
138
Let
ht W <
i D j1 .hs W < !1M i/
where
D j1 .!1M / D j2 .!1M /. Suppose that b 2 M and b !1M . Since
M MA!1
it follows that there exists t 2 M such that t almost disjoint codes b relative to
hs W < !1M i; i. e. b D j t \ s is innite. Therefore
j1 .b/ D <
j t \ t is innite D j2 .b/:
Therefore for each b 2 M such that b !1M , j1 .b/ D j2 .b/. The lemma follows.
t
u
The next two lemmas are key to proving many of the properties of the partial order
Pmax . Because we wish to apply them within the models occurring in conditions we
work in ZFC .
Lemma 4.36 (ZFC ). Suppose .M; I / is a countable transitive iterable model where
I 2 M is a normal uniform ideal on !1M and M ZFC . Suppose J is a normal
uniform ideal on !1 . Then there exists an iteration
j W .M; I / ! .M ; I /
such that:
(1) j.!1M / D !1 ;
(2) J \ M D I .
Proof. Fix a sequence hAk; W k < !; < !1 i of J -positive sets which are pairwise
disjoint. The ideal J is normal hence each Ak; is stationary in !1 . We suppose that
Ak; \ . C 1/ D ;.
Fix a function
f W ! !1M ! P .!1M / \ M n I
such that
(1.1) f is onto,
(1.2) for all k < !, f jk !1M 2 M,
(1.3) for all A 2 M if A has cardinality !1M in M and if A P .!1M / n I then
A ran.f jk !1M / for some k < !.
139
and it is easily veried that the range of j .f / is P .!1M /\M nI . This follows
from (1.3).
We construct an iteration of M of length !1 using the function f to provide a
book-keeping device for all of the subsets of !1 which belong to the nal model and
do not belong to the image of I in the nal model. More precisely construct an iteration
h.M ; I /; G ; j; W < !1 i such that for each < !1 , if !1M 2 Ak; then
j0; .f /.k; / 2 G .
The set C D j0; .!1M / j < !1 is a club in !1 . Thus for each B !1 such
that B 2 M!1 and B j0;!1 .I / there exists k < !; < !1 such that
.C n C 1/ \ Ak; B \ Ak; :
Further if B !1 , B 2 M!1 and B 2 j0;!1 .I / then B \ C D ;.
Thus J \ M!1 D I!1 .
t
u
Lemma 4.37 is the analog of Lemma 4.36 for iterable sequences. The proof is a
straightforward modication of the proof of Lemma 4.36.
Lemma 4.37 (ZFC ). Suppose h.Nk ; Jk / W k < !i is an iterable sequence such that
Nk ZFC for each k < !. Suppose J is a normal uniform ideal on !1 . Then there
exists an iteration
j W h.Nk ; Jk / W k < !i ! h.Nk ; Jk / W k < !i
such that:
(1) j.!1N0 / D !1 ;
(2) J \ Nk D Jk for each k < !.
t
u
140
t
u
Remark 4.39. Assuming that for every real x, x exists, it follows that the set of
conditions h.M; I /; ai 2 Pmax for which M ZFC is dense in Pmax . Thus in the
denition of Pmax the fragment of ZFC used is not really relevant provided it is strong
enough.
t
u
For the analysis of Pmax we need a much stronger existence theorem for conditions.
Lemma 4.40. Assume AD holds in L.R/. Suppose that X R and that X 2 L.R/.
Then there is a condition h.M; I /; ai 2 Pmax such that
(1) X \ M 2 M,
(2) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(3) .M; I / is X -iterable,
and further the set of such conditions is dense in Pmax .
Proof. We work in L.R/.
Suppose that for some X R with X 2 L.R/ no such condition
h.M; I /; ai 2 Pmax
exists. Then by standard reection arguments in L.R/ we may assume that X is
21
denable in L.R/. By the Martin-Steel theorem, Theorem 2.3, in L.R/ the pointclass
21 denable in L.R/
21 has the scale property. Thus any set X R R which is
is Suslin in L.R/ and so can be uniformized by a function which is
21 denable in
141
142
and
!1 D !1La :
143
This includes the proof that the nonstationary ideal on !1 is saturated in L.R/Pmax .
However we shall see in Chapter 5 that this assumption implies ADL.R/ .
This property for a set of reals, X , is really a regularity property which can be
established from a variety of different assumptions. For example, it can be established
quite easily from just the assumption that every set of reals which is projective in X is
weakly homogeneously Suslin.
Theorem 4.41. Suppose X R and that every set of reals which is projective in X is
weakly homogeneously Suslin. Then there is a condition
h.M; I /; ai 2 Pmax
such that
(1) X \ M 2 M,
(2) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(3) .M; I / is X -iterable.
Proof. Note that since there are nontrivial weakly homogeneously Suslin sets there
must exist a measurable cardinal. Let be the least measurable cardinal and let I be
a normal uniform ideal on such that I is maximal; i. e. the dual lter is a normal
measure.
By collapsing 2 to C if necessary we can assume that 2 D C . The generic
collapse of 2 to C preserves the hypothesis of the theorem and it adds no new reals
to V .
X is weakly homogeneously Suslin and so there exists a weakly homogeneous tree
S such that X D pS . The tree S is necessarily -weakly homogeneous.
Let S be a weakly homogeneous tree such that pS D R n X . Again S is
necessarily -weakly homogeneous and so if G P is V -generic where P is a partial
order of size less than then in V G, pS D R n pS .
Let Y be the set of reals which code elements of the rst order diagram of
hH.!1 /; X; 2i:
Y is weakly homogeneously Suslin since it is a countable union of weakly homogeneously Suslin sets. Similarly R n Y is also weakly homogeneous Suslin since it too is
the countable union of weakly homogeneously Suslin sets.
Therefore there exist weakly homogeneous trees T and T such that
pT D Y
and such that pT D R n Y . The trees T and T are each necessarily -weakly
homogeneous.
Thus if G P is V -generic where P is a partial order of size less than , then in
V G, pT D R n pT .
A key point is that is measurable and so this also holds if P is a partial order
which is -cc.
144
145
146
A P .!1 k / \ Nk n Jk ;
and A is dense;
(1.7) JkC1 \ Nk D Jk ;
(1.8) if C 2 Nk is closed and unbounded in !1N0 then there exists D 2 NkC1 such
that D C , D is closed and unbounded in C and such that
D 2 Lx
for some x 2 R \ NkC1 .
These properties are straightforward to verify, (1.6) follows from Lemma 4.10 and
(1.8) follows from Lemma 4.6.
By Corollary 4.20, the sequence h.Nk ; Jk / W k < !i is iterable. Let z be a real
which codes h.Nk ; Jk / W k < !i. Thus there is a condition h.M; I /; ai 2 Pmax such
that z 2 M. By Lemma 4.37, there is an iteration
j W h.Nk ; Jk / W k < !i ! h.Nk ; Jk / W k < !i
such that:
(2.1) j 2 M;
(2.2) j.!1N0 / D !1M ;
(2.3) I \ Nk D Jk for each k < !.
Let a D j.b/. Thus h.M; I /; a i 2 Pmax and h.M; I /; a i < h.Mk ; Ik /; ak i for all
k < !. This shows that Pmax is !-closed.
We nish by showing that Pmax is homogeneous. Suppose h.M0 ; I0 /; a0 i and
h.M1 ; I1 /; a1 i are conditions in Pmax . Let z be a real which codes the pair of these
conditions. Suppose h.M; I /; ai is a condition in Pmax such that z 2 M. Thus there
are iterations
j0 W .M0 ; I0 / ! .M0 ; I0 /
and
j1 W .M1 ; I1 / ! .M1 ; I1 /
such that:
(3.1) j0 2 M and j1 2 M;
(3.2) j0 .!1M0 / D !1M D j1 .!1M1 /;
(3.3) I \ M0 D I0 and I \ M1 D I1 .
147
Let a0 D j0 .a0 / and let a1 D j1 .a1 /. The key point is the following. Suppose that
h.N ; J /; bi 2 Pmax and h.N ; J /; bi < h.M; I /; ai. Let
j W .M; I / ! .M ; I /
be the unique iteration such that j.a/ D b. Then
h.N ; J /; j.a0 /i < h.M0 ; I0 /; a0 i
and
t
u
Using the iteration lemmas we prove two more lemmas which we shall use to
complete our initial analysis of Pmax . We begin with a denition that establishes some
key notation.
Denition 4.44. A lter G Pmax is semi-generic if for all < !1 there exists a
condition h.M; I /; ai 2 G such that < !1M .
Suppose G Pmax is semi-generic. Dene AG !1 by
AG D [a j h.M; I /; ai 2 G:
For each h.M; I /; ai 2 G let
jG W .M; I / ! .M ; I /
be the embedding from the iteration which sends a to AG .
Let
P .!1 /G D [P .!1 / \ M j h.M; I /; ai 2 G
and let
IG D [I j h.M; I /; ai 2 G:
t
u
Remark 4.45. (1) Suppose G Pmax is a semi-generic lter. Then Pmax is somewhat nontrivial. Strictly speaking, a lter G Pmax may be, for example, L.R/generic and not be semi-generic. We shall never consider lters in Pmax without
assumptions which guarantee that Pmax is nontrivial.
148
149
Proof. Let
h.p ; f / W < !1 i
be a sequence such that for all < < !1
(1.1) p0 < q0 ,
(1.2) .p ; f / 2 Y ,
(1.3) p < p ,
(1.4) f f ,
(1.5) dom.f /,
(1.6) J \ M D I ,
where .M ; I / is dened as follows. Let h.M ; I /; a i D p . Let
a D [a j < !1 :
Then for each there exists a unique iteration
j W .M ; I / ! .M ; I /
such that j .a / D a . This sequence is easily constructed using the properties of Y
and the proof of Lemma 4.36.
Let G be the lter generated by p j < !1 and let
f D [f j < !1 :
Thus G is a semi-generic lter and .G; f / has the desired properties.
t
u
The next lemma is simply the formulation of Lemma 4.10 for the special case we
are presently interested in. This is the case for structures of the form .M; I /; i. e. when
only one ideal is designated.
Lemma 4.47 (ZFC ). Suppose .M; I / is a countable transitive model where
I 2M
is a normal uniform ideal on !1M and M ZFC . Suppose that
j W .M; I / ! .M ; I /
t
u
150
Lemma 4.48 (ZFC ). Assume Pmax ; and that for each x 2 R, the set of
h.M; I /; ai 2 Pmax
such that x 2 M is dense in Pmax .
Suppose J is a normal uniform ideal on !1 and that Y H.!1 / is a .nonempty/
set of pairs .p; b/ such that:
(i) p 2 Pmax ;
(ii) b !1M , b 2 M, and b I ;
where p D h.M; I /; ai.
(iii) Suppose .h.M0 ; I0 /; a0 i; b0 / 2 Y and h.M1 ; I1 /; a1 i < h.M0 ; I0 /; a0 i. Then
.h.M1 ; I1 /; a1 i; b1 / 2 Y where b1 is the image of b0 under the iteration of
.M0 ; I0 / which sends a0 to a1 .
(iv) Suppose h.M0 ; I0 /; a0 i 2 Pmax , b0 2 M0 , b0 !1M0 and b0 I0 . Then there
exists .h.M1 ; I1 /; a1 i; b1 / 2 Y such that
h.M1 ; I1 /; a1 i < h.M0 ; I0 /; a0 i
and such that b1 j.b0 / where j is the embedding given by the iteration of
.M0 ; I0 / which sends a0 to a1 .
Then for each p0 2 Pmax there exists a semi-generic lter G Pmax such that p0 2 G,
J \ P .!1 /G D IG ;
jP .!1 /G j D !1 ;
and such that
!1 n 5A 2 J
where A is the set of j.b/ such that .h.M; I /; ai; b/ 2 Y , h.M; I /; ai 2 G, and
j W .M; I / ! .M ; I /
is the embedding given by the iteration of .M; I / which sends a to AG .
Proof. Let
S < !1 j is a limit ordinal
and x a partition hS W < !1 i of S into disjoint sets such that S D 5S j < !1
and such that S J for each < !1 . For any uniform normal ideal such a partition
exists.
We construct a sequence h.q ; b / W < !1 i of elements of Y such that for all
< < !1 , q < q < p0 and such that;
(1.1) for each < !1 there is a club C !1 such that S \ C j .b /,
(1.2) for each < !1 and for each d 2 P .!1M / \ M with d I there exists
< !1 such that < and b j; .d /,
151
in M such that j.!1N / D !1M and such that for all k < !, I \ Nk D Jk . Let
a D j.a /. Thus
h.M; I /; a i 2 Pmax
and for all <
152
Thus by property (iii) of Y there exists .q; b/ 2 Y such that q < h.M; I /; a i. Let
.q ; b / D .q; b/ and let d 2 P .!1M / \ M n I .
This completes the construction of the sequences. Notice that we have complete
freedom in the choice of d at each stage . Let G Pmax be the lter generated by
q j < !1 . We may assume by a routine book-keeping argument that
j .d / j < !1 D [M \ P .!1 / j < !1 n IG D P .!1 /G n IG :
We claim that G is the desired semi-generic lter. G is generated by a subset of
size !1 and so it follows that jAG j D !1 . All that needs to be veried is that 5AG is
of measure 1 relative to J and that IG D J \ P .!1 /G .
For each < !1 there is a club C !1 such that C \ S j .b /. Further by
denition j .b / 2 AG and so since S D 5S j < !1 it follows that there is a
club C !1 such that S \ C 5AG , take C D 4C j < !1 . However S is of
measure 1 relative to J and J is a uniform normal ideal. Hence C \ S is of measure 1
relative to J .
By the choice of hd W < !1 i it follows that
j .d / j < !1 D P .!1 /G n IG :
.bC1 / j .d /. Therefore every set in
However for each < !1 , jC1
P .!1 /G n IG is positive relative to J . Further every set in IG is nonstationary and
so
IG D J \ P .!1 /G :
t
u
Suppose G Pmax is L.R/-generic. We assume also that for all reals x, x exists
so that Pmax is nontrivial. Thus the lter G is semi-generic and so we have dened
AG !1 , P .!1 /G P .!1 /, and IG P .!1 /G .
The next theorem gives the basic analysis of Pmax .
Theorem 4.49. Suppose that for each set X R with X 2 L.R/, there is a condition
h.M; I /; ai 2 Pmax such that
(i) X \ M 2 M,
(ii) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(iii) .M; I / is X -iterable.
Suppose G Pmax is L.R/-generic. Then
L.R/G !1 -DC
and in L.R/G:
(1) P .!1 /G D P .!1 /;
(2) IG is a normal saturated ideal;
(3) IG is the nonstationary ideal.
153
Proof. We claim that for each set X R with X 2 L.R/ the set of such conditions
in Pmax which satisfy (i)(iii) is dense in Pmax . The point here is that given X and
a condition h.M0 ; I0 /; a0 i 2 Pmax dene a new set X R as follows. Fix a real z
which codes h.M0 ; I0 /; a0 i and dene X to be the set of reals which code a pair .z; t /
where t 2 X . We assume X is nonempty. Thus X 2 L.R/ and so there is a condition
h.M; I /; ai 2 Pmax such that
(1.1) X \ M 2 M,
(1.2) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(1.3) .M; I / is X -iterable.
By (1.2) it follows that z 2 M. Thus by Lemma 4.36, there is an iteration
j W .M0 ; I0 / ! .M0 ; I0 /
in M such that I D I \ M0 . Thus h.M; I /; j.a0 /i 2 Pmax and
h.M; I /; j.a0 /i < h.M0 ; I0 /; a0 i:
h.M; I /; j.a0 /i is the desired condition.
Therefore by Lemma 4.43, Pmax is !-closed and homogeneous.
We rst prove that if G Pmax is L.R/-generic then !1 -DC holds in L.R/G.
Since Pmax is !-closed it follows that DC holds in L.R/G. Every set in L.R/G
is denable from an ordinal, a real and G. Therefore to establish that !1 -DC holds
in L.R/G it sufces to show that if T 0; 1<!1 is an !-closed subtree, closed
under initial segments, and with no maximal elements, then there exists a function
F W !1 ! 0; 1 such that F j 2 T for all < !1 .
Fix a term 2 L.R/ for such a tree T and x a condition
h.M0 ; I0 /; a0 i 2 Pmax :
Clearly we may suppose for any q 2 Pmax , q forces that is an !-closed subtree of
0; 1<!1 , closed under initial segments and containing no maximal elements. We shall
apply Lemma 4.46 to obtain a term which is forced by a condition below h.M0 ; I0 /; a0 i
to be a term for a conal branch of the tree dened by .
Let Y be the set of all pairs .p; f / such that:
(2.1) p 2 Pmax ;
(2.2) p f 2 ;
(2.3) f 2 0; 1 for some < !1 .
Let X be the set of reals which code elements of Y . Thus since Y 2 L.R/,
X 2 L.R/. Let .M; I / be a countable iterable structure such that
(3.1) X \ M 2 M,
(3.2) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(3.3) .M; I / is X -iterable.
154
We may assume that M contains a real coding h.M0 ; I0 /; a0 i by the remarks above.
The following closure properties of Y can be expressed as rst order statements in
the structure hH.!1 /; X; 2i.
(4.1) Suppose .p; f / 2 Y and q < p. Then .q; f / 2 Y .
(4.2) Suppose .p; f / 2 Y and < dom.f /. Then .p; f j/ 2 Y .
(4.3) Suppose .p; f / 2 Y and < !1 . Then there exists .q; g/ 2 Y such that
q < p, f g and such that < dom.g/.
(4.4) Suppose p 2 Pmax , < !1 , is a limit ordinal and
f W ! 0; 1:
Then either .p; f / 2 Y or .p; f j/ Y for some < .
Since
it follows that
hH.!1 /M ; X \ Mi hH.!1 /; X i
hH.!1 /M ; Y \ H.!1 /M ; 2i hH.!1 /; Y; 2i:
.p; f j/ 2 Y
155
156
where
is dense:
Let Y H.!1 / be the set of all pairs .h.M; I /; ai; b/ such that,
(7.1) h.M; I /; ai 2 Pmax ,
(7.2) b 2 M and b !1M ,
(7.3) h.M; I /; ai b 2 ,
where if G Pmax is L.R/-generic and h.M; I /; ai 2 G then b is the image of b
under the iteration of .M; I / which sends a to AG .
Observe that because IG is the nonstationary ideal it follows that if
.h.M; I /; ai; b/ 2 Y
then necessarily b I .
The following properties of Y are easily veried.
(8.1) Suppose .h.M0 ; I0 /; a0 i; b0 / 2 Y and h.M1 ; I1 /; a1 i < h.M0 ; I0 /; a0 i. Then
.h.M1 ; I1 /; a1 i; b1 / 2 Y where b1 is the image of b0 under the iteration of
.M0 ; I0 / which sends a0 to a1 .
(8.2) Suppose h.M0 ; I0 /; a0 i 2 Pmax , b0 2 M0 , b0 !1M0 and b0 I0 . Then there
exists .h.M1 ; I1 /; a1 i; b1 / 2 Y such that
h.M1 ; I1 /; a1 i < h.M0 ; I0 /; a0 i
and such that b1 j.b0 / where j is the embedding given by the iteration of
.M0 ; I0 / which sends a0 to a1 .
157
The second of properties, (8.2), follows from the fact that if G Pmax is L.R/generic then in L.R/G
!1 n 5AG1 2 I1
where AG1 is the set of subsets of !1 given by evaluating using G1 and using
Y \ H.!1 /M :
Let
a10 D A
a10 D [b j h.N ; J /; bi 2 G1
h.M1 ; I1 /; a10 i < h.N ; J /; bi
158
Now suppose that G Pmax is L.R/-generic and that h.M1 ; I1 /; a10 i 2 G. Let
j W .M1 ; I1 / ! .M1 ; I1 /
be the embedding from the iteration which sends a10 to AG . The key point is that since
.M1 ; I1 / is X -iterable it follows that
j.Y M1 / D Y \ M1 :
Further suppose h.M; I /; ai < h.M1 ; I1 /; a10 i and let
k W .M1 ; I1 / ! .M1 ; I1 /
be the countable iteration of .M1 ; I1 / which sends a10 to a. By the properties of G1 in
M1 it follows that h.M; I /; ai < p for all p 2 k.G1 /. From these facts it follows that
j.AG1 / AG :
However 5AG1 is of measure 1 in M1 relative to I1 . Therefore 5j.AG1 / is of measure 1 relative to IG and so it contains a club in !1 since IG is the nonstationary ideal
in L.R/G.
Finally AG1 is of cardinality !1 in M1 and so j.AG1 / has cardinality !1 in
L.R/G. Thus j.AG1 / is the desired subset of AG .
This proves that IG is a saturated ideal in L.R/G and this completes the proof of
the theorem.
t
u
Combining Lemma 4.40 and Theorem 4.49 we obtain as an immediate corollary
the following theorem.
Theorem 4.50. Assume AD holds in L.R/. Suppose G Pmax is L.R/-generic. Then
L.R/G !1 -DC
and in L.R/G:
(1) P .!1 /G D P .!1 /;
(2) IG is a normal saturated ideal;
(3) IG is the nonstationary ideal.
t
u
We continue with our analysis of L.R/Pmax and prove that the conclusion of Corollary 3.48 holds in L.R/Pmax . This theorem can also be proved abstractly by using
Corollary 3.48 together with the absoluteness theorem, Theorem 4.64. But a proof
along these lines requires stronger hypotheses.
Remark 4.51. (1) It is possible to prove that INS is saturated in L.R/Pmax using
Lemma 4.52 instead of Lemma 4.48, see the proof of Theorem 10.54.
(2) There are Pmax -variations, P 2 L.R/, for which INS is not saturated in L.R/P .
However Lemma 4.52 will generalize to these models, yielding the semi-satut
u
ration of INS in these models, see Section 6.1.
159
Lemma 4.52. Assume ADL.R/ and suppose that G Pmax is L.R/-generic. Then in
L.R/G, for every set A 2 P .R/ \ L.R/ the set
X hH.!2 /; A; 2i j MX is A-iterable and X is countable
contains a club, where MX is the transitive collapse of X .
Proof. Suppose G Pmax is L.R/-generic. From the basic analysis of Pmax summarized in Theorem 4.49 it follows that
H.!2 /L.R/G D H.!2 /L.R/ AG :
We work in L.R/G. Fix A R with A 2 L.R/. Fix a countable elementary
substructure
X hH.!2 /; A; G; 2i:
Let hXi W i < !i be an increasing sequence of countable elementary substructures of
X such that
X D [Xk j k < !
and such that for each k 2 !, Xk 2 XkC1 . Therefore for each k < !, there exists
h.M; I /; ai 2 G \ XkC1 satisfying
(1.1) Xk \ P .!1 / M ,
(1.2) A \ M 2 M,
(1.3) .M; I / is A-iterable,
where M is the iterate of M given by the iteration of .M; I / which sends a to AG .
Let MX be the transitive collapse of X . We claim that MX is A-iterable. Given this
the lemma follows.
For each k < ! let h.Mk ; Ik /; ak i 2 G \ XkC1 be a condition satisfying the
requirements (1.1), (1.2) and (1.3). For each k < ! let
k
h.Mk ; Ik /; G ; j;
W < !1 i
k
be the iteration of .Mk ; Ik / such that j0;!
.ak / D AG . Thus for each k < !,
1
k
h.Mk ; Ik /; G ; j;
W < !1X i 2 MX
and further
MX D [Mk j k < !;
where D X \ !1 .
Suppose
j W MX ! N
is given by a countable iteration of MX . Let
D j.!1MX /:
For each k < ! let
.Nk ; Jk / D j..Mk ; Ik //
160
where D !1MX D !1X . Therefore for each k < !, .Nk ; Jk / is an iterate of .Mk ; Ik /
by an iteration of length which extends the iteration
k
h.Mk ; Ik /; G ; j;
W < i:
For each k < ! this is the (unique) iteration of .Mk ; Ik / which sends ak to
j.AG \ !1MX /. By induction on ,
N D [Nk j k < !
and so MX is iterable. The argument here is identical to proof that Pmax is !-closed,
cf. Lemma 4.43. We nish by analyzing
AQ D [j.B/ j B A and B 2 MX :
We must show that AQ D A \ N .
Let D !1MX . Thus MX D [Mk j k < !. For each k < !, .Mk ; Ik / is
A-iterable. Therefore
k
A \ MX D [j0;
.A \ Mk / j k < !:
For each k < ! let AQk be the image of A \ Mk under the iteration of .Mk ; Ik /
which sends ak to j.AG \ /. This is the iteration which denes Nk . Thus
AQ D [AQk j k < !
since for all B 2 MX , B A if and only if
B Mk \ A
for some k < ! and since for all k < !,
k
Mk \ A D j0;
.A \ Mk /:
161
Proof. (1) is an immediate corollary to Lemma 4.52 and Theorem 3.19. (1) also follows from (2).
We prove (2). Let D L.R/ and suppose
f W S ! Ord
where S is a stationary subset of !1 .
By the chain condition satised by Pmax in L.R/, there exists a set X Ord such
that X 2 L.R/, f S X and such that
jX j <
in L.R/.
Therefore we may suppose that
f WS !
for some
< . (2) now follows from Lemma 4.52 and Theorem 3.42.
t
u
t
u
Denition 4.55. Suppose that A !1 . The set A is L.R/-generic for Pmax if there
exists a lter G Pmax which is L.R/-generic and such that A D AG .
t
u
The following lemma shows that the generic for Pmax can be identied with the
subset of !1 it creates.
Lemma 4.56. Assume ADL.R/ . Suppose that A !1 is L.R/-generic for Pmax . Dene in L.R/A a subset F Pmax as follows.
h.M; I /; ai 2 F
if there exists an iteration
j W .M; I / ! .M ; I /
such that
(1) j.a/ D A,
(2) I D INS \ M .
Then F is a lter in Pmax , F is L.R/-generic and A D AF .
Proof. Fix a lter G Pmax such that G is L.R/-generic and such that
A D AG :
Note that for each h.M; I /; ai 2 G, the corresponding iteration
j W .M; I / ! .M ; I /
162
such that j.a/ D A can be computed in L.A; .M; I // and so M 2 L.R/A. Therefore by Theorem 4.50 it follows that
P .!1 /L.R/G L.R/A:
Thus the set F Pmax is the same computed in L.R/A or L.R/G.
Thus it sufces to show that in L.R/G,
F D G:
By Theorem 4.50, G F and so we need only show that F G, i. e. that
the requirement specifying membership in F fails for conditions which do not belong
to G.
Suppose h.M; I /; ai 2 Pmax and h.M; I /; ai G. We prove that
h.M; I /; ai F :
Let z 2 R code h.M; I /; ai. Therefore there is a condition h.N ; J /; bi 2 G such
that z 2 N and such that h.M; I /; ai and h.N ; J /; bi are incompatible. First suppose
there is no iteration of .M; I / which sends a to b. If there exists an iteration of .M; I /
which sends a to A then it is easily veried that there must be an iteration of .M; I /
which sends a to b. Therefore there is no iteration of .M; I / which sends a to A and
so
h.M; I /; ai F :
Therefore we may assume that there is an iteration
k W .M; I / ! .M ; I /
such that k.a/ D b. The iteration k is unique and k 2 N . If I D M \ J then
h.N ; J /; bi < h.M; I /; ai
which contradicts the incompatibility of these conditions. Therefore
I M \ J
in particular there must exist B 2 P .!1N / \ M n I such that B 2 J . Let
j W .N ; J / ! .N ; J /
be the iteration such that j .b/ D A. Thus
j .k/ W .M; I / ! .M ; I /
is the iteration of .M; I / which sends a to A. But j .B/ 2 M n I and
j .B/ 2 j .J /. Therefore j .B/ is nonstationary in L.R/G since j .J / IG
and IG is the nonstationary ideal. Thus
h.M; I /; ai F ;
and this proves F D G.
t
u
163
hZ W < !1 i
164
165
For example assume the nonstationary ideal on !1 is !2 -saturated, there is a measurable cardinal and there is a transitive inner model of ZF C DC containing the reals,
containing the ordinals, and in which the club lter on !1 is an ultralter. Then in
L.R/ every subset of !1 is constructible from a real.
Lemma 4.59. Assume that for every real x, x exists. Suppose
h.M; I /; ai 2 Pmax ;
d
!1M
and d 2 M.
;x/
(1.4) every subset of !1 kC1 which belongs to LdkC1 ; xk either contains or is disM
joint from a tail of the indiscernibles of LxkC1 below !1 kC1 .
where for each k < !, pk D h.Mk ; Ik /; ak i, dk D jk .d / and jk is the elementary
embedding from the unique iteration of .M; I / such that jk .a/ D ak . Implicit in (1.4)
M
M
is the fact that if A !1 k and if A 2 Mk then every subset of !1 k which is in LA
#
belongs to L A where D Mk \ Ord. This is because A 2 Mk which in turn
follows from the iterability of .Mk ; Jk /. We use this frequently.
Choose a condition h.N ; J /; bi 2 Pmax such that for all k < !,
h.N ; J /; bi < h.Mk ; Ik /; ak i:
For each k < ! let
166
be the unique iteration such that jk .ak / D b. Let d D jk .dk /. This is unambiguously
dened and we may apply Lemma 4.57 in N to obtain that there is a real t 2 N such
that d 2 Lt . The condition h.N ; J /; bi 2 D1 and h.N ; J /; bi < p.
The second case is that no such sequence
h.pk ; xk / W k < !i
exists. Notice that if
h.N1 ; J1 /; b1 i < h.N0 ; J0 /; b0 i
in Pmax and if
is the unique iteration such that j.b0 / D b1 then for every D 2 J a tail of indiscernibles of Lx below !1N1 is disjoint from D where x is any real in N1 which codes
N0 .
Therefore since the sequence h.pk ; xk / W k < !i does not exist it follows that there
exist a condition
h.N0 ; J0 /; b0 i < p;
a real x0 2 N0 , and a set D !1N0 , such that
(2.1) D 2 Lx0 ; d0 ,
(2.2) both D and !1N0 n D are positive relative to J0 ,
where d0 D j.d / and j W .M; I / ! .M ; I / is the unique iteration such that
j.a/ D b0 .
Fix a condition
h.N1 ; J1 /; b1 i < h.N0 ; J0 /; b0 i:
By modifying b1 we shall produce a condition in D1 below h.N0 ; J0 /; b0 i.
We work in N1 . Fix a real t which codes .N0 ; J0 /. Let C be the set
C D < !1N1 j L t Lt :
Therefore C is a club in !1N1 and C 2 Lt . Let XC be the elements of C which are
not limit points of C and let
W !1N1 ! XC
be the enumeration function of XC .
Fix A !1N1 such that A 2 N1 and !1N1 D !1LA . Let A C be the image of
A under . Working in N1 construct an iteration
j0 W .N0 ; J0 / ! .N0 ; J0 /
of length .!1 /N1 such that;
(3.1) J0 = J1 \ N0 ,
(3.2) j0 .D/ \ XC D A .
167
The iteration exists because the requirements given by (2.1) and (2.2) do not interfere. One achieves (2.1) by working on C n XC as in the proof of Lemma 4.36 and
(2.2) is achieved by working on XC .
Let b1 D j0 .b0 /, let d1 D j0 .d0 / and let D D j0 .D/. Thus
h.N1 ; J1 /; b1 i < h.N0 ; J0 /; b0 i
and
!1N1 D !1Lt;D
since A 2 Lt; D . However D 2 Lx; d0 and so D 2 Lx; d1 . Therefore
and so
h.N1 ; J1 /; b1 i
!1N1 D !1Lx;t;d1
2 D0 .
t
u
The next theorem reinforces the analogy between Pmax and Sacks forcing.
Theorem 4.60. Assume AD holds in L.R/. Suppose that G Pmax is a lter which
is L.R/-generic. Suppose that A !1 and that A 2 L.R/G n L.R/. Then A is
L.R/-generic for Pmax and
L.R/G D L.R/A:
Proof. This is immediate, the argument is similar to that for the homogeneity of Pmax
together with the analysis provided by Theorem 4.49 and Lemma 4.59.
Let G Pmax be L.R/-generic. Fix A !1 , A 2 L.R/G n L.R/. By
Theorem 4.49 there exists a condition
h.M0 ; I0 /; a0 i 2 G
such that for some d 2 M0 , j .d / D A where
j W .M0 ; I0 / ! .M0 ; I0 /
is the iteration which sends a to AG . By Lemma 4.59 we may assume that
M0 !1 D !1L.d;x/ for some real x:
Therefore there exists a real x 2 M0 such that
!1 D !1LA;x :
We rst show that L.R/G D L.R/A. Since
M0 MA!1
it follows, by Lemma 4.35, that there exists a real y 2 M0 with
AG 2 LA; y:
Therefore L.R/A D L.R/AG D L.R/G.
To nish we must prove that A is L.R/-generic for Pmax .
Let g Pmax be the lter generated by
h.N ; J /; A \ !1N i j h.N ; J /; bi 2 G and h.N ; J /; bi < h.M0 ; I0 /; a0 i:
It follows that g is L.R/-generic and that
A D [b j h.N ; J /; bi 2 gI
i. e. that A is the set AG computed from g.
t
u
Therefore A is L.R/-generic for Pmax .
168
The next theorem is the key for actually verifying that specic 2 sentences hold
in
Pmax
Pmax
9x .x/:
hH.!2 /M1 ; 2; I1 i
b1
The next theorem is simply a reformulation. This theorem strongly suggests that if
AD holds in L.R/ and if
G Pmax
is L.R/-generic then in L.R/G one should be able to analyze all subsets of P .!1 /
which are denable in the structure
hH.!2 /; 2; INS iL.R/G
by a 1 formula. Thus while a 2 sentence may fail in L.R/G one can analyze
completely the counterexamples.
169
170
There is a stronger absoluteness theorem that is true and this is the version which
we prove.
Theorem 4.65. Assume ADL.R/ and that there is a Woodin cardinal with a measurable above. Suppose that J is a normal uniform ideal on !1 , is a 2 sentence in the
language for the structure
hH.!2 /; 2; J i;
and that
hH.!2 /; 2; J i :
Then
Pmax
:
Proof. Let .x; y/ be a 0 formula such that D 8x9y: .x; y/ (up to logical
equivalence).
Assume towards a contradiction that
Pmax
::
hH.!2 /M ; 2; I i 8y b
where b D j.b0 / and j W .M0 ; I0 / ! .M0 ; I0 / is the iteration such that j.a0 / D a.
By Lemma 4.36, there is an iteration
j W .M0 ; I0 / ! .M0 ; I0 /
such that:
(1.1) j.!1M0 / D !1 ;
(1.2) J \ M0 D I0 .
Let B D j.b0 /. The sentence holds in V and so there exists a set D 2 H.!2 /
such that
hH.!2 /; 2; J i : B; D:
Let be a Woodin cardinal and be a measurable cardinal above . Let g be
a V -generic enumeration of J of length !1 . The poset is simply J <!1 ordered by
extension. In V g let
S D 5hS W < !1 i
be the diagonal union of the generic enumeration of J . Thus S is co-stationary in V g.
This is because J is a normal ideal in V .
171
172
is an iteration of .M0 ; I0 / of length .!1 /N0 and further I0 D J0 \ M0 . The latter
holds since I0 D J \ M0 and since in V gC G,
V gC G
J D INS
\ V:
hH.!2 /M1 ; I1 ; 2i 6 8y b
where b D j0 .b/ and j0 W .M0 ; I0 / ! .M0 ; I0 / is the iteration such that
j0 .a0 / D a1 .
By absoluteness, noting that V is 13 -correct in the generic extension, V gC G,
such a condition h.M; I /; ai must exist in V , which is a contradiction.
t
u
Denition 4.66. Suppose X 2 H.!1 /. Let Z be the transitive closure of X . Then
Q3 .X / is the set of all Y Z such that the following hold.
(1) There exists a transitive inner model M of ZFC such that:
a) Ord M;
b) X 2 M;
c) for some < !1 ; X 2 V , 2 M and is a Woodin cardinal in M;
d) Y 2 M.
(2) Suppose that M is a transitive inner model of ZFC such that:
a) Ord M;
b) X 2 M;
c) for some < !1 ; X 2 V , 2 M and is a Woodin cardinal in M.
Then Y 2 M.
t
u
The operation Q3 .X / has its origins in descriptive set theory. The exact denition
is given, and the basic theory is developed in work of Kechris, Martin and Solovay
.Kechris, Martin, and Solovay 1983/. The context for the work is
12 -Determinacy.
Q3 .!/ is the set of all subsets of ! which are recursive in some real which is a 12
singleton in a countable ordinal .Kechris, Martin, and Solovay 1983/.
173
174
(1.1) M 2 N ;
hH.!2 /; 2; INS iN 8y a
where a D j.a/.
Assume toward a contradiction that for some b 2 H.!2 /N ,
hH.!2 /; 2; INS iN : a ; b:
Choose a transitive set Y 1 N such that
(2.1) Y 2 N ,
(2.2) jY j D !1N in N ,
(2.3) h.M0 ; I0 /; a0 i; a ; b Y ,
(2.4) !1N Y ,
(2.5) hH.!2 /; 2; INS iY : a ; b.
The structure .M0 ; I0 / is an iterate of .M0 ; I0 / and the iteration is uniquely determined by a . Therefore M0 2 Y .
Let Z !1N be such that Z 2 N and Y 2 LZ.
Let N1 be a transitive inner model of ZFC such that N1 contains the ordinals,
Z 2 N1 ,
Q3 .Z/ D P .!1N / \ N1 ;
and such that there exists < !1 such that is a Woodin cardinal in N1 . We may
suppose that
N1 D LS
for some S .
Y
N
D INS
\ Y and so it follows that
Note that INS
hH.!2 /; 2; INS iN1 : a ; b
and that
N1
I0 D INS
\ M0 :
Let
175
P0 D Coll.!1 ; </N1
and let
Q D P0 P
be an iteration dened in N1 such that P is ccc in N1P0 ,
N1Q MA C :CH
and such that Q has cardinality in N1 .
Let g Q be N1 -generic with g 2 V . Therefore
hH.!2 /; 2; INS iN1 g : a ; b:
N1 g
and let < !1 be strongly inaccessible in N1 g. Since
Let I D INS
N1 D LS
for some S , it follows that !1 is a limit of indiscernibles of N1 g, and so exists.
Let
M D V \ N1 g:
By Theorem 4.63, the ideal I is precipitous in N1 g and N1 g contains the ordinals.
Therefore by Lemma 3.8 and Lemma 3.10, .M ; I / is iterable.
Thus h.M ; I /; a i 2 Pmax and
hH.!2 /; 2; INS iM :8y a :
However M0 2 M and I0 D I \ M . Therefore
h.M ; I /; a i < h.M0 ; I0 /; a0 i
and this is a contradiction.
t
u
In the case that the counterexample is actually in L.R/; i. e. is constructible from a
real, then a much stronger statement can be made.
Theorem 4.68. Suppose that there exists a Woodin cardinal with a measurable cardinal above.
Suppose that .x/ is a 1 formula in the language for the structure
hH.!2 /; 2; INS i:
Suppose that A !1 , x 2 R, and that A 2 Lx.
Suppose that M is a countable transitive model,
M ZFC C There is a Woodin cardinal with a measurable above;
x 2 M and that
M 1 V:
3
Then
hH.!2 /; 2; INS i A
if and only if
where AM D A \ !1M .
176
hH.!2 /; 2; INS i
and that for some partial order P ,
hH.!2 /; 2; INS iV
Then
:
Pmax
:
Proof. The theorem follows by a simple absoluteness argument, noting that from the
hypothesis that for every partial order P ,
1
VP
2 -Determinacy;
it follows that for every partial order P ,
V 1 V P I
4
1
i. e. that 4 statements with parameters from V are absolute between V and V P .
It sufces to prove that if is a 2 sentence in the language for the structure
hH.!2 /; INS ; 2i
and if
Pmax
;
:
j W .M; I / ! .M ; I /
a) .P .!1 //M N ;
b) !1N D !1M ;
c) Q3 .S / N , for each S 2 N such that S !1N ;
d) If S !1N , S 2 M and if S I then S is a stationary set in N .
Then
177
::
Then by the iteration lemmas, (1.1) must fail in V P . But this is a contradiction since
(1.1) is expressible as a 14 statement about t where t 2 R codes the condition
h.M; I /; ai.
t
u
A stronger form of the absoluteness theorem is actually true. This arises from
expanding the structure hH.!2 /; 2; INS i by adding predicates for each set of reals in
L.R/. The expanded structure
hH.!2 /; 2 ; INS ; X I X 2 L.R/; X Ri
is a natural one to consider in the presence of suitable large cardinals. In this case each
set X R with X 2 L.R/ has a canonical interpretation in any generic extension of V
just as borel sets have canonical interpretations. We shall need the following corollary
to the results in Section 2.2.
Lemma 4.70. Suppose X R and let Y R be the set of reals which code elements
of the rst order diagram of the structure
hH.!1 /; 2; X i:
Suppose S and T are trees on ! such that
(1) S and T are weakly homogeneous,
(2) X D pS and Y D pT .
Suppose P 2 V and G P is V -generic. Let XG D pS and let YG D pT , each
computed in V G. Then in V G,
hH.!1 /V ; 2; X i hH.!1 /V G ; 2; XG i:
Theorem 4.71. Assume AD holds in L.R/. Suppose
language for the structure
hH.!2 /; X; 2; INS i
9x .x/:
Then there is a condition h.M; I /; ai 2 Pmax and a set b 2 H.!2 /M such that if
h.M ; I /; a i h.M; I /; ai
and if .M ; I / is X -iterable then
hH.!2 /M ; X \ M ; 2; I i
t
u
b
where b D j.b/ and j W .M; I / ! .M ; I / is the iteration such that j.a/ D a .
178
t
u
Pmax
:
Proof. We sketch the argument which is really just a minor modication of the proof
of Theorem 4.65.
Let .x; y/ be a 0 formula such that D 8x9y: .x; y/ (up to logical equivalence). Clearly we may assume that contains only 1 unary predicate from those
additional predicates for the sets of reals we have added to the structure
hH.!2 /; 2; INS i:
Let X be the corresponding set of reals. Assume toward a contradiction that
Pmax
::
hH.!2 /M ; 2; I ; X \ M i 8y b
where b D j.b/ and j W .M; I / ! .M ; I / is the iteration such that
j.a/ D a .
By Theorem 4.41, we may assume by rening h.M; I /; ai if necessary, that .M; I / is
X -iterable and that X \ M 2 M.
By Lemma 4.36, there is an iteration
j W .M; I / ! .M ; I /
such that:
179
(2.1) j.!1M / D !1 ;
(2.2) J \ M D I ;
(2.3) X \ M D j.X \ M/.
Let B D j.b/ and let A D j.a/. The sentence holds in V and so there exists a
set D 2 H.!2 / such that
hH.!2 /; 2; INS ; X i : B; D:
Let be the least Woodin cardinal and let T be a C -weakly homogeneous tree
such that X D pT .
Let
Q D Coll.!1 ; </ P
be an iteration dened in V such that P is ccc in V Coll.!1 ;</ ,
V Q MA C :CH;
and such that Q has cardinality in V .
Let G Q be V -generic.
By Theorem 4.63, the nonstationary ideal on !1 is precipitous in V G. Since
.INS /V D .INS /V G \ V
and since
where XG D pT V G .
Let be the least strongly inaccessible cardinal above . For trivial reasons (there
are sets which are not
11 and which are C -weakly homogeneously Suslin) exists
C
and further X is -weakly homogeneously Suslin. Let IG be the nonstationary ideal
on !1 (computed in V G).
Let g Coll.!; / be V G generic. Thus by Lemma 3.8 and Lemma 3.10,
.V G; IG / is iterable in V Gg. Let X.G;g/ D pT V Gg . It follows that
.V G; IG / is X.G;g/ -iterable in V Gg.
Therefore in V Gg,
hH.!2 /; 2; INS ; XG iV G : B; D;
and so in V Gg;
(3.1) h.V G; IG /; Ai 2 Pmax ,
(3.2) h.V G; IG /; Ai < h.M; I /; ai,
(3.3) h.V G; IG /; Ai is X.G;g/ iterable,
(3.4) hH.!2 /; 2; INS ; XG iV G :8y j.b/, where
j W .M; I / ! .M ; I /
is the iteration such that j.a/ D A.
180
Pmax
:
Proof. Let be the least ordinal which is a limit of Woodin cardinals. Let be a
measurable cardinal above . By the results of .Koellner and Woodin 2010/ if X R
and X 2 L.R/ then X is < weakly homogeneously Suslin. Therefore the theorem
now follows from Theorem 4.72.
t
u
The strengthened absoluteness theorem has in some sense a converse. We rst
prove a technical lemma which, while not strictly necessary for the proof of Theorem 4.76, does simplify things a little.
Lemma 4.74. Assume that for some countable elementary substructure,
X0 H.!2 /;
MX0 is iterable where MX0 is the transitive collapse of X0 .
Suppose that h.M0 ; I0 /; a0 i 2 Pmax and that h.M1 ; I1 /; a1 i 2 Pmax are conditions
such that there exist iterations
j0 W .M0 ; I0 / ! .M0 ; I0 /
and
satisfying
(i) j0 .a0 / D j1 .a1 /,
(ii) I0 D M0 \ INS ,
(iii) I1 D M1 \ INS .
Then h.M0 ; I0 /; a0 i and h.M1 ; I1 /; a1 i are compatible in Pmax .
181
Proof. Suppose
X H.!2 /
is a countable elementary substructure such that
h.M0 ; I0 /; a0 i; h.M1 ; I1 /; a1 i; j0 ; j1 X:
Let MX be the transitive collapse of X . Let .j0X ; j1X / be the image of .j0 ; j1 / and let
IX be the image of INS \ X under the collapsing map.
By Theorem 3.16, .MX ; IX / is iterable.
By Theorem 3.22, for every x 2 R, x exists. Therefore by Lemma 4.38 there
exists h.M; I /; ai 2 Pmax such that
MX 2 H.!1 /M :
By Lemma 4.36, there exists in M an iteration
j W .MX ; IX / ! .MX ; IX /
such that IX D I \ MX .
Let b D j.j0X .a0 //. The iteration j.j0X / witnesses that
h.M; I /; bi < h.M0 ; I0 /; a0 i;
and the iteration
j.j1X /
witnesses that
h.M; I /; bi < h.M1 ; I1 /; a1 i:
t
u
Remark 4.75. Lemma 4.74 can be proved under a variety of assumptions. For example it follows from the assumption that there is a Woodin cardinal with a measurable
cardinal above.
t
u
Theorem 4.76. Assume ADL.R/ . Suppose that for each 2 sentence in the language
for the structure
hH.!2 /; 2; INS ; X I X 2 L.R/; X Ri
if
Pmax
then
hH.!2 /; 2; INS ; X I X 2 L.R/; X Ri :
Then L.P .!1 // D L.R/G for some G Pmax which is L.R/-generic.
Proof. It sufces to show that for all A 2 P .!1 / n L.R/, A is L.R/-generic for Pmax .
From this it follows by Theorem 4.60 that L.P .!1 // is a Pmax generic extension of
L.R/.
Suppose A !1 and A L.R/. Let
FA Pmax
be the set of all conditions h.M; I /; ai 2 Pmax such that there is an iteration
j W .M; I / ! .M ; I /
such that j.a/ D A and INS \ M D I .
182
t
u
Remark 4.77. We shall show in Chapter 10 that it is essential for Theorem 4.76 that
INS be a predicate of the structure even if one assumes in addition Martins Maximum
for partial orders of cardinality c, cf. Theorem 10.70. We shall also show that conally many sets in P .R/ \ L.R/ must also be added, (Theorem 10.90).
t
u
If one assumes in addition that R# exists then Theorem 4.76 can be reformulated
so as to refer only to a structure of countable signature; i. e. the structure of a countable
language.
For each n 2 ! let Un be a set which 1 denable in the structure
hL.R/; hi W i < ni; 2i
where hi W i < ni is an increasing sequence of Silver indiscernibles of L.R/, and such
that Un is universal. Clearly the denition of the set Un depends only on the choice of
the universal formula (and not on the choice of hi W i < ni).
183
Theorem 4.78. Assume ADL.R/ and that R# exists. Suppose that for each 2 sentence in the language for the structure
hH.!2 /; 2; INS ; Un I n < !i
if
Pmax
then
hH.!2 /; 2; INS ; Un I n < !i :
Then L.P .!1 // D L.R/G for some G Pmax which is L.R/-generic.
t
u
;
hH.!2 /; 2ihM;E i :
(2) V 12 -Determinacy.
t
u
Remark 4.80. (1) It follows from Theorem 4.67 that Theorem 4.79(2) cannot be
1
signicantly improved; i. e.
2 -Determinacy can fail in V .
(2) Suppose that for each 2 sentence if there exists a partial order P such that
hH.!2 /; 2iV
;
then
hH.!2 /; 2i :
Assume there is an inaccessible cardinal. Must ADL.R/ hold?
t
u
Chapter 5
Applications
5.1
185
t
u
Lemma 5.5. Suppose that for each set X R with X 2 L.R/, there is a condition
h.M; I /; ai 2 Pmax such that
186
5 Applications
(i) X \ M 2 M,
(ii) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(iii) .M; I / is X -iterable.
Suppose G Pmax is L.R/-generic. Then
L.R/G AC :
Proof. We work in L.R/G.
Necessarily,
P .!1 / L.R/G:
Suppose hSi W i < !i and hTi W i < !i are sequences of pairwise disjoint subsets
of !1 . Suppose the Si are stationary and suppose that !1 D [Ti j i < !.
Let h.M; I /; ai 2 G be such that hSi W i < !i; hTi W i < !i 2 M where
j W .M; I / ! .M ; I /
is the iteration such that j.a/ D AG . Let hsi W i < !i; hti W i < !i in M be such that
j..hsi W i < !i; hti W i < !i// D .hSi W i < !i; hTi W i < !i/:
Thus in M, hsi W i < !i and hti W i < !i are sequences of pairwise disjoint subsets of
!1M , the si are not in I , and
!1M D [ti j i < !:
Let D be the set of conditions h.N ; J /; bi < h.M; I /; ai such that in N there exist
< !2N and a continuous increasing function F W !1N ! with conal range such
that
F .tiN / sQiN
for each i < ! where tiN D k.ti /, siN D k.si / and k is the embedding of the iteration
of .M; I / which sends a to b. For each i < !, sQiN denotes the set AQ as computed in
N where A D siN .
It sufces to show that D is dense below h.M; I /; ai.
We show something slightly stronger. Suppose
h.N ; J /; bi < h.N0 ; J0 /; b0 i < h.M; I /; ai:
Then for some c 2 N , h.N ; J /; ci 2 D and
h.N ; J /; ci < h.N0 ; J0 /; b0 i < h.M; I /; ai:
si0
Let be the image of si under the iteration of .M; I / which sends a to b0 and let
ti0 be the image of ti under this iteration.
Let x 2 N be a real which codes N0 .
Working in N we dene an iteration of .N0 ; J0 / of length !1N . Let C be the set of
indiscernibles of Lx less than !1N . Let D C be the set of 2 C such that C \
has ordertype . Thus D is a closed unbounded subset of C . Let
h.N ; J /; G ; j; W < !1N i
be an iteration of .N0 ; J0 / in N such that
187
(1.1) for all 2 D and for all < , j0; .si0 / 2 G if 2 j0; .ti0 / where is the
th element of C above ,
(1.2) J! N D J \ N! N .
1
The iteration is easily constructed in N , the point is that the requirements given by
(1.1) and (1.2) do not interfere. The other useful observation is that if 2 C and if
k W .N0 ; J0 / ! .N0 ; J0 / is any iteration of length then D k.!1N0 /.
Let be the .!1N C !1N /th indiscernible of Lx. Let F be the function
F W !1N !
given by F ./ is the th indiscernible of Lx where D !1N C . Thus
(2.1) 2 N ,
(2.2) < !2N ,
(2.3) F 2 N ,
(2.4) F W !1N ! is continuous and strictly increasing.
Let siN D j0;! N .si0 / and let tiN D j0;! N .ti0 /. Let
1
c D j0;! N .b0 /:
1
Thus h.N ; J /; ci 2 Pmax and h.N ; J /; ci < h.N0 ; J0 /; b0 i. By the denition of the
iteration it follows that in N ,
F tiN sQiN
and so h.N ; J /; ci 2 D.
t
u
Lemma 5.5 yields, immediately, two corollaries, noting that by Lemma 4.40, the
hypothesis of Lemma 5.5 is a consequence of ADL.R/ .
Corollary 5.6. Assume ADL.R/ . Then L.R/Pmax ZFC.
t
u
Corollary 5.7. Assume ./ holds. Then AC holds in hH.!2 /; 2i.
t
u
5.2
We sketch a proof of the following lemma which we shall use to prove that Martins
Maximum implies AC .
Lemma 5.8. Assume Martins Maximum. Suppose that S !1 is stationary. Then
j 2 SQ and cof ./ D !
is stationary in !2 .
188
5 Applications
189
(4.1) 0 < 1 .
(4.2) Suppose 2 t2 and let
X D f .jk/ j k < !:
Then
a) X \ !1 D 0 ,
b) ordertype.X / D 1 .
To see this suppose that g P is V -generic with .s2 ; t2 / 2 G. Then in V g, g 2 t2 .
Let
Xg D f .g jk/ j k < !:
Thus Xg 2 Cg . By absoluteness it follows from (4.2(a)) and (4.2(b)) that
Xg \ !1 D 0 and that ordertype.Xg / D 1 . Therefore Xg 2 Zg S0 ; S1 and so
Cg \ Zg S0 ; S1 ;:
To nd .0 ; 1 / and .s2 ; t2 / we associate to each pair
.0 ; 1 / 2 S0 S1
with 0 < 1 a game, G .0 ; 1 /, as follows: Player I plays to construct a sequence
h.i ; iI / W i < !i
of pairs such that .i ; iI / 2 !2 1 . Player II plays to construct a sequence
h.bi ; ni ; iII / W i < !i
of triples .bi ; ni ; iII / 2 t1 ! 1 . Let hi W i < !i be the sequence such that for all
i < !, 2iC1 D iI and 2i D iII .
The requirements are as follows: For each i < j < !,
(5.1) bi biC1 and dom.bi / D i ,
(5.2) if s1 bi then biC1 D bi _ for some > i ,
(5.3) f .b2iC1 / < !1 if and only if i < 0 ,
(5.4) ni is odd and
f .bi / D f .bni /;
190
5 Applications
191
192
5 Applications
such that:
(9.1) D sup.i / j i < !.
(9.2) .i / j i < ! .
(9.3) C \ is conal in .
(9.4) Suppose X 2 P!1 . / and that F X <! X . Then ordertype.X / 2 S .
Q
Thus 2 C \ S.
t
u
From Lemma 5.8 and the results of .Foreman, Magidor, and Shelah 1988/ we
obtain the following corollary.
Theorem 5.9. Assume Martins Maximum. Then
H.!2 / AC :
Proof. The relevant result of .Foreman, Magidor, and Shelah 1988/ is the following.
Assume Martins Maximum. Suppose hTi W i < !i are pairwise disjoint subsets of !1
and that !1 D [Ti j i < !. Suppose hSi W i < !i are pairwise disjoint stationary
subsets of !2 such that for all i < !, Si C! where
C! D < !2 j cof D !:
Then there exists an ordinal < !2 and a continuous (strictly) increasing function
F W !1 ! with conal range such that
F Ti Si
for each i < !.
This together with the previous lemma yields that Martins Maximum implies AC .
t
u
Thus:
Theorem 5.10. Assume Martins Maximum. Then
L.P .!1 // ZFC:
5.3
The sentence
AC
t
u
AC
193
Further this sentence can be used in place of MA!1 in dening Pmax , an alternate approach which will be useful in dening some of the Pmax variations, cf. Denition 6.91.
We will also consider, in Section 7.2, versions of this sentence relativized to a normal
ideal on !1 .
Denition 5.11. AC : Suppose S !1 and T !1 are stationary, co-stationary, sets.
Then there exist < !2 , a bijection
W !1 ! ;
and a closed unbounded set C !1 such that
< !1 j ordertype./ 2 T \ C D S \ C:
t
u
Thus AC asserts that for each pair .S; T / of stationary, co-stationary, subsets of
!1 , there exists an ordinal < !2 such that
S INS D 2 j.T / in V B
where
B D RO.P .!1 /=INS /
and
j W V ! .M; E/ V B
is the corresponding generic elementary embedding. This implies (in ZF) that the
boolean algebra
P .!1 /=INS
can be wellordered (in length at most !2 ).
Lemma 5.12 (ZF + DC). Assume
AC
holds in
hH.!2 /; 2i:
Suppose hS W < !1 i is a partition of !1 into !1 many stationary sets. Then there is
a surjection
W !2 ! P .!1 /
which is 1 denable in
hH.!2 /; INS ; 2i
from hS W < !1 i.
Proof. For each set A !1 let
SA D [SC1 j 2 A
and let SA D S0 if A D ;.
The key point is that if A !1 , B !1 , and if A B then
SA M SB INS :
Dene
W !2 ! P .!1 /
194
5 Applications
t
u
The proof that Martins Maximum implies AC is actually much simpler then the
proof we have given that Martins Maximum implies AC . The reason is that our approach to proving AC from Martins Maximum was through Lemma 5.8 which established quite a bit more than is necessary. Here we take a more direct approach which
only requires a special case of the reection principle, SRP, an observation due independently to P. Larson. The special case is SRP for subsets of P!1 .!2 /, which can be
proved from just Martins Maximum.c/. This special case is discussed in Section 9.3.
Theorem 5.13. Assume Martins Maximum.c/. Then
H.!2 /
AC :
AC
195
Let
X D j0;2 !2V ;
the image of !2V under j0;2 . Thus
(2.1) ordertype.X / < j0;2 .!1V /,
(2.2) j0;2 .H /X <! X ,
(2.3) X \ j0;2 .!1V / 2 j0;2 .S /,
(2.4) X \ j0;2 .!1V / 2 j0;2 .S0 / if and only if ordertype.X / 2 j0;2 .T0 /.
By absoluteness such a set X must exist in M2 and so by the elementarity of j0;2 ,
it follows that in V
X 2 ZS j H X <! X ;:
Thus ZS is stationary in P!1 .!2 /.
By Martins Maximum.c/, there exists 0 < !2 such that !1 < 0 and such that
Z \ P!1 .0 /
is closed unbounded in P!1 .0 /. Let
0 W !1 ! 0
be a surjection. It follows that .0 ; 0 / witnesses that
AC
t
u
Larson has also noted that the proof of Theorem 5.13 easily adapts to show that
Martins Maximum.c/ implies AC . We note that Lemma 5.8 cannot be proved from
just Martins Maximum.c/. Therefore, for the proof that we have given that Martins
Maximum implies AC , Martins Maximum.c/ does not sufce. Finally Larson has
proved versions of Lemma 5.8 showing for example that Martins Maximum.c/ implies
that for each stationary set S !1 , SQ is stationary in !2 and that
SQ \ < !2 j cof./ D ! ;:
The sentence,
boolean algebra
AC ,
is (trivially) generated by the term for j.T /. This fact combined with Theorem 3.51
yields the following lemma as an immediate corollary.
Lemma 5.14. Suppose that
AC
holds. Then
2@0 D 2@1 D @2 :
t
u
The next lemma shows that AC serves successfully in place of MA!1 in the denition of Pmax . This lemma is really just a special case of the claim given at the beginning
of the proof of Theorem 3.51.
196
5 Applications
AC :
a !1M ;
and
M a is a stationary, co-stationary, set in !1 :
Suppose
j1 W M ! M1
and
j2 W M ! M2
are semi-iterations of M such that M1 is transitive, M2 is transitive and such that
j1 .a/ D j2 .a/:
Then M1 D M2 and j1 D j2 .
Proof. Fix a and suppose that
hM ; G ; j; j < i
is a semi-iteration of M such that M is transitive for all .
We prove that G0 , M1 and
j0;1 W M ! M1
are uniquely specied by j0; .a/ \ !2M .
We note that since G0 is an M-normal ultralter,
G0 D b !1M j b 2 M and !1M 2 j0;1 .b/:
Therefore since
M
AC
!1M n b .INS /M :
AC
197
j1 .!1M /
j2 .!1M /.
t
u
The proof that ./ implies AC is simplied by rst proving the following technical
lemma which isolates the combinatorial essence of the implication.
Lemma 5.16 (ZFC ). Suppose that x 2 R codes
h.M; I /; ai 2 Pmax
and that x # exists. Let C be the set of of the Silver indiscernibles of Lx below !1 and
let C 0 be the limit points of C . Suppose that
s; t .P .!1 //M n I
is such that both !1M0 n s I and !1M0 n t I . Suppose J is a normal, uniform, ideal
on !1 . Then there exists an iteration
j W .M; I / ! .M ; I /
of length !1 such that
and such that for all 2 C 0 ,
if and only if
I D J \ M
2 j.s/
C 2 j.t /
198
5 Applications
Proof. We modify the proof of Lemma 4.36 using the notation from that proof. The
modication is a minor one.
Choose the sequence hAk; W k < !; < !1 i of J positive, pairwise disjoint, sets
such that
[Ak; j k < ! and < !1 C 0 :
Following the proof of Lemma 4.36 construct the iteration
h.M ; I /; G ; j; W < !1 i
of .M; I / to satisfy the additional requirement that for all 2 C 0 ,
2 j0;C1 .s/
if and only if
j0; .t / 2 G
C
where D .
For each 2 C if
h.M ; I /; G ; j; W < i
is any iteration of length then
j0; .!1M / D
and so this additional requirement does not interfere with the original requirements
indicated in the proof of Lemma 4.36. Thus
j0;!1 W .M; I / ! .M!1 ; I!1 /
t
u
is as desired.
Lemma 5.17. Assume ./ holds. Then
AC
holds.
199
if and only if
C 2 j.t /
where for each 2 C , C denotes the least element of C above .
Thus for all 2 C 0 ,
2S
if and only if
C 2 T:
Let be the least Silver indiscernible of Lx0 above !1 and let
W !1 !
be a bijection.
Thus there exists a club D C 0 such that for all 2 D,
ordertype. / D C :
Therefore
< !1 j ordertype. / 2 T \ D D S \ D:
This proves the lemma.
5.4
t
u
We sketch a different presentation of Pmax . This leads to different proofs of the absoluteness theorems. This approach will be useful in proving absoluteness theorems for
some of the variations of Pmax that we shall dene, cf. Theorem 6.85.
Another feature of this approach is that it much easier to show that suitable conditions exist. This is because the generic iterations are based on elementary embeddings
associated to the stationary tower and not to an ideal on !1 . Thus no forcing arguments
are required to produce conditions.
Recall from Section 2.3 the following conventions. Suppose is a Woodin cardinal. Then Q< is the partial order given by the stationary tower restricted to V and
restricted to sets of countable sets.
We let I< denote the associated directed system of ideals. This is dened as
follows. For each < , let
I D a P!1 .V / j a is not stationary in P!1 .V /:
If < there is a natural map
; W I ! I
given by
; .c/ D 2 P!1 .V / j \ V 2 c:
I< is the directed system of ideals, hI ; ; W < < i. A set a 2 V is
I< -positive if a P!1 .[a/ and a is stationary. Thus I< can be naturally identied
with the set of a 2 V such that a P!1 .[ a/ and such that a is not stationary.
Q< is the set of a 2 V such that a is I< -positive.
We generalize the notions of iterability to the current context.
200
5 Applications
t
u
The next lemma, while not strictly necessary for what follows, does simplify some
of the denitions. The lemma justies the identication of an iteration of .M; I/ with
the resulting elementary embedding.
Lemma 5.19. Suppose M is a countable model of ZFC and I 2 M is the directed
system I< as computed in M for some 2 M which is a Woodin cardinal in M .
Suppose
h.M ; I /; G ; j; W < < i
is an iteration of .M; I/ of length . Then for each < the sequence
h.M ; I /; G ; j; W < < i
is uniquely determined by the elementary embedding,
j0; W .M0 ; I0 / ! .M ; I /:
201
t
u
MkC1
D !1
202
5 Applications
If is a limit ordinal then is the length of the iteration, otherwise the length of
the iteration is where C 1 D .
A sequence h.Mk ; Ik / W k < !i is an iterate of h.Mk ; Ik / W k < !i if it occurs in
an iteration of h.Mk ; Ik / W k < !i.
The sequence h.Mk ; Ik / W k < !i is iterable if every iterate is wellfounded.
t
u
The next two lemmas are the generalizations of Lemma 3.8 and Lemma 3.10 to
iterations based on the stationary tower. We omit the proofs which are straightforward
modications of the previous arguments.
Lemma 5.21. Suppose M and M are countable transitive models of ZFC such that
M 2 M :
. Let I D .I< /M
Suppose 2 M , is a Woodin cardinal in M , and MC1 D MC1
and suppose
h.M ; I /; G ; j; W < < i
t
u
203
Lemma 5.23. Suppose is a Woodin cardinal and that is the least strongly inaccessible cardinal above . Suppose
X V
is a countable elementary substructure. Let .M; I/ be the transitive collapse of
.X; I< \ X / and let X be the image of under the collapsing map. Suppose that
A 2 P .R/ \ X and that every set of reals which is projective in A is C -weakly
homogeneously Suslin. Suppose that
P 2 MX
is a partial order and that g P is an M -generic lter with g 2 V . Let
Ig D .I<X /M g :
Then:
(1) hV!C1 \ M g; A \ M g; 2i hV!C1 ; A; 2i;
(2) .M g; Ig / is A-iterable.
Proof. Fix
X V
such that X is countable. Clearly 2 X since is denable in V .
Let .M; I; X / be the image of .V ; I< ; / under the collapsing map.
Suppose A 2 X , A R and for all B R such that B is projective in A, B is
C -weakly homogeneously Suslin.
Let
D B R j B is projective in A:
Since is the least strongly inaccessible above , every set in is weakly homogeneously Suslin. Therefore if G0 is M -generic for a partial order P0 2 M , with
G0 2 V , then A \ M G0 2 M G0 and
hH.!1 /M G0 ; A \ M G0 ; 2i hH.!1 /; A; 2i:
This follows by Lemma 2.29.
Fix a partial order P 2 MX and suppose that
gP
is an M -generic lter with g 2 V . Let
Ig D .I<X /M g :
Thus
hH.!1 /M g ; A \ M g; 2i hH.!1 /; A; 2i;
and so (1) holds.
Suppose 2 X , < and
V ZFC:
204
5 Applications
Let X be the image of under the collapsing map. Let S and T be trees on ! 2
such that if G Coll.!; / then
.pS /V G D AV G
and
.pT /V G D .R n A/V G :
Since 2 X we may suppose that
S; T X:
Let .SX ; TX / be the image of .S; T / under the collapsing map.
Suppose G Coll.!; X / is M g-generic with G 2 V . Therefore by the remarks
above,
A \ M gG 2 M gG
and
hH.!1 /M gG ; A \ M gG; 2i hH.!1 /; A; 2i:
Suppose
However
pTX \ M gG D .R n pSX / \ M gG
and so
205
t
u
Remark 5.24. Lemma 5.23 can be proved with weaker requirements on the set A.
The approximate converse of this strengthened version of Lemma 5.23 is proved as
Lemma 6.59, in Section 6.2.2 where we consider the Pmax -variation, Qmax .
t
u
Denition 5.25. Suppose .M; I/ is a countable iterable structure where I is the directed system of nonstationary ideals, I< , for some 2 M such that is a Woodin
cardinal in M. Suppose
h.M ; I /; G ; j; W < !1 i
is an iteration of .M; I/ of length !1 . The iteration is full if for all < !1 and for all
a 2 M such that a is I -positive, the set
< !1 j j; .a/ 2 G
t
u
is stationary in !1 .
The next lemma accounts in part for the restriction to iterations which are full.
Lemma 5.26. Suppose .M; I/ is a countable iterable structure where I is the directed
system of nonstationary ideals, I< , for some 2 M such that is a Woodin cardinal
in M. Suppose
h.M ; I /; G ; j; W < !1 i
is an iteration of .M; I/ of length !1 .
(1) Suppose D 2 M!1 ,
D j0;!1 .QM
< /;
and that D is dense. Suppose A is stationary in P!1 .M!1 /. Then there exists
B A such that B is stationary in A and such that B b for some b 2 D.
(2) Suppose that the iteration is full. Then for each < !1 and for each a 2 M
such that a is I -positive, j;!1 .a/ is stationary set in V .
Proof. The proofs of (1) and (2) are similar. We rst prove (1). Fix A and D. Since
D 2 M!1 there exists < !1 and D 2 M such that
D D j;!1 .D /:
206
5 Applications
(1.1) X H.!2 /,
(1.2) X \ M!1 2 A,
(1.3) M 2 X ,
(1.4) h.M ; I /; G ; j; W < !1 i 2 X .
Thus A is stationary.
Suppose X 2 A . Let D X \!1 . G is M -generic and so G \j; .D / ;.
Therefore there exist X < and bX 2 MX such that
jX ; .bX / 2 G \ j; .D /:
Clearly X 2 X and bX 2 X . A is stationary and so there exist 0 and b0 2 M0
such that
B D X 2 A j X D 0 and bX D b0
is stationary in P!1 .H.!2 //. Let b D j0 ;!1 .b0 /. Thus b 2 D.
For each X 2 B let
bX D j0 ; .b0 /
where D X \ !1 . Thus for each X 2 B,
bX 2 G \ j; .D /
where again D X \ !1 .
We claim that B b. To prove this it sufces to show that for each X 2 B,
X \ .[b/ 2 b:
Fix X 2 B. Let D X \ !1 and let
D j; C1 .t / j t 2 [bX :
By the properties of the generic elementary embedding associated to the stationary
tower and since bX 2 G ,
2 j; C1 .bX /:
Further is countable in M C1 since
j; C1 .bX / 2 j0; C1 .QM
< /:
Therefore
j C1;!1 . / D j C1;!1 .t / j t 2
and so
j;!1 .t / j t 2 [bX 2 j;!1 .bX /:
However
X \ .[b/ D j;!1 .t / j t 2 [bX
since
b D j0 ;!1 .b0 / D j;!1 .j0 ; .b0 // D j;!1 .bX /:
207
Therefore
X \ [b 2 b:
Thus B b and this proves (1).
We now prove (2). Fix 0 < !1 and a 2 G0 . Let
S D < !1 j j0 ; .a/ 2 G :
Since the iteration is full, S is a stationary subset of !1 . It sufces to prove that for
each function
F W H.!2 /<! ! H.!2 /
there exists X H.!2 / such that
X \ .[j0 ;!1 .a// 2 j0 ;!1 .a/:
Fix F and let X H.!2 / be a countable elementary substructure such that
(2.1) X \ !1 2 S ,
(2.2) M0 2 X ,
(2.3) h.M ; I /; G0 ; j0 ; W 0 < !1 i 2 X .
Let 0 D X \ !1 . As above, since j0 ;0 .a/ 2 G0 ,
j0 ;!1 .t / j t 2 [j0 ;0 .a/ 2 j0 ;!1 .a/:
It follows that
X \ .[j0 ;!1 .a// D j0 ;!1 .t / j t 2 [j0 ;0 .a/:
Therefore
X \ .[j0 ;!1 .a// 2 j0 ;!1 .a/:
This proves (2).
t
u
(T)
Denition 5.27. We dene Pmax
by induction on M \ Ord where M is the countable
(T)
consists of pairs h.M; I/; X i such that
transitive model specied in the condition. Pmax
the following hold.
208
5 Applications
t
u
(T)
The precise relationship between Pmax and Pmax
is given in Theorem 5.39.
(T)
The analysis of Pmax requires a non-interference lemma.
and that
j W .M; I/ ! .M ; I /
209
Lemma 5.28 is really quite general. We state the version for iterable structures
.M; I / where I 2 M is an ideal on !1M . This we have already discussed in a different
context, see Remark 3.61. This lemma is required for the analysis of any variation of
Pmax in which one has dropped all the requirements on the models designed to recover
iterations from only the iterates. The proof of Lemma 5.29 is identical to the proof of
Lemma 5.28.
Lemma 5.29. Suppose M is a countable transitive model of ZFC and that .M; I / is
iterable. Suppose
j W .M; I / ! .M ; I /
is an iteration of .M; I / of length !1 . Then M M .
t
u
We shall also require a boundedness lemma for iterable structures of the form
.M; I/ where
I D .I< /M
and where is a Woodin cardinal of M. This lemma is the obvious generalization of
Lemma 3.15, and the proof is essentially the same.
Lemma 5.30. Suppose M is a countable model of ZFC and I 2 M is the directed
system I< as computed in M for some 2 M which is a Woodin cardinal in M.
Suppose that x 2 R codes M.
(1) Suppose that
j W .M; I/ ! .M ; I /
j W .M; I/ ! .M ; I /
t
u
The boundedness lemma, Lemma 5.30, yields as an immediate corollary the following lemma.
(T)
(T)
Lemma 5.31. Suppose h.M0 ; I0 /; X0 i 2 Pmax
, h.M1 ; I1 /; X1 i 2 Pmax
, and
h.M1 ; I1 /; X1 i < h.M0 ; I0 /; X0 i:
Let
j W .M0 ; I0 / ! .M0 ; I0 /
210
5 Applications
be the sequence such that for each i < !, .Mi ; Ii / is the iterate of .Mi ; Ii / obtained
by combining the iterations given by the h.Mj ; Ij /; Xj i for j > i .
Then:
(1) For each i < !,
,
a) Mi 2 MiC1
MkC1
(2) For each i < ! let Qi be the partial order of Ii -positive sets computed in Mi .
For each a 2 [Qi j i < ! there exists a sequence hgi W i < !i such that
a) a 2 [gi j i < !,
b) for each i < !, gi giC1 and gi is Mi -generic.
(3) The sequence
is iterable.
211
j i < j < !:
and
212
5 Applications
Remark 5.33. Lemma 5.32 has the following consequence which is really the key to
(T)
(cf. Theorem 5.39).
establishing the relationship between Pmax and Pmax
Suppose
h.Mi ; Ii / W i < !i
is as specied in Lemma 5.32. By Lemma 5.32 (1) and by Lemma 4.17, the sequence
hMi W i < !i
is iterable in the sense of Denition 4.15. This was noted in the proof of Lemma 5.32
and the observation is the basis for the reformulation of Pmax given in Section 5.5.
By Lemma 5.32(2), for each i < !,
t
u
(T)
can be carried in a fashion analogous to
Using Lemma 5.32, the analysis of Pmax
(T)
that for Pmax provided Pmax is sufciently nontrivial.
The proof of Lemma 5.37 requires Theorem 5.34 and Theorem 5.35; these are
proved in .Koellner and Woodin 2010/.
t
u
213
Theorem 5.36. Assume AD holds in L.R/. Suppose A R and A 2 L.R/. Then for
each n 2 ! there exist a countable transitive model M and an ordinal 2 M such
that the following hold.
(1) M ZFC.
(2) is the nth Woodin cardinal of M .
(3) A \ M 2 M and
hV!C1 \ M; A \ M; 2i hV!C1 ; A; 2i:
(4) A \ M is C -weakly homogeneously Suslin in M .
Proof. We work in L.R/.
Suppose the theorem fails. Then there exists A 2 L.R/ which is a counterexample
and such that A is 21 in L.R/. Let B R code the rst order diagram of
hV!C1 ; A; 2i:
is 21
Thus B
denable in L.R/. Therefore by the MartinSteel theorem, Theorem 2.3,
there exist (denable) trees S and T in L.R/ such that
B D pS
and
R n B D pT :
Therefore if N L.R/ is any transitive inner model of ZF such that
S; T N
then A \ N 2 N , B \ N 2 N and
hV!C1 \ N; A \ N; 2i hV!C1 ; A; 2i:
We claim that by Theorem 5.34, there exists a transitive inner model N L.R/
and an increasing sequence hi W i n C 1i of countable ordinals such that
(1.1) S; T N ,
(1.2) N ZFC,
(1.3) for each i n C 1, i is a Woodin cardinal in N .
We indicate how to nd N in the case that n D 0, in this case there are to be two
Woodin cardinals in N . We work in L.R/.
Since S; T are denable, S; T HOD.
Let Z0 Ord be such that HOD D LZ0 . Choose x0 such that
.!2 /LZ0 ;x0
is a Woodin cardinal in
0 ;x0
:
HODLZ
Z0
214
5 Applications
Let
0 D .!2 /LZ0 ;x0 :
Choose a 0 such that
and such that
0 ;x0
a 2 HODLZ
Z0
0 ;x0
\ V 0 :
La \ V0 D HODLZ
Z0
0 ;ay0
0 ;ay
D P .0 / \ HODLZ
:
P .0 / \ HODLZ
a;Z0
a;Z0
0 ;a;y0
:
HODLZ
Z0 ;a
Let
1 D .!2 /LZ0 ;a;y0
and let
0 ;a;y0
:
N D HODLZ
Z0 ;a
pT D pT :
Thus
215
pS \ N g D .R n pT / \ N g;
g 2 H.!1 /
is a lter which is M-generic for a partial order in M . Let Ig D .I< /Mg .
(1) A \ Mg 2 Mg.
(2) hH.!1 /Mg ; A \ Mgi hH.!1 /; Ai.
(3) .Mg; Ig / is A-iterable.
(4) is the nth Woodin cardinal of M.
(T)
Further the set of such conditions is dense in Pmax
.
216
5 Applications
(1.1) N ZFC.
(1.2) is the nth Woodin cardinal in N .
(1.3) B \ N 2 N and
hV!C1 \ N; B \ N; 2i hV!C1 ; B; 2i:
(1.4) B \ N is C -weakly homogeneously Suslin in N .
By Lemma 5.23 (applied in N ) there exists a condition
(T)
h.M; I/; ;i 2 .Pmax
/N
such that the following holds where M be the Woodin cardinal of M associated to I.
Suppose that M g is a generic extension of M for a partial order in MM such that
g 2 N . Let
Ig D .I<M /Mg :
Then
(2.1) A \ Mg 2 Mg,
(2.2) hV!C1 \ Mg; A \ Mg; 2i hV!C1 \ N; A \ N; 2i,
(2.3) .Mg; Ig / is A -iterable in N ,
(2.4) M is the nth Woodin cardinal of M.
Since
hV!C1 \ M; A \ M; 2i hV!C1 ; M; 2i;
it follows that h.M0 ; I0 /; X0 i 2 H.!1 /M .
Let
j W .M0 ; I0 / ! .M0 ; I0 /
be an iteration of .M0 ; I0 / such that j 2 M and such that j is full in M. Let
Y D .h.M0 ; I0 /; X0 i; j / [ j.X0 /.
(T)
Thus in N ; h.M; I/; Y i 2 Pmax
and
h.M; I/; Y i < h.M0 ; I0 /; X0 i:
Finally
hV!C1 \ N; B \ N; 2i hV!C1 ; B; 2i;
and so for any choice of g in V , (2.1)(2.3) hold in V .
(T)
Similarly in V ; h.M; I/; Y i 2 Pmax
and
h.M; I/; Y i < h.M0 ; I0 /; X0 i:
Thus h.M; I/; Y i is as required.
t
u
217
(T)
The analysis of Pmax
is now straightforward, following that of Pmax . In many ways
(T)
the analysis of Pmax is easier than that of Pmax . One reason is that no local forcing
arguments are required.
Assume ADL.R/ and suppose
(T)
G Pmax
L.R/G ZFC;
.P .!1 //
L.R/G
t
u
(T)
We leave the proof of this to the reader and simply prove that Pmax
is equivalent in
L.R/ to the iteration
Pmax .!2<!2 /:
(T)
Theorem 5.39. Assume AD holds in L.R/. Suppose G Pmax
is L.R/-generic. Then
L.R/G D L.R/gh
where g Pmax is L.R/-generic and
h .!2<!2 /L.R/g
is L.R/g-generic.
Proof. Suppose A !1 , A 2 L.R/G and
!1 D !1LA :
We prove that A is L.R/-generic for Pmax . This will prove that
.L.P .!1 ///L.R/G
(T)
is a Pmax generic extension of L.R/ noting that Pmax
is !-closed in L.R/.
Let FA be the set of h.N ; J /; bi 2 Pmax such that there exists an iteration
k W .N ; J / ! .N ; J /
with k.b/ D A and
J D INS \ N :
218
5 Applications
Fix D Pmax such that D is open, dense in Pmax and such that D 2 L.R/. Assume
toward a contradiction that
FA \ D D ;:
By Theorem 5.38, there exist h.M0 ; I0 /; X0 i 2 G and a0 2 M0 such that
!1M0 D !1La0
and such that j.a0 / D A where
j0 W .M0 ; I0 / ! .M0 ; I0 /
is the iteration of .M0 ; I0 / given by G.
We work in L.R/ and assume that
h.M0 ; I0 /; X0 i FA \ D D ;:
Let h.N0 ; J0 /; b0 i 2 Pmax be such that
M0 2 .H.!1 //N0
and such that
J0 D .INS /N0 :
Let
j1 W .M0 ; I0 / ! .M1 ; I1 /
be an iteration such that j1 2 N0 , j.!1M0 / D !1N0 , and such that j1 is full in N0 . Let
a1 D j1 .a0 /. Thus
!1La1 D !1N0
and so h.N0 ; J0 /; a1 i 2 Pmax .
Since D is open, dense in Pmax , there exists h.N1 ; J1 /; b1 i 2 D such that
h.N1 ; J1 /; b1 i < h.N0 ; J0 /; a1 i
and such that
J1 D .INS /N1 :
(T)
Let h.M2 ; I2 /; X2 i 2 Pmax
be such that
N1 2 H.!1 /M2 :
Let
be the iteration such that k0 .a1 / D b1 . By Lemma 4.36 there exists an iteration
k1 W .N1 ; J1 / ! .N2 ; J2 /
such that k1 2 M2 and such that
.INS /N2 D J2 D N2 \ .INS /M2 :
Let a2 D k1 .b1 /.
Thus k1 .k0 .j1 // is an iteration
k1 .k0 .j1 // W .M0 ; I0 / ! .k1 .k0 .j1 //.M0 /; k1 .k0 .j1 //.I0 //
219
and
h.M2 ; I2 /; X i < h.M0 ; I0 /; X0 i
where
X D k1 .k0 .j1 //.X0 / [ .h.M0 ; I0 /; X0 i; k1 .k0 .j1 ///:
By genericity we may assume
h.M2 ; I2 /; X i 2 G:
Let
j2 W .M2 ; I2 / ! .M2 ; I2 /
be the iteration given by G. Thus
j2 .k1 / W .N1 ; J1 / ! .N1 ; J1 /
is an iteration such that
M
L.R/G
.INS /N1 D J1 D N \ INS 2 D N \ INS
:
Further
A D j0 .a0 / D j2 .k1 .k0 .j1 .a0 //// D j2 .k1 .b1 // D j2 .a2 /
and
j2 .k1 .b1 // D j2 .k1 /.j2 .b1 // D j2 .k1 /.b1 /:
Therefore h.N1 ; J1 /; b1 i 2 FA which contradicts the choice of D and A.
Therefore
.L.P .!1 ///L.R/G
is a Pmax generic extension of L.R/.
Fix g Pmax such that g is L.R/-generic and
.L.P .!1 ///L.R/G D L.R/g:
Let P be the following partial order dened in L.R/g.
(T)
P is the set of pairs .h.M; I/; X i; j / such that h.M; I/; X i 2 Pmax
and such that
O I/
O
j W .M; I/ ! .M;
is an iteration which is full in L.R/g. Suppose .h.M0 ; I0 /; X0 i; j0 / 2 P and that
.h.M1 ; I1 /; X1 i; j1 / 2 P . Then
.h.M1 ; I1 /; X1 i; j1 / < .h.M0 ; I0 /; X0 i; j0 /
if
.h.M0 ; I0 /; X0 i; j0 / 2 j1 .X1 /:
The two relevant properties of P are the following.
220
5 Applications
L.R/G0 D L.R/gh0 :
By the denability of forcing it follows that there exists h P such that h is
L.R/g-generic and such that
L.R/G D L.R/gh:
t
u
5.5 Pmax
221
5.5 Pmax
(T)
We dene a second reformulation of Pmax . This version is quite closely related to Pmax
and it involves a reformulation of the sentence AC .
Denition 5.40. AC
: Suppose that hS W < !1 i and hT W < !1 i are each
sequences of stationary, co-stationary sets. Then there exists a sequence h W < !1 i
of ordinals less than !2 such that for each < !1 there exists a bijection
W !1 ! ;
and a closed unbounded set C !1 such that
< !1 j ordertype./ 2 T \ C D S \ C:
t
u
AC
M
AC :
if and only if
The reason for introducing
iterable sequence and that
AC
[Mi j i < !
AC :
AC :
AC .
Denition 5.41. Pmax
is the set of pairs .hMk W k < !i; a/ such that the following
hold.
(3) Mk 2 MkC1 ; !1
MkC1
D !1
AC .
222
5 Applications
b) M0 jX j D !1 ,
M0
.
c) for all S; T 2 X , if S T then S \ T 2 INS
is analogous to Pmax . A condition
The ordering on Pmax
(3) .INS /MkC1 \ Mk D .INS /N1 \ Mk for all k < !.
Remark 5.42.
t
u
(1) One can strengthen (4) by requiring that for all k < !,
Mk
MkC1
D Mk \ INS
:
INS
The proof of Lemma 5.15 easily adapts to prove the following lemma which is the
analog of Lemma 4.35.
. Suppose that
Lemma 5.43. Suppose that .hMk W k < !i; a/ 2 Pmax
5.5 Pmax
223
and such that for all k < !, Z INS kC1 and .!1M0 n Z/ INS kC1 .
Therefore arguing as in the proof of Lemma 5.15, if j1 .Z/ D j2 .Z/ then
hMk1 W k < !i D hMk2 W k < !i
and j1 D j2 .
The sequence hMk1 W k < !i is iterable and so it follows that for all b !1M0 , if
b 2 [Mk j k 2 !
then
b # 2 [Mk j k 2 !:
Therefore .x; a/# 2 [Mk j k 2 !. Thus since j1 .a/ D j2 .a/ it follows that
t
u
j1 .Z/ D j2 .Z/ noting that necessarily j1 .!1M0 / D j2 .!1M0 /.
The basic iteration lemma required for the analysis of Pmax
is a modication of
Lemma 4.37. The proof is a minor variation of the proof of Lemma 4.36.
. Suppose that
Lemma 5.44 (ZFC ). Suppose that .hMk W k < !i; a/ 2 Pmax
hS W < !1 i
is a sequence of pairwise disjoint stationary subsets of !1 . Then there is an iteration
j W hMk W k < !i ! hMk W k < !i
such that for all S !1 , if
M
then S n S 2 INS
t
u
As a corollary to Lemma 5.44 we obtain the following iteration lemma. It is for the
proof of this lemma that the requirement (7) in the denition of Pmax
is essential.
Lemma 5.45. Suppose that
.hMk W k < !i; a/ 2 Pmax
;
.hNk W k < !i; b/ 2 Pmax ;
and that
N1
INS
\ .[Mk j k < !/ D [INS k j k < !:
224
5 Applications
Proof. Since
.hNk W k < !i; b/ 2 Pmax
there exists a sequence hS W < !1N0 i 2 N0 such that for all < < !1N0 ,
S !1N0 ;
N1
S INS
;
N0
.
and such that S \ S 2 INS
With this sequence the lemma follows by Lemma 5.44.
t
u
is a routine generalization of
Using the iteration lemmas the basic analysis of Pmax
the analysis of Pmax provided suitable iterable structures exist.
t
u
. It is (notationally)
We prove a very general existence lemma for conditions in Pmax
(T)
convenient to refer to Pmax in the statements of the two preliminary lemmas that we
require; note that the assumption
(T)
h.M0 ; I0 /; ;i 2 Pmax
5.5 Pmax
225
and that
X0 V
is a countable elementary substructure such that
M0 D MX0
where MX0 is the transitive collapse of X0 . Suppose that a0 !1M0 is a set in M0
such that
.!1 /La0 D .!1 /M0
Then there exists
.hMO k W k < !i; a/
O 2 Pmax
226
5 Applications
such that:
(1) there exists a countable iteration
j W .M0 ; I0 / ! .M0 ; I0 /
O
such that j.a0 / D a.
O
h.Mk ; Ik / W k < !i
of elements of M together with iterations
jk W .Mk ; Ik / ! .Mk ; Ik /
and elements .Fk.S/ ; Fk.T / / 2 Mk as follows. We simultaneously dene an increasing
sequence hXk W k < !i of countable elementary substructures of V such that for each
k < !, Mk is the transitive collapse of Xk .
.M0 ; I0 / and X0 are as given.
Suppose that Xk and .Mk ; Ik / have been dened. We dene .Fk.S/ ; Fk.T / /, jk ,
XkC1 and .MkC1 ; IkC1 /.
Let k 2 Mk be the Woodin cardinal of Mk corresponding to Ik and let
Qk D .Q<k /Mk ;
be the associated stationary tower. Let hk W < !1 i be the increasing enumeration
of the ordinals 2 !1 n Mk such that is a cardinal in L.Mk /. Let
Ck D j k D :
Choose .Fk.S/ ; Fk.T / / 2 Mk such that
(1.1) Fk.S/ W .!1 /Mk ! P .!1 /Mk n .INS /Mk ,
(1.2) Fk.T / W .!1 /Mk ! P .!1 /Mk n .INS /Mk .
By Lemma 5.47 there is an iteration
jk W .Mk ; Ik / ! .Mk ; Ik /
of length !1 such that
5.5 Pmax
227
M
M
[ jk .f / j f W !1 k ! P .!1 k / \ Mk n .INS /Mk and f 2 Mk j k < !
which is easily achieved.
Let X D [Xk j k < ! and for each k < ! let
.MO k ; YOk /
be the image of .Mk ; jk .Yk // under the transitive collapse of X . Let
aO D j0 .a0 / \ X:
We claim that
.hMO k W k < !i; a/
O 2 Pmax
228
5 Applications
Combining Lemma 5.48 with Theorem 5.36 we obtain the following fairly general
. A more general version is given in Secexistence theorem for conditions in Pmax
tion 10.4; cf. Theorem 10.152. The version here sufces for our immediate purposes.
Suppose that is a 2 sentence such that it is a theorem of
ZFC C There is a Woodin cardinal
that there exists a boolean algebra B such that
V B :
For example could be any of the following.
(1) .
(2) MA C :CH.
(3) MA!1 C INS is !2 -saturated.
Theorem 5.49. Assume AD holds in L.R/. Then for each set A R with
A 2 L.R/;
such that
there is a condition .hMk W k < !i; a/ 2 Pmax
M0 ZFC C ;
and such that
(1) A \ M0 2 M0 ,
(2) hH.!1 /M0 ; A \ M0 i hH.!1 /; Ai,
(3) hMk W k < !i is A-iterable,
.
and further the set of such conditions is dense in Pmax
Proof. As usual the density of the desired conditions follows on abstract grounds (by
changing A and applying the iteration lemma, Lemma 5.44).
Fix A and let B0 be the set of x 2 R such that x codes an element of the rst order
diagram of the structure
hV!C1 ; A; 2i:
Let B1 be the set of x 2 R such that x codes an element of the rst order diagram of
the structure
hV!C1 ; B0 ; 2i:
Thus B1 2 L.R/.
By Theorem 5.36, there exist a countable transitive model M and an ordinal 2 M
such that the following hold.
(1.1) M ZFC.
(1.2) is the second Woodin cardinal in M .
5.5 Pmax
229
(1.3) B1 \ M 2 M and
hV!C1 \ M; B1 \ M; 2i hV!C1 ; B1 ; 2i:
(1.4) B1 \ M is C -weakly homogeneously Suslin in M .
Let 0 be the least Woodin cardinal of M and let
0 be the least strongly inaccessible cardinal of M above 0 . Thus since
M0 ZFC C There is a Woodin cardinal;
there exists a partial order P 2 M0 such that
M P :
Let g P be an M -generic lter (with g 2 V ). Thus:
(2.1) M g ZFC;
(2.2) is a Woodin cardinal in M ;
(2.3) B0 \ M g 2 M g and
hV!C1 \ M g; B0 \ M g; 2i hV!C1 ; B0 ; 2iI
(2.4) B0 \ M g is C -weakly homogeneously Suslin in M g;
noting that (2.3) and (2.4) follow from Lemma 2.29.
Let be the least strongly inaccessible cardinal of M g above . By (2.3),
1
B0 \ M g is not
1 in M g and so by (2.4), exists.
Let
X0 M g
be an elementary substructure structure such that X0 2 M g, B0 \ M g 2 X0 , and
such that X0 is countable in M g. Let M0 be the transitive collapse of X0 . Let a !1
be a set in X0 such that
!1 D .!1 /La
and let a0 D a \ X0 .
By Lemma 5.48 there exists
O 2 .Pmax
/M g
.hMO k W k < !i; a/
such that in M g,
(3.1) there exists a countable iteration
j W .M0 ; I0 / ! .M0 ; I0 /
such that j.a0 / D aO and such that
M0 D MO 0 ;
(3.2) hMO k W k < !i is B0 \ M g-iterable.
230
5 Applications
t
u
such that
.
Then D1 is dense in Pmax
t
u
Suppose G Pmax
is L.R/-generic. We assume ADL.R/ so that by Theorem 5.49,
is nontrivial. We associate to the generic lter G, a subset of !1 , AG and an ideal,
Pmax
IG . This is just as in case of Pmax .
The next theorem gives the basic analysis of Pmax
.
5.5 Pmax
231
Theorem 5.51. Assume AD L.R/ . Then Pmax
is !-closed and homogeneous.
Suppose G Pmax
is L.R/-generic. Then
L.R/G ZFC
and in L.R/G:
(1) P .!1 /G D P .!1 /;
(2) IG is a normal saturated ideal;
(3) IG is the nonstationary ideal.
t
u
Theorem 5.51 can be proved following the proof of the analogous theorem for Pmax .
One can also obtain the theorem as an immediate corollary of the following theorem
together with the analysis of L.R/Pmax .
Theorem 5.52. Assume AD L.R/ . Suppose G Pmax
is L.R/-generic. Then there
exists a lter H Pmax such that H is L.R/-generic and
L.R/G D L.R/H :
such that
Proof. Let DPmax be the set of .hMk W k < !i; a/ 2 Pmax
(1) M0 MA!1 ,
(2) M0 INS is saturated,
(3) .INS /M0 D .INS /M1 \ M0 .
By Theorem 5.50, DPmax is dense in Pmax
. Dene
G 2 L.R/H
and so L.R/G D L.R/H .
t
u
232
5 Applications
0
5.6 Pmax
0
0
We dene a fourth presentation of Pmax , this is Pmax
. The partial order Pmax
is essen
without the requirement that AC
hold in the models associated to the
tially just Pmax
conditions. This requires that history be added to the conditions as was done in the
(T)
.
denition of Pmax
0
is generalizing the following two theorems which are
Our purpose in dening Pmax
corollaries of the results of Chapter 3.
Theorem 5.53 (MA!1 ). Assume INS is !2 -saturated and that P .!1 /# exists. Then
there is a semi-generic lter
G Pmax
such that P .!1 /G D P .!1 /.
Proof. By Lemma 3.14 and Theorem 3.16, if
X H.!2 /
is a countable elementary substructure then MX is iterable where MX is the transitive
collapse of X .
Fix a set A !1 such that
!1 D !1LA :
Following the proof of Theorem 4.76, let
FA Pmax
be the set of all conditions h.M; I /; ai 2 Pmax such that there is an iteration
j W .M; I / ! .M ; I /
such that j.a/ D A and INS \ M D I .
Let
G D FA :
By Corollary 3.13, for each countable elementary substructure
X H.!2 /
such that A 2 X ,
h.MX ; IX /; AX i 2 G
P .!1 /G D P .!1 /:
A similar argument proves the corresponding theorem for Pmax
.
t
u
0
5.6 Pmax
233
Theorem 5.54 ( AC ). Assume INS is !2 -saturated and that P .!1 /# exists. Then there
is a semi-generic lter
G Pmax
such that P .!1 /G D P .!1 /.
t
u
(2) Mk 2 MkC1 ; !1
MkC1
D !1
c) j0 W hNk W k < !i ! hNk W k < !i is an iteration such that for all k < !,
234
5 Applications
d) j0 .X0 / X ;
e) if ..hNk W k < !i; X0 /; j1 / 2 X then j0 D j1 .
0
is dened as follows. Suppose
The order on Pmax
0
:
.hMk W k < !i; X /; .hNk W k < !i; Y / Pmax
Then
.hMk W k < !i; X / < .hNk W k < !i; Y /
if there exists an iteration
j W hNk W k < !i ! hNk W k < !i
such that ..hNk W k < !i; Y /; j / 2 X .
t
u
0
(T)
is a combination of the denitions of Pmax
and
Remark 5.58. The denition of Pmax
Pmax . One important item in the denition of Pmax has been eliminated, this is clause
0
is such that X ;
(7). The relevant observation is that if .hMk W k < !i; X / 2 Pmax
then (7) holds, i. e. there exists a set Z 2 M0 such that
M1
(1) Z P .!1 /M0 n INS
,
(2) M0 jZj D !1 ,
M0
.
(3) for all S; T 2 Z, if S T then S \ T 2 INS
t
u
0
extension of L.R/, assuming ADL.R/ , is a routine modiThe analysis of the Pmax
(T)
extension. As for the analysis of the Pmax
-extension,
cation of the analysis of the Pmax
a non-interference lemma is required. The proof is similar to that of Lemma 5.28 which
(T)
.
is the corresponding lemma for Pmax
and that
j W hMk W k < !i ! hMk W k < !i
is an iteration of hMk W k < !i of length !1 . Then hMk W k < !i M0 .
t
u
such that x 2 M0 .
0
(1) A lter G Pmax
is semi-generic if for each < !1 there exists
0
such that < .!1 /M0 .
.hMk W k < !i; X / 2 Pmax
0
5.6 Pmax
235
0
(2) Suppose that G Pmax
is a semi-generic lter. Then for each p 2 G, with
p D .hMk W k < !i; X /,
t
u
0
The existence of conditions as required for the analysis of the Pmax
extension is an
.
immediate corollary of Theorem 5.49 which is the corresponding theorem for Pmax
Theorem 5.61. Assume AD holds in L.R/. Then for each set A R with A 2 L.R/;
0
such that
there is a condition .hMk W k < !i; X / 2 Pmax
(1) A \ M0 2 M0 ,
(2) hH.!1 /M0 ; A \ M0 i hH.!1 /; Ai,
(3) hMk W k < !i is A-iterable,
0
.
and further the set of such conditions is dense in Pmax
t
u
0
The basic analysis of Pmax
is given in the following two theorems. The second
0
, the proof is similar.
theorem is the version of Theorem 5.39 for Pmax
0
Theorem 5.62. Assume AD L.R/ . Then Pmax
is !-closed and homogeneous.
0
Suppose G Pmax is L.R/-generic. Then
L.R/G ZFC
and in L.R/G:
(1) P .!1 /G D P .!1 /;
(2) IG is a normal saturated ideal;
(3) IG is the nonstationary ideal.
t
u
0
Theorem 5.63. Assume AD holds in L.R/. Suppose G Pmax
is L.R/-generic. Then
L.R/G D L.R/gh
where g Pmax is L.R/-generic and
h .!2<!2 /L.R/g
is L.R/g-generic.
t
u
236
5 Applications
The generalization of Theorem 5.53 and Theorem 5.54 that we seek is the following.
Theorem 5.64. The following are equivalent.
(1) 2@0 D 2@1 and there exists a countable elementary substructure
X H.!2 /
such that the transitive collapse of X is iterable.
(2) There exists a semi-generic lter
0
G Pmax
0
5.6 Pmax
237
given by G where p D .hM.p;k/ W k < !i; X.p/ /i. By the denition of P .!1 /G ,
P .!1 /G D [P .!1 / \ M.p;0/
j p 2 G;
and so by (2), 2@0 D 2@1 . To nish we must prove that (2) implies that there exists a
countable elementary substructure
X H.!2 /
such that the transitive collapse of X is iterable. By (2), for every x 2 R there exists
an iterable sequence,
hNk W k < !i;
such that x 2 N0 . Thus arguing as in the proof of Theorem 3.19, for every x 2 R,
x # exists. Thus the existence of X H.!2 /, countable and with iterable transitive
collapse, is essentially an immediate corollary of Lemma 4.22 and Theorem 3.19. u
t
238
5 Applications
5.7
The Axiom
We prove that the Pmax -extension can be characterized by a certain kind of generic
homogeneity. This property generalizes to L.P .!1 // a well known property which
characterizes L.R/ in the case that L.R/ is computed in LG where G is L-generic
for adding uncountably many Cohen reals to L. This is the symmetric extension of L
given by innitely many Cohen reals.
Suppose that L.R/ is a symmetric extension of L for adding innitely many Cohen
reals. Then the following hold.
(1) There is an L-generic Cohen real.
(2) Let
X R
be a nonempty set which is ordinal denable in L.R/. Then there exists a term
2 L such that for all L-generic Cohen reals c,
.c/ 2 X;
where .c/ is the interpretation of by the generic lter given by c.
It is straightforward to show that the converse is also true: If (1) and (2) hold then
L.R/ is a symmetric extension of of L for adding innitely many Cohen reals. The
point is that (1) and (2) imply that for every x 2 R n L,
Lx D Lc
for some c 2 R which is an L-generic Cohen real. Further (1) and (2) also imply that
for every x 2 R, there is an Lx-generic Cohen real.
We generalize this to L.P .!1 // in Theorem 5.67. This gives a reformulation of
the axiom ./ which seems more suited to the investigation of the consequences of this
axiom. As we have indicated above, this property characterizes the Pmax -extension.
We x some notation. Recall that the partial order Coll.!; <!1 / consists of nite
partial functions
p W ! !1 ! !1
such that for all k 2 ! and for all 2 !1 ,
p.k; / < 1 C :
Denition 5.65. Suppose
g Coll.!; <!1 /
is a lter. For each < !1 let
Sg D j for some p 2 g; p.0; / D :
Suppose !1 Coll.!; <!1 /, then
Ig . / D j .; p/ 2 for some p 2 g:
Ig . / is the interpretation of by g.
t
u
239
Remark 5.66. The sequence hSg W < !1 i, dened in Denition 5.65, can be dened
using any reasonable sequence h W < !1 i of terms for pairwise disjoint stationary
subsets of !1 . The only important requirement is that
h W < !1V i 2 L:
t
u
M
240
5 Applications
(1.1) j 2 Lt g,
(1.2) j.!1M0 / D !1 ,
(1.3) for all
if S D j.hi /./ then
t
u
241
t
u
one need only consider sets
X P .!1 /
which are denable (without parameters).
Lemma 5.69 (For all t 2 R, t # exists). The following are equivalent.
(1) .
(2) Suppose X P .!1 /, X ;, and that X is denable by a 2 formula.
Then there exist t 2 R and a term
!1 Coll.!; <!1 /
such that
2 Lt
and such that for all lters
g Coll.!; <!1 /
if g is Lt -generic and if for each < !1 , Sg is stationary, then
Ig ./ 2 X:
Proof. We have only to prove that (2) implies .
each z ! let z be the least ordinal such that there exists a counterexample
For
to which is denable in Vz from z and ordinal parameters. Let Xz be the least
such counterexample in the natural order of sets which are denable in Vz from z and
ordinal parameters.
If no such counterexample exists then set Xz D ;.
Let X be the set of A !1 such that
A 2 Xz
242
5 Applications
If fails then X ;.
Assume toward a contradiction that X ;.
X is denable by a 2 formula. Therefore by (2) there exist t 2 R and a term
!1 Coll.!; <!1 /
such that
2 Lt
and such that for all lters
g Coll.!; <!1 /
if g is Lt -generic and if for each < !1 , Sg is stationary, then
Ig ./ 2 X:
#
243
244
5 Applications
1
It is not difcult to show, assuming AD, that every !1 -borel set is
3 . This is an
immediate consequence of the fact that assuming AD, the
13 sets are closed under !1
unions. In fact, Lemma 5.70 can be proved assuming AD, and so, assuming AD, the
following are equivalent:
(1) A is !1 -borel,
(2) A is effectively !1 -borel,
(3) there exist x 2 R, < !2 , and a formula , such that !1 < and such that
A D y 2 R j L x; y x; y:
t
u
Theorem 5.72. Assume . Suppose that A R and that A is denable from ordinal
and real parameters. Then the following are equivalent.
(1) A is !1 -borel.
1
(2) A is
3 and
L.R/ A is !1 -borel:
(3) There exist x 2 R, < !2 , and a formula , such that !1 < and such that
A D y 2 R j L x; y x; y:
t
u
Proof. (2) trivially implies (1) and by Lemma 5.70, (3) also implies (1).
We assume (1) and prove (3).
By Lemma 5.70 there exist S0 !1 , 0 < !2 , and a formula 0 .x0 ; x1 /, such that
!1 < 0 and such that
A D y 2 R j L0 S0 ; y 0 S0 ; y; !1 :
Clearly we can suppose that 0 is less than the least ordinal such that !1 < and
such that L S0 is admissible.
Fix 0 and 0 . Let X P .!1 / be the set of S such that
(1.1) A D y 2 R j L0 S; y 0 S; y,
(1.2) 0 < where is the least ordinal above !1 such that L S is admissible.
Thus X ; and since A is denable from ordinal and real parameters, so is X .
By there exist x 2 R and a term
!1 Coll.!; <!1 /
such that
2 Lx
and such that for all lters
g Coll.!; <!1 /
245
246
5 Applications
holds. Then:
(1) 12 D !2 .
(2) Every club in !1 contains a club which is constructible from a real.
(3) Suppose A !1 is conal. There is a wellordering <A of !1 and a club C !1
such that for all 2 C , rank.<A j / 2 A.
(4) Suppose A !1 is conal. The set A contains a conal subset which is constructible from a real.
(5) Suppose A !1 and that A is not constructible from a real. There exists a real
x and a lter g Coll.!; <!1 / such that g is Lx-generic and such that
Lxg D LxA:
Proof. Clearly (2) implies (1). Further assuming (2) and that for all x 2 R, x # exists,
(3) and (4) are equivalent. We note that (3) simply asserts that AQ ; whenever A is a
conal subset of !1 .
(2) is a trivial consequence of .
We prove (5), the proof of (4) is similar though much easier.
Let X0 P .!1 / be the set of counterexamples to (5). Assume toward a contradiction that
X0 ;.
By there exist t0 2 R,
0 !1 Coll.!; <!1 /;
such that for all lters g Coll.!; <!1 /, if g is Lt0 -generic and if
Sg j < !1 P .!1 / n INS ;
then Ig .0 / 2 X0 .
Since t0# exists, we can suppose by modifying t0 if necessary that 0 is denable in
Lt0 from t0 and !1 .
Let g0 Coll.!; <!1 / be a lter such that g0 is Lt0# -generic and such that
Sg0 j < !1 P .!1 / n INS :
Let 0 be the least indiscernible of Lt0 . Let 0 be the least ordinal such that 0 < 0
and such that there exist p; q 2 Coll.!; <!1 / satisfying the following.
(1.1) pj.! 0 / 2 g0 .
(1.2) qj.! 0 / 2 g0 .
(1.3) p 0 2 Ig0 .0 /.
(1.4) q 0 Ig0 .0 /.
247
Let p0 and q0 be the Lt0 -least such conditions relative to the canonical wellordering of Lt0 given by t0 .
Since Ig0 .0 / 2 X0 , 0 , p0 and q0 exist. Since 0 is denable in Lt0 from t0 and
!1 ,
0 ; p0 ; q0 L 1 t0
where 1 is the least indiscernible of Lt0 above 0 .
Let h W < !1 i be the increasing enumeration of the indiscernibles of Lt0
below !1 .
For each < !1 let
j W Lt0 ! Lt0
be the canonical elementary embedding such that
j.0 / D
and such that
Lt0 D j .f /.s/ j f 2 Lt0 ; s 2 <! :
These requirements uniquely specify j .
Since 0 is denable in Lt0 from .t0 ; !1 /, for each < !1 ,
j .0 / D 0 :
For each < !1 let
D j .0 /;
let
p D j .p0 /;
and let
q D j .q0 /:
Suppose that g Coll.!; <!1 / is an Lt0 -generic lter such that
g \ Coll.!; <0 / D g0 \ Coll.!; <0 /:
Then for each < !1 , the elementary embedding j lifts to an elementary embedding
jO W Lt0 g0 \ Coll.!; <0 / ! Lt0 g \ Coll.!; < /:
Thus
(2.1) p j.! / 2 g,
(2.2) q j.! / 2 g,
(2.3) p 2 Ig .0 /,
(2.4) q Ig .0 /.
We work in Lt0# g0 and dene a lter g Coll.!; <!1 / which is Lt0 -generic.
Construct g \ Coll.!; < / by induction on such that
g \ Coll.!; <0 / D g0 \ Coll.!; <0 /
and such that for all < !1 , if is a limit ordinal then
248
5 Applications
(3.1) p W ! ! j p 2 g0 g,
(3.2) i 2 ! j CiC1 2 Ig .0 / codes g0 \ Coll.!; <C! /,
(3.3) g \ Coll.!; < / 2 Lt0# g0 \ Coll.!; < /.
This is easily done. There are two relevant points.
(4.1) For each < !1 if is a limit ordinal and if
h0 Coll.!; <Ci /
is Lt0 -generic then the lter h0 can be enlarged to an Lt0 -generic lter
h1 Coll.!; <CiC1 /
with either pCi 2 h1 or qCi 2 h1 as desired.
(4.2) The are strongly inaccessible in Lt0 . Therefore if < ! is a limit ordinal
and h Coll.!; < / is a lter such that for all < ,
h \ Coll.!; < /
is Lt0 -generic, then h is Lt0 -generic.
With these observations in hand we sketch the inductive step. The limit steps are immediate, the uniformity of the construction at successor steps ensures that (3.3) holds.
Suppose < !1 , is a limit ordinal and that g \ Coll.!; < / is given. Since
g \ Coll.!; < / 2 Lt0# g0 \ Coll.!; < /;
we can dene g \ Coll.!; / satisfying (3.1) and preserving Lt0 -genericity.
For each i ! let
i D C2Ci :
Thus for each i < !,
.pCiC1 ; qCiC1 / 2 L i t0 :
We work in Lt0# g0 \ Coll.!; <! /.
Let x ! be such that
x 2 Lt0# g0 \ Coll.!; <! /
and such that x codes g0 \ Coll.!; <! /.
We choose x to be the Lt0# g0 \ Coll.!; <! /-least such set in the canonical
wellordering of Lt0# g0 \ Coll.!; <! / given by .t0# ; g0 \ Coll.!; <! //.
It is straightforward to construct, using (4.1),
g \ Coll.!; <i /
by induction on i < ! such that
i < ! j pCiC1 2 g D x
and such that
i < ! j qCiC1 2 g D ! n x:
By (4.2)
249
is as desired. We dene
g \ Coll.!; <! /
g D pj.! ! ; !1 // j p 2 g
t
u
Remark 5.74. (4) can be generalized to give the following. Suppose T !1<!
is a wellfounded subtree such that for all s 2 T if s is not of rank 0 in T then
j s _ 2 T has size !1 . Then there is a subtree T T such that:
(1) T is constructible from a real;
(2) rank.T / D rank.T /;
(3) for all s 2 T if s is not of rank 0 in T then j s _ 2 T has size !1 .
This in fact follows just from the assertion that the nonstationary ideal on !1 is saturated together with the assertion that every set A !1 contains a conal subset which
is constructible from a real.
t
u
250
5 Applications
The following theorem which we shall prove in Section 6.2.4 shows that the axiom
implies that there are no weak Kurepa trees. The relevant theorem of Section 6.2.4
is Theorem 6.124 and
is actually a little stronger than Theorem 5.75. Note
that by
Theorem 5.73(5), implies that for every B !1 , B # exists. Therefore implies
that for every B !1 ,
jP .!1 / \ LBj D !1 :
Theorem 5.75. Assume
that
. Suppose that A !1 . Then there exists B !1 such
(1) A 2 LB,
(2) for all Z !1 if
Z \ 2 LB
t
u
Remark 5.76. (1) The rst 4 consequences of given in Theorem 5.73 follow
from Martins Maximum though the proofs seem more involved.
(2) We do not know if (5) of Theorem 5.73 can be proved from Martins Maximum.
This problem seems very likely related to the problem of the relationship of Martins Maximum and the axiom ./. Similarly we do not know if Theorem 5.75
can be proved from Martins Maximum. The two problems are likely closely
related. Note that if B !1 satises the condition (2) of Theorem 5.75 and if
t
u
B # exists then there must exist x 2 LB \ R such that x # LB.
Lemma 5.77. Assume
. Suppose that y 2 R and that
M 2 H.!2 /
Sg
251
Let
g1 Coll.!; <!1 /
and
g2 Coll.!; <!1 /
be Lz-generic lters such that
(1.1) Sg1 is stationary for each < !1 ,
(1.2) Sg2 is stationary for each < !1 ,
(1.3) g2 is Lzg1 -generic.
Such a pair of lters is easily constructed using z # .
Let A1 D Ig1 . / and let A2 D Ig2 ./.
Thus A1 2 X and A2 2 X .
Therefore
M 2 Lzg1 \ Lzg2 :
It follows that M 2 Lz.
t
u
Lemma 5.77 has the following corollary. This is also a corollary of Theorem 5.73
and Theorem 3.22, but the proof we give is more direct.
Corollary 5.78. Assume
. Then for all x 2 R, x exists.
Proof. Fix x 2 R.
By Theorem 5.73(2) and Lemma 5.77,
F \ P .!1 / \ HODx
is an HODx -ultralter where F is the club lter on !1 .
Therefore !1 is a measurable cardinal in HODx and
F \ P .!1 / \ HODx
is a normal measure on !1 .
Let
N D LF ; x:
Since !1 is a measurable cardinal in N ,
jV!1 \ N j D !1
and so by Lemma 5.77 there exists y0 2 R such that
V!1 \ N 2 Ly0 :
Since y0# exists, for all A !1 , if
A \ 2 Ly0
for all < !1 then A 2 Ly0 .
252
5 Applications
Thus
P .!1 / \ N Ly0
and so
jP .!1 / \ N j D !1 :
Hence by Lemma 5.77 again there exists y1 2 R such that
P .!1 / \ N 2 Ly1
and such that
F \ P .!1 / \ N 2 Ly1 :
Thus
N Ly1
#
y1#
t
u
. Suppose
A 2 P .!1 / n L.R/:
h.M; I /; ai 2 Pmax
253
254
5 Applications
I D INS \ M :
g is Lt -generic and
A D Ij.g / ./:
Moreover for each < ,
.Sg /N gh I0 :
Thus, by the elementarity of j j0 , for each < !1 ,
Sj.g
Since
/
I :
I D INS \ M ;
This is a contradiction.
Corollary 5.80. Assume . Then MA!1 .
Proof. Let A 2 P .!1 / n L.R/ code a pair
.P ; D/ 2 H.!2 /
such that
255
j W .M; I / ! .M ; I /
such that
I D INS \ M
M MA!1 :
t
u
. Suppose
A 2 P .!1 / n L.R/:
256
5 Applications
Proof. Suppose
A 2 P .!1 / n L.R/:
Fix x 2 R such that
!1 D .!1 /LA;x :
B 2 P .!1 /G :
Therefore
P .!1 / D P .!1 /G
and so G satises (2).
We nish by proving that
HODP .!1 / D HODR G:
Let P be the predicate dened as follows. .A; ; ; b/ 2 P if
(1.1) 2 Ord and > !2 ,
(1.2) b 2 R Ord,
(1.3) A 2 P .!1 /,
(1.4) is a formula in the language for set theory,
(1.5) V A; b.
It is an elementary fact that
HODP .!1 / D L.P; P .!1 //:
Therefore it sufces to show that for all 2 Ord,
P \ V 2 HODR G:
Let Q be the following predicate. .t; ; ; ; b/ 2 Q if
257
Using the results of the next section, Section 5.8, the assumption in the Corollary 5.82 that
L.R/ AD
can be eliminated; i. e. if
then ./ and
.
t
u
implies that a perfect set theorem holds for denable subsets
258
5 Applications
Theorem 5.83. Assume holds. Suppose X P .!1 / and that X is denable in
L.P .!1 // from real and ordinal parameters. Suppose there exists A 2 X such that
A L.R/. Then there exists a function
W 2<!1 ! !1 !
and a conal set
T !1
such that for all s 2 2<!1 and for all t 2 2<!1 :
(1) if s t then .s/ .t / and
.t / \ D .s/ \
for all 2 .s/;
(2) for all 2 T \ dom.s/,
2 .s/
if and only if
s./ D 1I
and such that for all F 2 2!1 ,
[.F j/ j < !1 2 X:
Proof. The proof is quite similar to the proof of Theorem 5.73(5).
Let X0 D X n L.R/. Thus X0 is denable from ordinal and real parameters, and
X0 ;.
By there exist t0 2 R,
0 !1 Coll.!; <!1 /;
such that for all lters g Coll.!; <!1 /, if g is Lt0 -generic and if
Sg j < !1 P .!1 / n INS ;
then Ig .0 / 2 X0 .
Since t0# exists, we can suppose by modifying t0 if necessary that 0 is denable in
Lt0 from t0 and !1 .
Let F t0 be the set of lters g Coll.!; <!1 /, such that g is Lt -generic and such
that
Sg j < !1 P .!1 / n INS :
Fix g0 2 F t0 . Since Ig0 .0 / 2 X0 ,
Ig0 .0 / Lt0 :
Let 0 be the least indiscernible of Lt0 . Let 0 be the least ordinal such that
0 < 0 and such that there exist p; q 2 Coll.!; <!1 / satisfying the following.
(1.1) pj.! 0 / 2 g0 .
(1.2) qj.! 0 / 2 g0 .
259
D j .0 /;
let
p D j .p0 /;
and let
q D j .q0 /:
Suppose that g 2 F t0 and that
g \ Coll.!; <0 / D g0 \ Coll.!; <0 /:
Then for each < !1 , the elementary embedding j lifts to an elementary embedding
jO W Lt0 g0 \ Coll.!; <0 / ! Lt0 g \ Coll.!; < /:
Thus
(2.1) p j.! / 2 g,
(2.2) q j.! / 2 g,
(2.3) p 2 Ig .0 /,
(2.4) q Ig .0 /.
260
5 Applications
Let
T D C1 j < !1 :
For each < !1 let
F t0
2 S .s/
261
and so is as desired.
Remark 5.84. Thus subsets of 2!1 which are denable in L.R/Pmax are either in L.R/
t
u
or contain copies of 2!1 .
The reformulation of ./ as taken together with the results of Chapter 4 strongly
suggests that, assuming ./, one should be able to analyze sets
X P .!1 /
which are denable in the structure
hH.!2 /; 2; INS i
by a 1 formula.
We explore the possibilities for classifying specic denable subsets of P .!1 /.
For this we assume that the axiom ./ holds and we focus on attempting to classify
partitions
W !1 2 ! 0;1
for which there is no homogeneous rectangle for 0, of (proper) cardinality @1 . Here we
adopt the convention that if Z D A B !1 !1 is a rectangle, then Z has proper
cardinality @1 if both A and B have cardinality @1 .
This is related to the following variation of a question of S. Todorcevic.
Is it consistent that for any partition
W !1 2 ! 0;1;
either there is a homogeneous rectangle for for 0, of (proper) cardinality @1 ,
or there is no such homogeneous rectangle in any generic extension of V which
preserves !1 ?
262
5 Applications
Remark 5.85.
L.a/
Let b be a set in L.a/ which codes a. One can show that M3 .a/ is precisely the
set of all sets, c, which can be coded by a set z such that z 2 Q3 .b/.
1
Denition 5.86 (
2 -Determinacy). Suppose that
W !1 2 ! 0;1:
(1) Suppose that X !1 . Let E
.3/ X be the set of < !1 such that there exists
Z1 Z2
such that
263
t
u
Assume there exists a Woodin cardinal with a measurable cardinal above. Then
1
by .Martin and Steel 1989/,
2 -Determinacy holds and so Denition 5.86 applies. If
the partition given by has a homogeneous rectangle for 0, of (proper) cardinality @1 ,
then necessarily E
.3/ X; A contains a club in !1 .
Another trivial observation is that if for some .X; A/ the set E
.3/ X; A is nonstationary then there exists Y !1 such that
E
.3/ Y; A D ;:
With the notation as above we have the following lemma.
Lemma 5.87. Assume there is a Woodin cardinal with a measurable above. Then
(1) E
.3/ X contains a closed unbounded set or E
.3/ X is nonstationary,
(2) E
.3/ X; A contains a closed unbounded set or E
.3/ X; A is nonstationary.
264
5 Applications
and such that S exists. By the hypothesis of the lemma, and S exist.
Let
N D LS :
Let
Y N
be a countable elementary substructure containing innitely many Silver indiscernibles
of N .
We prove that
Y \ !1 2 E
.3/ X :
Let NY be the transitive collapse of Y . Let SY be the image of S under the collapsing map and let Y be the image of .
Let D NY \ Ord.
Thus
NY Y is a Woodin cardinal
and
NY D L SY :
Since Y contains innitely many indiscernibles of N ,
L SY LSY :
The key points are that
.E
.3/
XY /NY D E
.3/ X \ Y \ !1
Y
and that
XY is stationary;
NY E
.3/
Y
where XY and Y are the images of X and under the collapsing map.
By elementarity,
.E
.3/
XY /LSY D E
.3/ X \ Y \ !1
Y
and
XY is stationary:
LSY E
.3/
Y
Let a D .E
.3/
XY /LSY and let D !1LSY D Y \ !1 .
Y
Let
G .Q<Y /LSY
be an LSY -generic lter for the stationary tower such that a 2 G.
Let
265
j W LSY ! LSY
t
u
Theorem 5.88. Assume ./. Suppose that A !1 . Then there exists a transitive inner
model M containing the ordinals and the set A such that
M ZFC C There exist ! many Woodin cardinals:
Proof. We sketch the argument. We require the following strengthening of Theorem 5.73(5). For each x 2 R let
Nx D HODxL.R/ :
Suppose A !1 . Then there exist x 2 R and G Coll.!; <!1V / such that
(1.1) G is Nx -generic,
(1.2) A 2 Nx G.
266
5 Applications
:
Let X be the set of A !1 for which x and G do not exist satisfying (1.1) and (1.2).
By there exist t 2 R and 2 Lt such that
(2.1) !1 Coll.!; <!1 /,
(2.2) for each Lt -generic lter
g Coll.!; <!1V /;
if Sg .g/ INS for each < !1 then Ig ./ 2 X .
Let g Coll.!; <!1 / be an N t -generic lter such that
Sg j < !1 \ INS D ;;
and let A D Ig . /. Thus A 2 X , but
A 2 N t g
which is a contradiction. The lter g is easily constructed since there is a closed unbounded set
C !1V
of ordinals which are strongly inaccessible in N t .
We now prove the theorem. Fix A !1 . Fix x 2 R and G Coll.!; <!1V / such
that
(3.1) G is Nx -generic,
(3.2) A 2 Nx G.
By the results of .Koellner and Woodin 2010/, there is an inner model M of Nx such
that
(4.1) Ord M ,
(4.2) M ZFC C There exist ! many Woodin cardinals,
(4.3) P .!1V / \ Nx D P .!1V / \ M .
By .Koellner and Woodin 2010/, !1V is the least measurable cardinal of Nx . Therefore
M G ZFC C There exist ! many Woodin cardinals:
However A 2 M G and so M G is as required.
Combining Lemma 5.87 and Theorem 5.88 we obtain:
t
u
267
268
5 Applications
then
N j.0 / has no homogeneous rectangle for 0, of (proper) cardinality @1 .
Let A D hS W < !1 i be an enumeration of
.P .!1 //M!1 n I!1 :
Since
I!1 D INS \ M!1 ;
for each < !1 , S is a stationary subset of !1 .
Let X !1 be a set which codes M!1 .
Suppose
Y H.!2 /
is a countable elementary substructure with
h.M ; I /; G ; j; W < !1 i 2 Y
and let D Y \ !1 .
Then
and X \ codes M .
Suppose that N is a countable transitive model of ZFC such that !1N D and such
that X \ 2 N . Then
.P .!1 //M N :
Therefore by the choice of h.M0 ; I0 /; a0 i,
E
.3/ X; A
and so E
.3/ X; A D ;.
t
u
There is a version of Lemma 5.90 for dealing with the existence of homogeneous
sets for partitions
W !1 2 ! 0;1:
This requires the obvious adaptation of Denition 5.86.
1
Denition 5.91 (
2 -Determinacy). Suppose that
W !1 2 ! 0;1:
(1) Suppose that X !1 . Let F
.3/ X be the set of < !1 such that there exists
Z
such that
a) Z has ordertype ,
b) .; / D 0 for all .; / 2 Z Z with < ,
c) Q3 .Z .X \ / j2 / ;,
d) M3 .Z .X \ / j2 / D !1 .
269
Z has ordertype ,
.; / D 0 for all .; / 2 Z Z with < ,
Q3 .a/ ;,
M3 .a/ D !1 ,
for each < , S \ 2 M3 .a/ and S \ is a stationary set within
M3 .a/,
where
a D Z .X \ / j2 :
t
u
t
u
270
5 Applications
Remark 5.94. Todorcevics theorem is actually stronger, Theorem 5.93 is simply the
version relevant to our discussion. Note that Theorem 5.93(1) asserts in effect that
t
u
cannot have a homogeneous set of cardinality @1 for 1.
By combining Theorem 5.67 and Lemma 5.90 we obtain the next theorem. Suppose
g Coll.!; <!1 /
is a maximal lter. Let
Ag D hSg j < !1 i:
Suppose
L!1 Coll.!; <!1 /:
Then
Ig . / D z j .z; p/ 2 g for some p 2 g:
This generalizes the denition of Ig ./ given previously where !1 Coll.!;<!1 /.
Theorem 5.95. Assume ./. Suppose
W !1 2 ! 0;1
is a partition with no homogeneous rectangle for 0 of .proper/ cardinality @1 . Then
there exist t !,
L!1 Coll.!; <!1 /;
and a lter
g Coll.!; <!1 /;
such that
(1) 2 Lt ,
(2) g is Lt -generic, 2 Lt g and D Ig ./,
(3) Sg j < !1 P .!1 / n INS ,
(4) for all lters
g Coll.!; <!1 /;
if g is Lt -generic and if
271
Proof. Assuming , this follows easily from Lemma 5.90. By Theorem 5.67,
u
t
holds in L.P .!1 //.
The key question is the following.
Assume ./. Suppose
W !1 2 ! 0;1
272
5 Applications
g Coll.!; <!1 /;
if g is Lt -generic then
a) W !1 2 ! 0;1,
b) E
.3/
X D ;,
t
u
The connection with the question of Todorcevic is given in the Theorem 5.97 and
Theorem 5.98 below. We state these without giving the proofs for they require some
additional machinery which is beyond the scope of this presentation, particularly in the
case of Theorem 5.98.
The proof of Theorem 5.97 is completely straightforward given that (in the notation
of the theorem)
.E
.3/ X /V D .E
.3/ X /V Z
which is true by an absoluteness argument.
The proof of Theorem 5.98 requires some inner model theory and genericity iterations.
Theorem 5.97. Suppose is a Woodin cardinal, there is a measurable cardinal above
, and that
W !1 2 ! 0;1:
Suppose Z1 !1 and Z2 !1 are conal sets such that
273
t
u
t
u
In Chapter 7 we shall consider Pmax -extensions of inner models other than L.R/;
i. e. inner models satisfying stronger determinacy hypotheses. Using these results one
can show, for example, that if
ZFC C There are ! 2 many Woodin cardinals
is consistent then
ZFC C
is consistent.
In particular it is consistent for ./ to hold and for there to exist a Woodin cardinal with a measurable above. Therefore by Theorem 5.97 if ./ implies that for all
partitions
W !1 2 ! 0;1
either there is a homogeneous rectangle for 0 of (proper) cardinality !1 or there exists
a set X !1 such that E
.3/ X is nonstationary, then the answer to Todorcevics
question is yes.
274
5 Applications
Remark 5.99. (1) There is no evidence to date that Todorcevics question involves
large cardinals at all.
(2) One can dene other versions of E
.3/ X . For example dene E
.1/ X modifying the denition of E .3/ X by replacing M3 .a/ by M1 .a/ where for each
transitive set a 2 H.!1 /,
M1 .a/ D L .a/
where is the least ordinal such that L .a/ is admissible. E
.2/ X is dened
using
M2 .a/ D La:
(3) Assume ./. Suppose
W !1 2 ! 0;1
is a partition such that for some X !1 , E
.3/ X is nonstationary.
Is E
.2/ X nonstationary for some X !1 ?
Is E
.1/ X nonstationary for some X !1 ?
5.8
t
u
Assume ./ holds. We shall show in Section 6.1 that it does not necessarily follow that
the nonstationary ideal on !1 is !2 -saturated in V . This suggests that the structure of
the quotient algebra
P .!1 /=INS
is necessarily somewhat complicated.
The following lemma, which is well known, shows that assuming MA!1 , if I is a
normal, uniform, ideal on !1 which is !2 -saturated then the boolean algebra,
P .!1 /=I
is rigid.
Lemma 5.100 (MA!1 ). Suppose that I0 ; I1 are are normal, uniform, saturated ideals
on !1 and that
G0 .P .!1 / n I0 ; /
is V -generic. Suppose that
G1 .P .!1 / n I1 ; /
is a V -generic lter such that G1 2 V G0 . Then G0 D G1 .
275
a0 \
276
5 Applications
< !1 j f ./ \ is innite
277
t
u
t
u
278
5 Applications
Denition 5.102. Suppose I is a normal, uniform, ideal on !1 . The ideal I is quasihomogeneous if the following holds. Suppose that
X0 P .!1 /
is ordinal denable with parameters from I [ R. Suppose that there exists A0 2 X0
such that
A0 ; !1 n A0 \ I D ;:
Then for all A 2 P .!1 / n I if !1 n A I there exists B 2 X0 such that
A M B 2 I:
t
u
t
u
Theorem 5.105.
279
Proof. The theorem follows from the following claim which is an immediate corollary
of Lemma 4.36.
Suppose .M; I / is iterable, b !1M , b 2 M, b I and that !1M n b I . Suppose
S !1 is stationary and co-stationary. Then there exists an iteration
j W .M; I / ! .M ; I /
of .M; I / of length !1 such that C \ S D C \ j.b/ for some club C !1 and such
that for all d !1 , if d 2 M n I then d is stationary.
Suppose that
X0 P .!1 / n INS
and that X0 is ordinal denable from x where x 2 R. We suppose that X0 is nonempty
and that for all A 2 X0 , A is co-stationary.
Let Z0 be the set of pairs .t; / such that
(1.1) t 2 R, !1 Coll.!; <!1 /, 2 Lt ,
(1.2) for all lters
g Coll.!; <!1 /;
if g is Lt -generic and if for all < !1 , Sg is stationary, then Ig ./ 2 X0 .
Suppose that A0 2 X0 .
Therefore A0 L.R/ for otherwise, by Lemma 5.77, there exists z 2 R such that
A0 2 Lz. This contradicts that A0 is both stationary and co-stationary.
By Theorem 5.81 there exists a lter G0 Pmax such that G0 is HODR -generic
and such that
HODP .!1 / D HODR G0
and such that A0 D AG0 .
Since X0 is ordinal denable from a real parameter,
Z0 2 HODR :
Further there must exist a condition
h.M0 ; I0 /; a0 i 2 G0
such that in HODR ,
h.M0 ; I0 /; a0 i Pmax AG 2
280
5 Applications
t
u
Remark 5.106. An immediate corollary of Theorem 5.105 and Theorem 5.67 is that
assuming ./, the nonstationary ideal is quasi-homogeneous in L.P .!1 //.
This shows that MA!1 is consistent with the existence of a saturated ideal on !1
which is quasi-homogeneous. In Chapter 7, we shall improve this result, replacing
t
u
MA!1 by Martins MaximumCC .c/.
By Theorem 4.49, the basic analysis of the Pmax extension can be carried out just
assuming that for each set B R such that B 2 L.R/, there exists a condition
h.M; I /; ai 2 Pmax such that
(1) B \ M 2 M,
(2) hH.!1 /M ; B \ Mi hH.!1 /; Bi,
(3) .M; I / is B-iterable.
We now prove that this in fact holds, assuming . Our goal is to show that assum
ing , the nonstationary ideal is saturated in L.P .!1 //. By Theorem 5.104 it will
follow that implies ADL.R/ .
We rst prove that the conclusion of Lemma 4.52 follows from .
281
. Suppose B R and B 2 HODR . Then the set
282
5 Applications
j .B \ bk / D B \ j .bk /
283
j W .M; I / ! .M0 ; I 0 /
such that
(3.1) h.M; I /; ai < h.Mk ; Ik /; ak i,
(3.2) j.B \ j .bk // B \ j.j .bk // where
j W .Mk ; Ik / ! .Mk ; Ik /
is the iteration such that j .ak / D a,
(3.3) h.M0 ; I 0 /; a0 i 2 G where a0 D j.a/.
But this contradicts the fact that jk .B \ bk / D B \ Nk .
Therefore for each k < ! there exists qk 2 G such that
p < qk j p 2 Pmax Dk :
Note that Dk is denable in the structure
hH.!2 /; B; G; 2i
from bk . Therefore we can suppose that qk 2 Zk , for such a condition must exist in
Zk . This implies that
h.MkC1 ; IkC1 /; akC1 i 2 Dk :
For each k < n < !, let
jk;n W .Mk ; Ik / ! .Mkn ; Ikn /
be the iteration such that
jk;n .ak / D an
and let
jk;! W .Mk ; Ik / ! .Mk! ; Ik! /
be the iteration such that
jk;! .ak / D [an j n < !:
284
5 Applications
!
!
; IkC1
//
jO..MkC1
MO X D [jO.NkX / j k < !
and so
jO.B \ MX / D B \ MO X :
t
u
Therefore MX is B-iterable.
As a corollary to Lemma 5.107 we obtain that
of Pmax .
Theorem 5.108. Assume
. Suppose B R and that B 2 HODR . Then there exists
h.M; I /; ai 2 Pmax
such that
(1) B \ M 2 M,
(2) hH.!1 /M ; B \ Mi hH.!1 /; Bi,
(3) .M; I / is B-iterable.
285
Proof. Fix G Pmax such that G is HODR -generic and such that
P .!1 /G D P .!1 /:
Suppose 2 Ord,
and that
< L.B;R/ :
By Corollary 5.80,
L .B; R/G MA!1 :
Let A R be such that A 2 HODR and such that
<
11 .A/:
By Lemma 5.107, there exists a countable elementary substructure X H.!2 / such
that
hX; A \ X; 2i hH.!2 /; A; 2i
and such that the transitive collapse of X is A-iterable.
Therefore by Lemma 4.24, there exists a countable elementary substructure
Y L .B; R/G
such that B; AG Y and such that MY is strongly iterable where MY is the transitive collapse of Y .
Since B 2 Y , it follows by Theorem 3.34, that H.!2 /MY (which is the transitive
collapse of Y \ H.!2 /) is B-iterable.
MY
. Thus the structure .MY ; IY / is B-iterable. Let a be the image of
Let IY D INS
AG under the collapsing map. Thus
h.MY ; IY /; ai 2 Pmax
t
u
and is as required.
Corollary 5.109. Assume
. Then
t
u
286
5 Applications
t
u
t
u
Chapter 6
Pmax variations
In this chapter we dene several variations of Pmax . These yield models which, like
those dened in the next chapter, are conditional versions of the Pmax -extension.
The models obtained in this chapter condition the Pmax -extension by varying the
structure,
hH.!2 /; INS ; 2i;
relative to which the absoluteness theorems are proved.
One of these is the Qmax -extension which we shall dene in Section 6.2.1. This extension has two interpretations as a conditional extension. By modifying the structure,
hH.!2 /; INS ; 2i;
the Qmax -extension is the Pmax -extension conditioned on a form of . A very interesting feature of the Qmax -extension is that in it the nonstationary ideal on !1 is !1 -dense.
Further it also can be interpreted as the Pmax -extension conditioned by this, i. e. the
Qmax -extension realizes every 2 sentence in the language for the structure
hH.!2 /; INS ; 2i
which is (suitably) consistent with proposition that the nonstationary ideal is !1 -dense.
CH fails in the Qmax -extension so we also obtain as a corollary consistency of an
!1 -dense ideal on !1 together with :CH. Finally the Qmax -extension is a generic extension of L.R/ and ADL.R/ is sufcient to prove things work. This substantially lowers the upper bound for the consistency strength of the existence of an !1 -dense ideal
on !1 and in fact provides the optimal upper bound, the two theories are equiconsistent. Previous unpublished results of Woodin required the consistency of the existence
of an almost huge cardinal or the consistency of
ZF C ADR C is regular;
see Foreman .2010/ for a survey of results related to saturated ideals and generic elementary embeddings.
There is an important difference in the results here. The previous methods produced
models in which there is an !1 -dense ideal on !1 and in which CH holds. In the
context of CH the consistency of the existence of an !1 -dense ideal on !1 is quite
strong, much stronger than that of AD. This provides an example of a combinatorial
proposition whose consistency strength varies depending on whether one requires that
CH holds.
We also prove that the existence of an !1 -dense ideal on !1 implies that there is
a nonregular ultralter on !1 without assuming CH, this is a theorem of Huberich
.1996/. Combining these results also gives a new upper bound for the consistency
strength of the existence of a nonregular ultralter on !1 .
288
6 Pmax variations
6.1
Pmax
t
u
6.1
2P
max
289
This shows that in L.R/Pmax the quotient algebra P .!1 /=INS is not absolutely saturated and it answers the question above. The point here is that if the nonstationary ideal
is saturated then every normal ideal on !1 is of the form INS _ S for some S !1 .
Remark 6.3. It may seem strange that PNS could ever be nontrivial and yet be .!1 ; 1/distributive, or more generally that by forcing with an .!1 ; 1/-distributive partial order
it is possible to create a saturated ideal on !1 .
However suppose that
P .!1 /=INS RO.B Coll.!; !1 //
where B is a complete boolean algebra which is .!1 ; 1/-distributive and !2 -cc. Then
it is not difcult to show that
RO.PNS / B:
Further if G PNS is V -generic then in V G,
P .!1 /=IG RO.Coll.!; !1 //I
i. e. in V G, IG is an !1 -dense ideal.
One can show it is relatively consistent
P .!1 /=INS RO.B Coll.!; !1 //
where B is a complete, nonatomic, boolean algebra which, as above, is !2 -cc and
.!1 ; 1/-distributive; i. e. where B is the regular open algebra corresponding to a Suslin
tree on !2 . This can be proved by constructing a Qmax variation where Qmax is the
partial order constructed in Section 6.2.1.
However the example indicated in Lemma 6.2 is more subtle.
t
u
Remark 6.4. In Chapter 9 we shall consider the Pmax -extensions of inner models of
AD strictly larger than L.R/. We shall prove that if P .R/ is a pointclass such
that
L.; R/ ADR C is regular;
then if G Pmax Coll.!2 ; !2 / is L.; R/-generic,
L.; R/G ZFC C Martins MaximumCC .c/:
The proof of Lemma 6.2 easily generalizes to show both that
290
6 Pmax variations
t
u
Lemma 6.6 (ZFC ). Suppose that I J are normal uniform ideals on !1 and that
I J . Suppose that h.M0 ; I0 ; J0 /; a0 i 2 2 Pmax . Then there is an iteration
j W .M0 ; I0 ; J0 / ! .M0 ; I0 ; J0 /
such that j.!1M0 / D !1 , I0 D I \ M0 and J0 D J \ M0 .
6.1
2P
max
291
Proof. The proof is quite similar to the proof of Lemma 4.36, which is the analogous
lemma for Pmax .
Fix a set X !1 such that X 2 J n I .
Fix a sequence hAk; W k < !; < !1 i of pairwise disjoint subsets of !1 such that
the following conditions hold.
(1.1) If < !1 is an even ordinal then Ak; X and Ak; is I -positive.
(1.2) If < !1 is an odd ordinal then Ak; !1 n X and Ak; is J -positive.
Fix a function
f W ! !1M0 ! M0
such that
(2.1) f is onto,
(2.2) for all k < !, f jk !1M0 2 M0 ,
(2.3) for all A 2 M0 if A has cardinality !1M0 in M0 then
A ran.f jk !1M0 /
for some k < !.
The function f is simply used to anticipate subsets of !1 in the nal model.
Suppose
j W .M0 ; I0 ; J0 / ! .M0 ; I0 ; J0 /
is an iteration. Then we dene
j .f / D [j .f jk !1M / j k < !
and it is easily veried that the range of j .f / is M0 . This follows from (2.3).
We construct an iteration of .M0 ; I0 ; J0 / of length !1 using the function f to
provide a book-keeping device for all of the subsets of !1 which belong to the nal
model. More precisely construct an iteration
h.M ; I ; J /; G ; j; W < !1 i
such that for each < !1 , if is even and if for some < !1 ,
(3.1) !1M 2 Ak; ,
(3.2) < !1M ,
(3.3) j0; .f /.k; / !1M ,
(3.4) j0; .f /.k; / I ,
then G is M -generic for P .!1M / \ M n I and j0; .f /.k; / 2 G . If is odd
and if for some < !1 ,
6 Pmax variations
292
be the iteration obtained by combining the iterations given by the conditions pi for
i > k. Thus jk is uniquely specied by the requirement that
jk .ak / D [ai j i < !:
For each k < !
6.1
2P
max
293
each k < ! pkC1 < pk . Let hpk W k < !i be the associated sequence of 2 Pmax
conditions and for each k < !, let
h.Mk ; Ik ; Jk /; ak i D pk :
Then there is an iteration
j W h.Mk ; Ik ; Jk / W k < !i ! h.Mk ; Ik ; Jk / W k < !i
M
Jk D J \ Mk :
t
u
The next lemmas record some of the relevant properties of the partial order 2 Pmax .
First we note that the nontriviality of Pmax immediately gives the nontriviality of 2 Pmax .
Lemma 6.8. Assume that for each set X R with X 2 L.R/, there is a condition
h.M; I /; ai 2 Pmax such that
(i) X \ M 2 M,
(ii) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(iii) .M; I / is X -iterable.
Then for each set X R with X 2 L.R/, there is a condition
h.M; I; J /; ai 2 2 Pmax
such that
(1) X \ M 2 M,
(2) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(3) .M; I; J / is X -iterable.
Proof. This is immediate by the following observation. Suppose
h.M; I /; ai 2 Pmax :
!1M
Let S
be such that S 2 M, S I , and !1M n S I . Let J 2 M be the ideal
generated by I [ S . Then h.M; I; J /; ai 2 2 Pmax . The point is that any iteration of
.M; I; J / is an iteration of .M; I /.
t
u
294
6 Pmax variations
t
u
Remark 6.10. The analysis of 2 Pmax can be carried out abstractly just assuming: For
each set X R with X 2 L.R/, there is a condition
h.M; I; J /; ai 2 2 Pmax
such that
(1) X \ M 2 M,
(2) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(3) .M; I; J / is X -iterable.
t
u
t
u
Theorem 6.13. Suppose that INS is saturated or that sat.INS / is saturated. Then INS
is semi-saturated.
Proof. Clearly we may suppose that INS is not saturated and so sat.INS / is saturated.
Suppose V G is a generic extension of V and that U 2 V G is a V -normal
ultralter on !1V .
If
U .P .!1 / n sat.INS //V
6.1
2P
max
295
S 2 .sat.INS //V :
JG D [J j h.M; I; J /; ai 2 G;
The next lemma gives the basic analysis of 2 Pmax . It shows that IG is the nonstationary
ideal, JG is a saturated ideal in L.R/G and JG D sat.IG /. This implies that the
ideal IG is presaturated in a very strong sense. Recall that a normal ideal I on !1 is
presaturated if for any sequence
hAi W i < !i
of antichains of P .!1 / n I and for any A 2 P .!1 / n I , there exists B A such that
B I and such that for each i < !,
jX 2 Ai j X \ B I j !1 :
Lemma 6.14. Assume ADL.R/ . Then 2 Pmax is !-closed and homogeneous.
Suppose G 2 Pmax is L.R/-generic. Then
L.R/G !1 -DC
and in L.R/G:
296
6 Pmax variations
6.1
2P
max
297
Proof. In L.R/G let F be the set of h.M; I /; ai 2 Pmax such that there exists an
iteration
j W .M; I / ! .M ; I /
such that j.a/ D AG and such that
I D INS \ M :
By Lemma 6.14(1), in L.R/G every club in !1 contains a club which is constructible from a real. Therefore by Theorem 3.19, if
X H.!2 /
is a countable elementary substructure then the transitive collapse of X is iterable.
Thus by Lemma 4.74, the conditions in F are pairwise compatible in Pmax .
Therefore we have only to show that
F \D ;
for all D Pmax such that D is dense and D 2 L.R/.
Suppose h.M; I; J /; ai 2 G and that D Pmax is an open, dense set with
D 2 L.R/.
Let h.M0 ; I0 /; a0 i 2 Pmax be such that
M 2 H.!1 /M0 :
Let B0 !1M0 be a set in M0 such that both B0 and !1M0 are I0 -positive. Let J0 2 M0
be the uniform normal ideal on !1M0 dened by I0 [ B0 .
By Lemma 6.6 there exists an iteration
j W .M; I; J / ! .M ; I ; J /
such that j 2 M0 , I D I0 \ M , and J D J0 \ M .
Thus h.M0 ; I0 /; j.a/i 2 Pmax . Let h.M1 ; I1 /; a1 i 2 D be a condition such that
h.M1 ; I1 /; a1 i < h.M0 ; I0 /; a0 i
and let
j0 W .M0 ; I0 / ! .M0 ; I0 /
be the iteration such that j0 .j.a// D a1 .
Thus h.M1 ; I1 ; J1 /; a1 i 2 2 Pmax and
h.M1 ; I1 ; J1 /; a1 i < h.M0 ; I0 /; a0 i
where J1 2 M1 is the (normal) ideal on !1M1 dened by I1 [ j0 .B0 /.
Note that B0 and !1M0 n B0 are I0 -positive, and so j0 .B0 / and !1M1 n j0 .B0 / are
I1 -positive.
By genericity we may suppose that
h.M1 ; I1 ; J1 /; a1 i 2 G:
But then h.M1 ; I1 /; a1 i 2 F and so
F \D ;
for all D Pmax such that D is dense and D 2 L.R/.
This proves (1).
The second claim of the lemma follows from Lemma 6.14(1).
t
u
6 Pmax variations
298
The next lemma gives the basic relationship between Pmax and 2 Pmax .
Lemma 6.16. Assume ADL.R/ . Suppose G
L.R/G:
6.1
2P
max
299
j0 ..t0 ; b0 // D .X0 ; S0 /
where
6 Pmax variations
300
J2 \ t2 D J1 \ t2 :
and let
X0 D P .!1 / \ M0
X1 D P .!1 / \ M1 :
6.1
2P
max
301
y0 D .P .!1 //M0
and let
y1 D .P .!1 //M1 :
6 Pmax variations
302
t
u
t
u
There are absoluteness theorems for 2 Pmax analogous to those for Pmax . The proofs
are straightforward modications of those for the Pmax theorems. We prove the absoluteness theorem for 2 Pmax which corresponds to Theorem 4.64. The proof is quite
similar to that of Theorem 4.64.
Of course in this theorem the ideal I could be simply the nonstationary ideal.
Theorem 6.18. Assume ADL.R/ and that there is a Woodin cardinal with a measurable above. Suppose is a 2 sentence in the language for the structure
hH.!2 /; I; J; 2i
where I J are normal uniform ideals on !1 and I J . Suppose
hH.!2 /; I; J; 2i :
Then
2P
max
:
Proof. Let .x; y/ be a 0 formula such that D 8x9y: .x; y/ (up to logical
equivalence).
Assume towards a contradiction that
hH.!2 /; IG ; JG ; 2i
2P
max
::
6.1
such that if
2P
max
303
then
304
6 Pmax variations
is a stationary subset of !1 . This follows from the observation that in V , for each set
A 2 P .!1 / n I ,
ZA P!1 .H.!2 //
is stationary in P!1 .H.!2 // where ZA is the set of countable
X H.!2 /
such that
X \ !1 2 A
and such that for all B 2 X \ I , X \ !1 B.
If B 2 I then
B \ .!1 n S /
is countable.
Similarly T is co-stationary and
.INS _ T /V G \ V D J
Since I J , it follows that
.INS _ S /V G .INS _ T /V G ;
and since I J ,
.INS _ S /V G .INS _ T /V G :
6.1
2P
max
305
where the union ranges over iterations of arbitrary length and k0 is the embedding
given by
X .
Thus N is a transitive inner model of ZFC containing the ordinals and
NX D NY :
Let IY be the image of .INS _S /V GH under the collapsing map and let JY be the
image of .INS _T /V GH . Since INS is precipitous in V GH , it follows that .INS /NX
is precipitous in NX . Therefore .INS /N is precipitous in N . Thus by Lemma 3.8 and
Lemma 3.10, the structure
.NY ; .INS /NY /
is iterable in V GH . Therefore the structure
.NY ; IY ; JY /
is iterable.
Let jY be the image of j under the collapsing map. Thus jY 2 NY ,
jY W .M0 ; I0 ; J0 / ! .M0 ; I0 ; J0 /
is an iteration of .M0 ; I0 ; J0 / of length .!1 /NY ,
I0 D IY \ M0
and
J0 D JY \ M0 :
The latter two identities hold since,
I0 D I \ M0 D .INS _ S /V GH \ M0
and
J0 D J \ M0 D .INS _ T /V GH \ M0 :
Thus h.NY ; IY ; JY /; jY .a0 /i 2 2 Pmax and
h.NY ; IY ; JY /; jY .a0 /i < h.M0 ; I0 /; a0 i:
Finally let BY be the image of B0 under the collapsing map and let DY be the
image of D0 . Thus
hH.!2 /NY ; IY ; JY ; 2i : BY ; DY
and so
hH.!2 /NY ; IY ; JY ; 2i .:8y /BY :
However BY D jY .b0 /.
Thus in V GH there is a condition
h.M1 ; I1 ; J1 /; a1 i h.M0 ; I0 ; J0 /; a0 i
such that
hH.!2 /M1 ; I1 ; J1 i 6 8y b1
where b1 D j0 .b0 / and
j0 W .M0 ; I0 ; J0 / ! .M0 ; I0 ; J0 /
is the iteration such that j0 .a0 / D a1 .
By absoluteness, noting that V is 13 -correct in the generic extension, V GH ;
t
u
such a condition h.M1 ; I1 ; J1 /; a1 i must exist in V , which is a contradiction.
306
6 Pmax variations
6.2
6.2.1 Qmax
We shall be concerned with !1 -dense, normal, uniform ideals on !1 . Recall that if
I P .!1 / is a normal, uniform, ideal, then the ideal I is !1 -dense if the set of
nonzero elements of the boolean algebra,
P .!1 /=I;
contains a dense subset of cardinality @1 . We have previously proved using the core
model induction that the existence of such an ideal on !1 implies ADR and so we
obtain the equiconsistency, see Theorem 6.149 and the subsequent corollary.
We dene our next variation on Pmax which we shall call Qmax . This we shall use
to show that it is consistent that the nonstationary ideal on !1 is !1 -dense. The rst
proof we give here will require the consistency of a huge cardinal. We do this version
rst because it is relatively easy and it illustrates the basic method which can be used
to obtain a variety of results. We then reduce the hypothesis to simply the consistency
of AD by modifying the denition of Qmax . This version, which is somewhat more
technical to dene, we shall denote by Qmax . The denition of Qmax is analogous to
.
that of Pmax
In summary, we shall dene a partial order Qmax . Assuming the existence of an
huge cardinal we shall prove that
L.R/Qmax ZFC
and
L.R/Qmax The nonstationary ideal on !1 is !1 dense:
In fact we shall prove that if there is a normal, uniform, !1 -dense ideal on !1 and
there exist innitely many Woodin cardinals with a measurable cardinal above; then
L.R/Qmax ZFC
and
L.R/Qmax The nonstationary ideal on !1 is !1 dense:
Thus we abstractly obtain the consistency that the nonstationary ideal on !1 is !1 dense from the consistency that there is an !1 -dense, normal, uniform ideal on !1
(together with the appropriate large cardinals).
After the initial analysis of L.R/Qmax we shall dene Qmax and complete the anal
ysis of L.R/Qmax assuming only ADL.R/ . Finally we shall obtain as a corollary that,
assuming only ADL.R/ ,
RO.Qmax / D RO.Qmax /:
We shall also prove several absoluteness theorems for L.R/Qmax . One of these,
(Theorem 6.85), shows that the Qmax extension is simply the Pmax -extension conditioned on a form of . Another, Theorem 6.87, is a related theorem which shows that
satises a restricted form of the homogeneity condition, formalized
the Qmax -extension
in axiom , that characterizes the Pmax -extension.
We x some notation.
307
t
u
t
u
308
6 Pmax variations
Remark 6.21.
and that
I0 D I1 \ M0 ;
and so (3) in the denition of the order on Qmax is implied by the other conditions.
(2) If we modify the denition of Qmax to require that M ZFC we obtain an equivalent forcing notion provided for each real x there exists a condition h.M; I /; f i
with x 2 M (and M ZFC). This is true under mild assumptions. For example
if AD holds in L.R/ and there is an inaccessible then it is true. Unlike the situation for Pmax , the fragment of ZFC that the models occurring in the conditions
satisfy is important insofar as what theory one needs to work in to prove existence of conditions. The underlying point is that the existence of a precipitous
ideal on !1 is weak in terms of consistency strength, whereas the existence of an
t
u
!1 -dense ideal on !1 is equiconsistent with AD.
Lemma 6.22. Suppose h.M; I /; f i 2 Qmax . Suppose
j1 W .M; I / ! .M1 ; I1 /
and
j2 W .M; I / ! .M2 ; I2 /
are iterations of .M; I / such that j1 .f / D j2 .f /. Then M1 D M2 and j1 D j2 .
Proof. This lemma is the analogue of Lemma 4.35 and the proof is a routine adaptation
of the proof of Lemma 4.35. The function f plays the role of the set a.
We rst examine an iteration
j W .M; I / ! .M ; I /
corresponding to a single generic ultrapower. We prove that j is completely determined by j.f /.!1M /. The lemma follows by induction on the length of the iteration.
Let
U .P .!1 //M
be the M-ultralter corresponding to j . Thus
U D a !1M j a 2 M and !1M 2 j.a/:
309
t
u
The next lemma is the basic iteration lemma for conditions in Qmax . Because we
wish to apply it within the models occurring in conditions we assume only the relevant
fragment of ZFC.
Lemma 6.23 (ZFC ). Suppose that I is a normal !1 -dense ideal on !1 and that
f 2 YColl .I /:
Suppose that
h.M0 ; I0 /; f0 i 2 Qmax :
Then there is an iteration k W .M0 ; I0 / ! .M ; I / such that:
(1) k.!1M0 / D !1 ;
(2) I \ M D I ;
(3) k.f0 / D f modulo I .
Proof. This lemma corresponds to Lemma 4.36, however the proof in this case is easier.
First we note that (1) and (2) follow from (3). We prove (3).
Let h.M ; I /; G ; k; W < !1 i be any iteration of .M0 ; I0 / such that for
all < !1 if
k0; .!1M0 / D
and if f ./ is an M -generic lter for Coll.!; / then k; C1 is the corresponding
generic elementary embedding.
We claim that
h.M ; I /; G ; k; W < !1 i
is as desired.
Suppose A !1 and A codes the iteration
h.M ; I /; G ; k; W < !1 i:
Then the set of < !1 such that
h.M ; I /; G ; k; W < i 2 LA \
310
6 Pmax variations
contains a club in !1 . Further the set of < !1 such that k0; .!1M0 / D also contains
a club. Because of the relationship between f and I ,
< !1 j f ./ is not a LA \ -generic lter for Coll .!; / 2 I:
This is easily veried by using the generic elementary embedding corresponding to I ,
noting that if j is the generic elementary embedding then j.A/ \ !1V D A.
Therefore
!1 n j f ./ is M -generic 2 I:
Let X !1 be the set of < !1 such that f ./ is an M -generic lter for
Coll.!; / and such that k0; .!1M0 / D . Thus !1 n X 2 I . However by the properties
of the iteration,
X j k0; C1 .f0 /./ D f ./
and so k0;!1 .f0 / D f modulo I . This proves (3).
t
u
The analysis of the Qmax -extension requires the generalization of Lemma 6.23 to
sequences. Here (unlike for 2 Pmax ) we state the general lemma cf. Lemma 6.7. The
reference in the hypothesis to conditions in Qmax is simply a device to shorten the
statement of the lemma.
Lemma 6.24 (ZFC ). Suppose I is a normal !1 -dense ideal on !1 and that
f 2 YColl .I /:
Suppose hpk W k < !i is a sequence of conditions in Qmax such that for each k < !
pk D h.Mk ; Ik /; fk i
and for all k < !
(i) pk 2 MkC1 ,
(ii) jMk jMkC1 D .!1 /MkC1 ,
Mk
(iii) !1
MkC1
D !1
(iv) fk D f0 ,
(v) IkC1 \ Mk D Ik ,
(vi) if C 2 Mk is closed and unbounded in !1M0 then there exists D 2 MkC1 such
that D C , D is closed and unbounded in C and such that
D 2 Lx
for some x 2 R \ MkC1 .
Then there is an iteration
j W h.Mk ; Ik / W k < !i ! h.Mk ; Ik / W k < !i
such that j.!1M0 / D !1 , such that
< !1 j f ./ j.f0 /./ 2 I;
and such that for all k < !,
Ik D I \ Mk :
311
6 Pmax variations
312
and
and so
pSX \ M k.A \ M /:
Similarly
However
pTX \ M g D .R n pSX / \ M g
and so
since R \ N D R \ M .
Thus in M g, the structure .N; I / is A \ M g-iterable. Finally
.N; I /
is countable in M g and
hH.!1 /M g ; A \ M g; 2i hH.!1 /; A; 2i:
Therefore .N; I / is A-iterable in V .
The set of 2 M such that
M ZFC
t
u
Lemma 6.26 can be used to obtain the existence of suitably nontrivial conditions
in Qmax .
Theorem 6.27. Suppose there is a normal !1 -dense ideal on !1 . Suppose there are
! many Woodin cardinals with a measurable above them all. Suppose X R and
that X 2 L.R/.
313
t
u
314
6 Pmax variations
315
The next theorem gives the basic results concerning Qmax . It shows that IG is an
ideal, it is !1 -dense, and it is the nonstationary ideal in L.R/G.
Theorem 6.30. Suppose for each set X R with X 2 L.R/, there is a condition
h.M; I /; f i 2 Qmax such that
(i) X \ M 2 M,
(ii) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(iii) .M; I / is X -iterable.
Then Qmax is !-closed and homogeneous. Suppose G Qmax is L.R/-generic. Then
L.R/G !1 -DC
and in L.R/G:
(1) P .!1 /G D P .!1 /;
(2) IG is a normal !1 -dense ideal on !1 ;
(3) IG is the nonstationary ideal.
Proof. The only claim here that does does not have a counterpart in the Pmax case is
(2).
The other claims are proved by simply adapting the proofs of the corresponding
claims for the Pmax -extension. As in the case for 2 Pmax , the proof that Qmax is !-closed
requires proving Lemma 6.23 for the sequences that arise. Again the sequences satisfy
the conditions of Corollary 4.20 and so are iterable in the sense of Denition 4.8.
To prove (2) we work in L.R/G. Suppose
A 2 P .!1 / n IG :
By (1) and the denitions, there exists h.M; I /; f i 2 G such that
A 2 M n I ;
where .M ; I / is the image of .M; I / under the iteration which sends f to fG .
Therefore by the properties of .I ; fG / in M , there exists p 2 Coll.!; !1 / such
that
< !1 j p 2 fG ./ I
6 Pmax variations
316
.!1 n A/ \ < !1 j p 2 fG ./ 2 I :
However I D M \ IG .
Therefore for each set
S 2 P .!1 / n IG
there exists p 2 Coll.!; !1 / such that
< !1 j p 2 fG ./ n S 2 IG :
t
u
There remains the question of whether the axiom of choice holds in L.R/Qmax . We
show that it does and in fact for the same reason it holds in L.R/Pmax . Recall that AC
is the 2 sentence for
hH.!2 /; 2i
which we used to show that AC holds in L.R/Pmax .
Theorem 6.31. Suppose for each set X R with X 2 L.R/, there is a condition
h.M; I /; f i 2 Qmax such that
(i) X \ M 2 M,
(ii) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(iii) .M; I / is X -iterable.
Then
Qmax
hH.!2 /; 2iL.R/
AC :
Proof. The proof is a minor modication of the proof that AC holds in L.R/Pmax .
Fix G Qmax such that G is L.R/-generic.
Suppose hSi W i < !i and hTi W i < !i are sequences of pairwise disjoint subsets
of !1 . Suppose the Si are stationary and suppose that
!1 D [Ti j i < !:
Let h.M; I /; f i 2 G be such that hSi W i < !i; hTi W i < !i 2 M where
j W .M; I / ! .M ; I /
is the iteration such that j.f / D fG .
Let hsi W i < !i; hti W i < !i in M be such that
j..hsi W i < !i; hti W i < !i// D .hSi W i < !i; hTi W i < !i/:
Thus in M, hsi W i < !i and hti W i < !i are sequences of pairwise disjoint subsets of
!1M , the si are not in I , and
!1M D [ti j i < !:
317
Let D be the set of conditions h.N ; J /; gi < h.M; I /; f i such that in N there
exists < !2N and a continuous increasing function F W !1N ! with conal range
such that
F .tiN / sQiN
for each i < ! where tiN D k.ti /, siN D k.si / and k is the embedding of the iteration
of .M; I / which sends f to g. For each i < !, sQiN denotes the set AQ as computed in
N where A D siN .
It sufces to show that D is dense below h.M; I /; f i.
We show something slightly stronger. Suppose
h.N ; J /; gi < h.N0 ; J0 /; g0 i < h.M; I /; f i:
Then for some h 2 N , h.N ; J /; hi 2 D and
h.N ; J /; hi < h.N0 ; J0 /; g0 i < h.M; I /; f i:
si0
Let be the image of si under the iteration of .M; I / which sends f to g0 and let
ti0 be the image of ti under this iteration.
Let x 2 N be a real which codes N0 .
Working in N we dene an iteration of .N0 ; J0 / of length !1N . Let C be the set of
indiscernibles of Lx less than !1N . Let D C be the set of 2 C such that C \
has ordertype . Thus D is a closed unbounded subset of C . Let
h.N ; J /; G ; j; W < !1N i
be an iteration of .N0 ; J0 / in N such that
(1.1) For all 2 D and for all < , j0; .si0 / 2 G if 2 j0; .ti0 / where is the
th element of C above ,
(1.2) J! N D J \ N! N .
1
Condition (1.3) is the additional requirement special to the case of Qmax . The condition
(1.3) is satised by constructing the iteration using g. / to dene the generic ultralter
at stage whenever possible provided 2 D.
The iteration is easily constructed in N , the point is that the requirements given by
(1.1), (1.2), and (1.3) do not interfere. The other useful observation is that if 2 C
and if k W .N0 ; J0 / ! .N0 ; J0 / is any iteration of length then D k.!1N0 /.
Let be the th indiscernible of Lx where D !1N C!1N . Let F be the function
F W !1N ! given by F ./ is the th indiscernible of Lx where D !1N C . Thus
(2.1) 2 N ,
(2.2) < !2N ,
(2.3) F 2 N ,
(2.4) F W !1N ! is continuous and strictly increasing.
6 Pmax variations
318
Let siN D j0;! N .si0 / and let tiN D j0;! N .ti0 /. Let
1
h D j0;! N .g0 /:
1
Thus h.N ; J /; hi 2 Qmax and h.N ; J /; hi < h.N0 ; J0 /; g0 i. By the denition of the
iteration it follows that in N ,
F .tiN / sQiN
and so h.N ; J /; hi 2 D.
t
u
t
u
Qmax
:
Proof. Let be the supremum of the rst ! Woodin cardinals and let be the least
strongly inaccessible cardinal above .
Since there is a measurable cardinal above , every set
A 2 P .R/ \ L.R/
is <-weakly homogeneously Suslin.
Therefore by Theorem 6.27, for each set X R with X 2 L.R/, there is a
condition h.M; I /; f i 2 Qmax such that
319
(1.1) X \ M 2 M,
(1.2) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(1.3) .M; I / is X -iterable.
Thus the basic analysis of Qmax as given in Theorem 6.30 and Theorem 6.31 applies.
Suppose X V is a countable elementary substructure containing I and f . Let
hMX ; JX ; FX i be the image of hX; J; F i under the transitive collapse. Then .MX ; JX /
is iterable. Further for each A 2 X \ P .R/ \ L.R/, .MX ; JX / is A-iterable.
Thus hMX ; JX ; FX i 2 Qmax . The theorem follows from an argument similar to the
absoluteness theorem for Pmax . The situation here is actually simpler since no forcing
over V is required. The only other relevant point is that if h.M; I /; f i 2 Qmax then
there is an iteration
j W .M; I / ! .M ; I /
such that j.!1M / D !1 and such that
j j.f /./ F ./ 2 J:
t
u
As with the case for Pmax this expanded absoluteness theorem has a converse.
This requires two preliminary lemmas the rst of which is a very weak analogue of
Lemma 4.60.
Lemma 6.34. Assume that for each set X R with X 2 L.R/, there is a condition
h.M; I /; f i 2 Qmax such that
(i) X \ M 2 M,
(ii) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(iii) .M; I / is X -iterable.
Suppose that G Qmax is L.R/-generic. Suppose h 2 L.R/G and that in L.R/G;
h is a function such that
h W !1 ! H.!1 /
and such that the set
j fG ./ D h./
contains a club in !1 . Then there is a lter G Qmax such that G is L.R/-generic,
fG D h and such that
L.R/G D L.R/G :
320
6 Pmax variations
t
u
Lemma 6.35. Assume that for each set X R with X 2 L.R/, there is a condition
h.M; I /; f i 2 Qmax such that
(i) X \ M 2 M,
(ii) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(iii) .M; I / is X -iterable.
Suppose that G Qmax is L.R/-generic. Then
L.R/G D L.R/fG :
321
Proof. Let F be the set of h.M; I /; f i 2 Qmax such that there exists an iteration
j W .M; I / ! .M ; I /
such that j.f / D fG . The iteration j is unique and
j 2 L.fG ; h.M; I /; f i/:
Thus F 2 L.R/fg and G F .
It remains to show that the conditions in F are pairwise compatible in Qmax . This
follows easily from Theorem 6.30(1).
Suppose h.M0 ; I0 /; f0 i 2 F and that h.M1 ; I1 /; f1 i 2 F . Then by Theorem 6.30(1) there exists h.M; I /; f i 2 G such that
.M0 ; M1 / 2 H.!1 /M
and such that there exist iterations
k0 W .M0 ; I0 / ! .MO 0 ; IO0 /
and
t
u
Theorem 6.36. Suppose there are ! many Woodin cardinals with a measurable above
them all. Suppose I is a normal !1 -dense ideal on !1 and that
F 2 YColl .I /:
Suppose that for each 2 sentence, , in the language for the structure
hH.!2 /; I; F I ; X; 2 W X R; X 2 L.R/i
if
Qmax
hH.!2 /; I; F I ; X; 2 W X R; X 2 L.R/i :
Then there exists G Qmax such that:
(1) G is L.R/-generic;
(2) P .!1 / L.R/G;
(3) fG D F ;
(4) IG D I .
322
6 Pmax variations
t
u
323
The absoluteness theorems suggest that in the model L.R/Qmax one should have
all the consequences of the largest fragment of Martins Maximum which is consistent
with the existence of an !1 -dense ideal on !1 . This seems to generally be the case and
we shall prove a theorem along these lines. However in Section 6.2.4 we shall prove
that there is a weak Kurepa tree on !1 in L.R/Qmax . By varying the order on Qmax we
shall also produce a model in which the nonstationary ideal on !1 is !1 -dense and in
which there are no weak Kurepa trees, this is the subject of Section 6.2.5.
Many of the combinatorial consequences of the existence of an !1 -dense ideal can
be factored through a variant of . We dene three such variants, the rst is .!1<! /
which is easily seen to follow from .
Denition 6.37. .!1<! /: There is a function
f W !1 ! H.!1 /
such that for all sequences hD W < !1 i of dense subsets of Coll.!; !1 /,
< !1 j f ./ is a lter in Coll.!; / and f ./ \ D ; for all <
t
u
is stationary in !1 .
Denition 6.38. + .!1<! /: There is a function
f W !1 ! H.!1 /
such that for all dense sets,
D Coll.!; !1 /;
the set
< !1 j f ./ is a lter in Coll.!; / and f ./ \ D ;
contains a closed unbounded subset of !1 .
t
u
The next two lemmas give useful reformulations of .!1<! / and + .!1<! /.
Lemma 6.39. Suppose f W !1 ! H.!1 / is a function witnessing .!1<! / and that
M is a transitive set such that
H.!2 / M:
Then
X 2 P!1 .M / j f .X \ !1 / is MX -generic for Coll.!; X \ !1 /
is stationary in P!1 .M / where for each X 2 P!1 .M /, MX denotes the transitive
collapse of X .
Proof. Clearly we may suppose that M D H.!2 / since
P .Coll.!; !1 // H.!2 /:
Fix a function
H W H.!2 /<! ! H.!2 /:
324
6 Pmax variations
!1 N
and such that H N N . Clearly N is transitive.
Let hD W < !1 i enumerate the dense subsets of Coll.!; !1 / which belong to
N.
Let
S D < !1 j f ./ \ D ; for all < :
<!
325
t
u
326
6 Pmax variations
327
c2 \ . C 1/ D c1
where D [c1 .
Suppose is an ordinal and f W !1 ! H.!1 / is a function. Let P .f; / denote
the countable support iteration where for all < ,
P .f; C 1/ D P .f; / P .f /
and P .f / is as computed in V P .f;/ .
We note that P .f; / is not in general a semiproper partial order.
Lemma 6.44. Suppose that for all A !1 , A# exists, and that
f W !1 ! H.!1 /
is a function which witnesses .!1<! /. Suppose that is strongly inaccessible. Then
(1) P .f; / is .!; 1/-distributive,
(2) f witnesses ++ .!1<! / in V P .f;/ .
Proof. The partial order P .f; / is -cc and so (2) follows from (1) using the standard
analysis of terms and the denition of P .f; /.
We prove (1). Fix p 2 P .f; / and suppose hDk W k < !i is a sequence of dense
subsets of P .f; /. Let D C !.
By Lemma 6.43, the set,
X VC1 j f .X \ !1 / is L.MX /-generic for Coll.!; X \ !1 /;
is stationary in P!1 .VC1 / where for each X 2 P!1 .VC1 /, MX is the transitive
collapse of X . Note that if X VC1 then 2 X . Therefore P .f; / 2 X if f 2 X .
Therefore there exists X VC1 such that
(1.1) X is countable,
(1.2) p 2 X ,
(1.3) hDk W k < !i 2 X ,
(1.4) f .X \ !1 / is L.MX /-generic for Coll.!; X \ !1 /.
328
6 Pmax variations
t
u
t
u
329
330
6 Pmax variations
Proof. Let
f W !1 ! H.!1 /
be a function which witnesses .!1<! /. We may assume that for all < !1 , f ./ is a
lter in Coll.!; /.
(1) is immediate, in fact one can show that
R \ L.f /
is not meager. To see this suppose that N is a countable transitive set. Then since f
witnesses .!1<! / there exists a countable elementary substructure
X H.!2 /
such that N 2 X and such that f ./ is MX -generic for Coll.!; X \ !1 / where MX is
the transitive collapse of X .
Let D X \ !1 and let g W ! ! be the generic map given by f ./. Dene
c W ! ! 0;1 by c.i / D 0 if g.i / D 0 and c.i / D 1 if g.i / 0. Then c 2 L.f / and
c is Cohen generic over N .
Therefore R \ L.f / is not meager. This proves (1).
The proof of (2) is an easy modication of the standard construction of a Suslin
tree using .
Suppose .T; <T / is a tree. A branch is a maximal chain. A branch is rank conal
if the ranks of elements of the branch are conal in the rank of T . The tree is uniform
if for all a 2 T the ranks of b 2 T such that a <T b are conal in the rank of T .
Suppose <T is a partial order on such that .; <T / is a uniform tree and suppose
h W ! ! is a surjection.
The function h denes naturally a branch of .; <T / as follows. Let
hni W i < !i
be the sequence dened by n0 D 0 and niC1 is the least k < ! such that
h.ni / <T h.k/. Thus h.ni / j i < ! is chain in .; <T / which denes (uniquely) a
branch. If h is sufciently generic then this branch is rank conal.
Suppose g Coll.!; / is a lter with sufcient genericity to dene a function
hg W ! ! :
By xing a (recursive) bijection W ! ! ! ! we can dene from g a sequence
hhgi W i < !i of functions from ! ! where hgi .j / D hg ..i; j //. If g is sufciently
generic then
(1.1) for each i < !, the branch dened by hgi is rank conal in .; <T /,
(1.2) the union of the branches dened by the functions hgi contains every < .
Let <T be any order on !1 such that .!1 ; <T / is a normal, uniform, tree with
countable levels and such that for all countable elementary substructures
X H.!2 /;
331
332
6 Pmax variations
Proof. It sufces to show that there must exist a normal !1 -dense ideal on !1 . This is
a relatively standard fact.
Let I be a uniform, countably complete, !1 -dense ideal on !1 . Then
P .!1 /=I RO.Coll.!; !1 //:
Therefore I denes a boolean-valued elementary embedding
j WV !M VB
where B D RO.Coll.!; !1 //.
Dene
W P .!1 / ! B
V
by .A/ D [[!1 2 j.A/]].
Let I0 be the set of A such that .A/ D 0.
Thus I0 is a normal saturated ideal on !1 and induces a boolean isomorphism of
P .!1 /=I0 with a complete subalgebra of B.
It follows that I0 is a normal !1 -dense ideal.
t
u
Lemma 6.52 (ZFC ). Suppose I is a normal !1 -dense ideal on !1 . Let
f W !1 ! H.!1 /
be such that f induces a boolean isomorphism
W P .!1 /=I ! RO.Coll.!; !1 //:
Suppose c is Cohen generic over V and in V c let I.c/ be the normal ideal generated
by I . Then in V c, I.c/ is normal !1 -dense ideal on !1 and f induces a boolean
isomorphism
.c/ W P .!1 /=I ! RO.Coll.!; !1 //:
Proof. Suppose G Coll.!; !1 / is V c-generic. Then G is certainly V -generic and
so there exists a generic elementary embedding
j W V ! M V G
V
such that j.f /.!1 / D G.
Since c is Cohen generic over V G and since Cohen forcing is ccc, the embedding
j lifts to a generic elementary embedding
j W V c ! M c V Gc:
The induced ideal is easily veried to be I.c/ . The generic elementary embedding j
shows that I.c/ and f have the desired properties.
t
u
Theorem 6.53. Assume that for each set X R with X 2 L.R/, there is a condition
h.M; I /; f i 2 Qmax such that
(i) X \ M 2 M,
(ii) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(iii) .M; I / is X -iterable.
Then the following hold in L.R/Qmax .
(1) Every set of reals of cardinality !1 is of measure 0.
(2) The reals cannot be decomposed as an !1 union of meager sets.
333
.R/LA D R \ .LA/M1 :
Thus c is Cohen generic over LA.
t
u
6 Pmax variations
334
6.2.2 Qmax
We dene the variant of Qmax for which the analysis can be carried out assuming
just ADL.R/ . The modication is obtained by replacing the model M in a condition
with an ! sequence of models. With this we can improve Theorem 6.32 to obtain the
consistency of
ZFC C The nonstationary ideal on !1 is !1 dense
from simply the consistency of ZF C AD. This is best possible.
The denition of Qmax is motivated by the proof that Qmax is !-closed and the
denition is closely related both to Denition 4.15 and to the denition of Pmax
. In fact
there is a dense subset of Qmax which is a suborder of Pmax .
Suppose hpk W k < !i is a sequence of conditions in Qmax such that for all k < !,
pkC1 < pk . Suppose that for each k < !,
pk D h.Mk ; Ik /; fk i:
Let
G [P ./ \ Mk j k < !
is a lter such that G \ Mk is Mk -normal for all k < !, then for all k < !, G \ Mk
is Mk -generic. The same point applies to iterates of hMk W k < !i.
Therefore hMk W k < !i is iterable.
Denition 6.54. Qmax is the set of pairs .hMk W k < !i; f / such that the following
hold.
(1) f 2 M0 and
f W !1M0 ! M0
(3) Mk 2 MkC1 ; !1
(4) .INS /
MkC1
MkC1
D !1
\ Mk D .INS /
MkC2
335
\ Mk .
p 2 Coll.!; !1M0 /
t
u
As in the denition of the order on Qmax , clause (3) in the denition of the order on
Qmax follows from clauses (1) and (2).
The next lemma claries the effect of (6) and (7) in Denition 6.54.
Lemma 6.55. Suppose that .hMk W k < !i; f / 2 Qmax .
(1) Suppose that
g Coll.!; !1M0 /
6 Pmax variations
336
t
u
Denition 6.56. Suppose .hMk W k < !i; f / 2 Qmax and suppose X R. Then
hMk W k < !i is X -iterable if
(1) X \ M0 2 [Mk j k < !,
(2) for any iteration
j W hMk W k < !i ! hNk W k < !i
of hMk W k < !i, j.X \ M0 / D X \ N 0 .
t
u
If Qmax is sufciently nontrivial then Qmax and Qmax are equivalent as forcing
notions. More precisely if for every real x there exists a condition h.M; I /; f i 2 Qmax
such that x 2 M and such that
I D .INS /M
then
RO.Qmax / RO.Qmax /:
The proof of this is implicit in what follows.
We shall need a slight variant of iterability.
Denition 6.57. Let A R and .M; I; / 2 H.!1 / be such that
(i) M is a transitive model of ZFC,
(ii) is a Woodin cardinal in M and
I D .I< /M
is directed system of ideals associated to .Q< /M ,
(iii) .M; I/ is iterable,
(iv) A \ M 2 M and hH.!1 /M ; A \ M; 2i hH.!1 /; A; 2i.
The pair .M; I/ is strongly A-iterable if for all countable iterations
j W .M; I/ ! .M ; I /I
(1) j.A \ M / D A \ M ,
t
u
337
t
u
338
6 Pmax variations
Proof. This is a special case of the general theorem for producing Suslin representations from various forms of generic absoluteness and correctness in the context of a
Woodin cardinal. We briey sketch the argument which involves elementary aspects
of the stationary tower.
The key point is the following. Let R0 be the set of triples .P ; ; p/ such that
(1.1) P 2 M is a partial order,
(1.2) 2 MP is a term for a real,
(1.3) for all M -generic lters
g P;
if p 2 g then Ig . / A, where Ig ./ is the interpretation of by g.
Similarly let R1 be the set of triples .P ; ; p/ such that
(2.1) P 2 M is a partial order,
(2.2) 2 MP is a term for a real,
(2.3) for all M -generic lters
g P;
if p 2 g then Ig . / 2 A.
Then for each partial order P 2 M and for each term 2 MP ,
p j .P ; ; p/ 2 R0 [ R1
is dense in P . Further .R0 ; R1 / 2 M .
The verication is a routine consequence of A-iterability noting that if P 2 M is
a partial order and if g P is an M -generic lter then there exists an M -generic lter
h .Q< /M
such that g 2 M h.
We now work in M . Let AM D A \ M and x R0 and R1 as specied above.
Fix a strongly inaccessible cardinal, , of M which is below . A countable elementary substructure
X M
is AM -good if for each partial order P 2 X the following holds. Suppose 2 X \ M P
is a term for a real and that
g X \P
is an X -generic lter; i. e. g is a lter such that if D P is a dense set such that
D 2 X then
g \ D ;:
Let x 2 R be the interpretation of by g. Then
x 2 AM
339
j W .M; / ! .M ; /
is M -generic. Then
a) A \ M G D pj.T / \ M G,
b) hH.!1 /M
G
6 Pmax variations
340
Proof. For each n 2 ! let An be the set of x 2 R which code an element of the n
diagram of the structure
hV!C1 ; A; 2i:
Thus for each n 2 !, .M; I/ is strongly An -iterable.
For each n 2 ! let Tn 2 M be a tree on ! such that for all M -generic lters,
g Coll.!; </;
pTn \ M g D An \ M g.
The trees Tn exist by Lemma 6.59.
We rst prove that for any countable iteration
k W .M;
/ ! .M ;
/;
for each n 2 !,
pk.Tn / An :
Q
Q
\k.a/
j a 2
2 k.
/:
Q n /:
pk.Tn / pk.T
Finally k.Tn / is countable in MQ and
hMQ \ V!C1 ; A \ MQ ; 2i hV!C1 ; A; 2i:
Therefore if
pk.Tn / 6 An
then there must exists
x 2 pk.Tn / n An
Q n / and this contradicts that
Q
such that x 2 M . But then x 2 pk.T
Q n \ M / D An \ MQ :
k.A
This proves that for each n,
pk.Tn / An :
Finally let
j W .M; / ! .M ; /
341
be M -generic.
Let A D pj.T / \ M G and for each n < ! let An D pj.Tn / \ M G.
By the elementarity of j it follows that in M G, for each n < !, the set An is
the set of x 2 R which code an element of the n -diagram of
hV!1 \ M G; A ; 2i:
Further A A and for each n < !, An An . Therefore
A D A \ M G
and
t
u
t
u
We now come to the main theorem for the existence of conditions in Qmax .
First we x some notation and prove an easy preliminary lemma. This lemma is
really the technical key for producing conditions in Qmax from our assumptions.
Suppose S Ord is a set of ordinals. Then Coll .!; S / is the partial order of
nite partial functions p W ! S ! Ord such that p.i; / < 1 C . The order is by
extension and so Coll .!; S / is the natural restriction of the Levy collapse.
1
1
We also x some coding. A partial function f W H.!1 / ! H.!1 / is
1 if it is
1
1
in the codes. More precisely f is
1 if the set
x 2 R j x codes .a; b/ and b D f .a/
is
1.
342
6 Pmax variations
343
Let .M; I/ be strongly X -iterable with .z; t / 2 M . It follows that .M; I/ is strongly
Y -iterable.
For every real z there exists .M; I/ such that .M; I/ is strongly X -iterable and such
that z 2 M . In particular, .M; I/ is iterable and so for every z 2 R, z # exists.
We next prove that every subset of !1 which is coded by a set projective in X is
constructible from a real.
Suppose A !1 be such that A is coded by a set which is projective in X . Let
Y R be the set of reals which code elements of A. Therefore Y is projective in X .
Let .M; I/ be such that .M; I/ is strongly Y -iterable. Let z be a real which codes M .
Thus by absoluteness it follows that A 2 Lz.
We now prove the theorem. Fix X R. We are assuming that there exists .M; I/
such that .M; I/ is strongly X -iterable.
We dene sequences hNk ; fk ; Ck ; xk I k < !i and hMk ;
k ;
k I k < !i as follows.
Set C0 D !1 . Choose
.M; I/ 2 H.!1 /
such that .M; I/ is strongly X -iterable and let be the Woodin cardinal of M associated to I.
Let T0 2 M be a tree on ! such that for all M-generic lters
g Coll.!; </;
pT0 \ Mg D X \ Mg. The existence of the tree T0 follows from Lemma 6.59.
Let M0 D M and let
0 2 M be a normal measure on
0 < . Let N0 be the
image of .M0 /0 under the !1th iteration of .M0 ;
0 /. Let C1 be the critical sequence
of this iteration. Thus
(1.1) N0 ZFC,
(1.2) Ord \ N0 D !1 ,
(1.3) C1 is a club in !1 consisting of inaccessible cardinals of N0 .
Further for any < with ; 2 C1 there exists a canonical elementary embedding
j W N0 ! N 0
1
such that cp.j / D and j./ D . Let y0 2 R be an index for a
1 function f0 with
the following property. If .; g; h/ is such that;
(2.1) 2 C1 ,
(2.2) g is N0 -generic for Coll .!; </,
(2.3) h is N0 g-generic for Coll .!; /,
then f0 .; g; h/ is an N0 hg-generic for Coll .!; S / where S is the interval
.; / and is the least element of C1 above . Let x0 be a real such that
N0 ; M0 ; y0 ; C1 2 Lx0 with M0 countable in Lx0 .
344
6 Pmax variations
1
(5.1) 2 CkC2 ,
(5.2) g is NkC1 -generic for Coll .!; CkC1 \ /,
(5.3) h is NkC1 g-generic for Coll .!; /,
then fkC1 .; g; h/ is an NkC1 hg-generic for Coll .!; .; / \ CkC1 / where is
the least element of CkC2 above . Let xkC1 be a real such that
NkC1 ; MkC1 ; ykC1 ; CkC2 LxkC1
and such that MkC1 countable in LxkC1 .
345
6 Pmax variations
346
For each k < ! let k be the least element of CkC1 above 0 . Thus k is strongly
inaccessible in Nk and 1 D sup k j k < !.
Since 1 D sup k j k < !, for all k < !, Gk is Nk -generic. This follows by an
argument similar to that for the genericity of Gk .
We now come to the key points. For each k < ! let
jk W Nk ! N k
be the canonical embedding with critical point 0 and such that jk . 0 / D 1 . Thus for
all k < m < !;
(9.1) jm .Nk / D Nk and jm .Ck / D Ck ,
(9.2) jm jNk D jk ,
(9.3) jk j.Nk /k 2 Nm ,
where as above jm .Nk / and jm .Ck / are dened in the obvious way.
For each k < ! the embedding jk lifts to dene an embedding
jk W Nk Gk ! Nk Gk :
It follows that for all k < m < !;
(10.1) jm jNk Gk D jk ,
(10.2) jk j.Nk Gk /k 2 Nm Gm G.
For each k < ! let Uk be the Nk ultralter on 0 which is the image of
k under
the iteration of .Mk ;
k / which sends
k to 0 . It is straightforward to show that for
all k < !, Uk 2 NkC1 and that Uk D F \ Nk where F is the club lter on 0 as
computed in NkC1 .
The ultrapower of h.Nk ; Uk / W k < !i is dened as follows. Let
U D [Uk j k < !
and for each k < ! let
where
Nk D Nk0 =U
Nk0 D h W 0 ! Nk j h 2 [Ni
Let h.Nk ; Uk / W k < !i be the ultrapower of h.Nk ; Uk / W k < !i and let
j W [Nk j k < ! ! [Nk j k < !
be the induced embedding. Thus j is a 0 -embedding whose restriction to Nk is fully
elementary for each k < !. It follows that for each k < !, j jNk D jk .
347
Iterations of h.Nk ; Uk / W k < !i are dened in the natural fashion. As in the case
of iterating !-sequences of models (see Denition 4.8) the embeddings that arise
j W [Nk j k < ! ! [Nk j k < !
are 0 elementary embeddings whose restrictions to Nk are fully elementary.
It is easy to verify that h.Nk ; Uk / W k < !i is iterable and in fact for all k < !, Nk
is the image of Nk under any countable iteration of h.Nk ; Uk / W k < !i.
For each k < ! let Ik D INS \ Nk Gk where INS is the nonstationary ideal on
!1 .D 0 / as computed in NkC1 GkC1 .
Thus for all k < !:
(11.1) Nk Gk ZFC;
N Gk
D !1 kC1
GkC1
D 0 ;
t
u
348
6 Pmax variations
t
u
As a corollary to Theorem 6.64 we obtain Lemma 6.68 which in some weak sense
corresponds to Lemma 6.47. As we have already noted, Lemma 6.47 cannot be proved
just assuming ADL.R/ . The basic method for proving Lemma 6.68 can be used to
prove many similar results, it is also related to additional absoluteness theorems we
shall prove for Qmax cf. Theorem 6.85.
We need two preliminary lemmas. The rst is a corollary of Lemma 6.40.
Lemma 6.66. Suppose .M; I/ is a countable iterable structure such that
(i) M ZFC,
(ii) I 2 M and I is the tower of ideals I< as computed in M where is a Woodin
cardinal in M .
Suppose f 2 M , f witnesses + .!1<! / in M and for all p 2 .!1<! /M the set
< !1M j p 2 f ./
is stationary within M . Suppose g .!1<! /M is M -generic. Then there exists
G QM
<
such that G is M -generic and such that j.f /.!1M / D g where
j W M ! N M G
is the generic elementary embedding corresponding to G.
Proof. Suppose G QM
is M -generic and let
<
j W M ! N M G
be the generic elementary embedding corresponding to G. By Lemma 6.40, it follows
that j.f /.!1 / is M -generic for Coll.!; !1M /.
For each p 2 Coll.!; !1M / let
Sp D < !1M j p 2 f ./:
The set Sp is stationary within M and so Sp 2 QM
. If Sp 2 G then
<
p 2 j.f /.!1M /:
The lemma follows by the denability of forcing.
t
u
349
350
6 Pmax variations
M g
B0 \ MZ g 2 WH Z ;
C MZ
/ .
where D .Z
C
Thus in MZ g every set projective in A \ MZ g is Z
-weakly homogeneously
Suslin.
Let 2 MZ g be the least strongly inaccessible cardinal above Z .
By standard arguments, Z is a Woodin cardinal in MZ g.
Let Y .MZ g/ be an elementary substructure such that A \ MZ g 2 Y ,
Y 2 MZ g and Y is countable in MZ g.
Let N be the transitive collapse of Y and let I be the image of .I<Z /MZ g under
the collapsing map. Thus
N ZFC C C ++ .!1<! /;
and
hV!C1 \ N; A \ N; 2i hV!C1 ; A; 2i:
By Lemma 5.23, the structure .N; I/ is A \ MZ g iterable in MZ g.
However
hV!C1 \ MZ g; B0 \ MZ g; 2i hV!C1 ; B0 ; 2i
and N is countable in MZ g. Therefore the structure .N; I/ is A-iterable in V .
Lemma 6.68. Assume AD holds in L.R/. Then for every set
A 2 P .R/ \ L.R/;
there is a condition .hMk W k < !i; f / 2 Qmax such that
(1) A \ M0 2 M0 ,
(2) hH.!1 /M0 ; A \ M0 ; 2i hH.!1 /; A; 2i,
(3) hMk W k < !i is A-iterable,
(4) holds in M0 ,
(5) f witnesses ++ .!1<! / in M0 .
t
u
351
j W .N; I/ ! .N ; I /
6 Pmax variations
352
Since N 2 M0 ,
and is as required.
With Theorem 6.65 the analysis of Qmax can easily be carried out as in the case of
Qmax . We summarize the results of this in the next two theorems.
First we prove the main iteration lemmas for conditions in Qmax .
Lemma 6.69. Suppose .hMk W k < !i; f / 2 Qmax . Suppose
j W hMk W k < !i ! hMk W k < !i
and
t
u
353
(2) for all k < !, I \ Mk D INS \ Mk D .INS /MkC1 \ Mk ;
(3) j.f / D h modulo I .
Proof. This lemma corresponds to Lemma 6.23.
Let hhMk W k < !i; G ; j; W < !1 i be any iteration of hMk W k < !i
such that for all < !1 if
j0; .!1M0 / D
and if h./ is an [Mk j k < !-generic lter for Coll.!; / then j; C1 is the
corresponding generic elementary embedding.
We claim that
hhMk W k < !i; G ; j; W < !1 i
is as desired.
Suppose A !1 . By assumption
< !1 j h./ is not a LA \ -generic lter for Coll.!; / 2 I:
Suppose A !1 and A codes the iteration
hhMk W k < !i; G ; j; W < !1 i
Then the set of < !1 such that
hhMk W k < !i; G ; j; W < i 2 LA \
contains a club in !1 . Further the set of < !1 such that
j0; .!1M0 / D
also contains a club.
Let X !1 be the set of < !1 such that h./ is an [Mk j k < !-generic
lter for Coll.!; / and such that j0; .!1M0 / D . Thus !1 n X 2 I . However by the
properties of the iteration,
X j j0; C1 .f /./ D h./
and so j0;!1 .f / D h modulo I .
t
u
354
6 Pmax variations
p 2 .H.!1 //N0 :
I .INS /N1 ;
(1.2) for all k < !, I \ Mk D .INS /N0 \ Mk D .INS /MkC1 \ Mk ;
(1.3) j.f / D g modulo I .
Let h D j.f /. Thus .hNk W k < !i; h/ 2 Qmax and
.hNk W k < !i; h/ < p:
t
u
As a corollary to Lemma 6.71 and Lemma 6.68, we obtain the set of conditions
specied in Lemma 6.68 is dense in Qmax .
355
Qmax .
such that:
(1) A \ N0 2 N0 ;
(2) hH.!1 /N0 ; A \ N0 ; 2i hH.!1 /; A; 2i;
(3) hNk W k < !i is A-iterable;
(4) holds in N0 ;
(5) g witnesses ++ .!1<! / in N0 ;
(6) .hNk W k < !i; g/ < p.
Proof. Let B R be the set of x 2 R such that x codes a pair .y; p/ where y 2 A.
By Lemma 6.68, there exists
.hNk W k < !i; g0 / 2 Qmax
such that:
(1.1) B \ N0 2 N0 ;
(1.2) hH.!1 /N0 ; B \ N0 ; 2i hH.!1 /; B; 2i;
(1.3) hNk W k < !i is B-iterable;
(1.4) holds in N0 ;
(1.5) g0 witnesses ++ .!1<! / in N0 .
By (1.2), p 2 .Qmax /N0 . By Lemma 6.71, there exists g 2 N0 such that
(2.1) .hNk W k < !i; g/ 2 Qmax ,
(2.2) .hNk W k < !i; g/ < p,
(2.3) < !1N0 j ./ g0 ./ 2 .INS /N1 .
Thus .hNk W k < !i; g/ is as desired.
Theorem 6.73. Assume AD holds in L.R/. Then Qmax is !-closed.
t
u
356
6 Pmax variations
Proof. This is one theorem about Qmax that is actually much simpler than the corresponding theorem about Qmax or Pmax .
The !-closure of Qmax is essentially built into its denition.
Suppose hpi W i < !i is a strictly decreasing sequence of conditions in Qmax and
that for each i < !,
pi D .hMki W k < !i; fi /:
Let fO D [fi j i < !. For each i < ! let
ji W hMki W k < !i ! hMO ki W k < !i
be the iteration such that ji .fi / D fO. This iteration exists since hpi W i < !i is a
strictly decreasing sequence in Qmax .
By Lemma 4.22, hMO kk W k < !i satises the hypothesis of Lemma 4.17 and so by
Lemma 4.17, hMO kk W k < !i is iterable.
Let
O0
M
Mi
D !1 0 D sup !1 0 j k < !:
For q 2 Coll.!; / let
Sq D < j q 2 fO./:
Then for each q 2 Coll.!; /,
Sq 2 MO 00 ;
and further
Ok
Sq .INS /Mk
for each k 2 !.
Fix k 2 !. Suppose that A and that
O kC1
A 2 MO kk n .INS /MkC1 :
Then for some q 2 Coll.!; /,
O kC1
Sq n A 2 .INS /MkC1 :
Finally if q1 q2 in Coll.!; / then Sq1 Sq2 .
It follows that .hMO kk W k < !i; fO/ 2 Qmax . Further if q 2 Qmax and
q < .hMO kk W k < !i; fO/;
then q < pi for all i < !.
By Corollary 6.72 there exists q 2 Qmax such that
q < .hMO kk W k < !i; fO/;
and so there exists q 2 Qmax such that
q < pi
for all i < !.
t
u
357
The basic analysis of Qmax is straightforward, the results are given in the next two
theorems.
Theorem 6.74. Assume AD holds in L.R/. Suppose
G Qmax
is L.R/-generic. Then
L.R/G !1 -DC
and in L.R/G:
(1) P .!1 /G D P .!1 /;
(2) IG is a normal !1 -dense ideal on !1 ;
(3) IG is the nonstationary ideal.
Proof. The proof is essentially the same as the proof of Theorem 6.30. Here one uses
Lemma 6.70 and the proof of Theorem 6.73.
t
u
Theorem 6.75. Assume AD holds in L.R/. Then
L.R/Qmax AC :
Proof. The proof for Qmax naturally generalizes.
t
u
t
u
358
6 Pmax variations
As a corollary to the basic analysis of Qmax we obtain Theorem 6.80 which shows
that ADL.R/ implies that Qmax is nontrivial in the sense required for the basic analysis
summarized in Theorem 6.30.
We require a preliminary lemma.
Lemma 6.77. Assume ADL.R/ and suppose
G Qmax
is L.R/-generic. Then in L.R/G, for every set A 2 P .R/ \ L.R/ the set
X hH.!2 /; A; 2i j MX is A-iterable and X is countable
contains a club, where MX is the transitive collapse of X .
Proof. This is the Qmax version of Lemma 4.52.
Suppose G Qmax is L.R/-generic. From the basic analysis of Qmax summarized
in Theorem 6.73 and Theorem 6.74, it follows that
H.!2 /L.R/G D H.!2 /L.R/ G:
We work in L.R/G. Fix A R with A 2 L.R/. Fix a stationary set
S P!1 .H.!2 //
and x a countable elementary substructure
X hH.!2 /; A; G; 2i
such that X \ H.!2 / 2 S . Let hXi W i < !i be an increasing sequence of countable
elementary substructures of X such that
X D [Xi j i < !
and such that for each i 2 !, Xi 2 XiC1 . Therefore for each i < !, there exists
.hMk W k < !i; f / 2 G \ XiC1 satisfying
(1.1) Xi \ P .!1 / j.M0 /,
(1.2) A \ M0 2 M0 ,
(1.3) hMk W k < !i is A-iterable,
where
359
t
u
As a corollary to Lemma 6.77 and Lemma 4.24 we obtain the following theorem
which easily yields a plethora of conditions in Qmax .
Theorem 6.78. Assume AD holds in L.R/. Suppose G Qmax is L.R/-generic.
Then in L.R/G the following holds. Suppose 2 Ord,
L .R/G ZFC ;
and that
L .R/ 1 L.R/:
Suppose
X L .R/G
is a countable elementary substructure with G 2 X . Let MX be the transitive collapse
of Y and let
IX D .INS /MX :
Then for each A R such that A 2 X \ L.R/, .MX ; IX / is A-iterable.
6 Pmax variations
360
MX
t
u
361
(1) !3 D L.R/ .
(2) 12 D !2 .
(3) Suppose S !1 is stationary and
f W S ! Ord:
Then there exists g 2 L.R/ such that 2 S j f ./ D g./ is stationary.
Proof. (2) follows from Theorem 3.16 and Lemma 6.77.
By Theorem 6.75,
L.R/ !3
in L.R/G since c D !2 in L.R/G.
Qmax satises the following chain condition in L.R/. Suppose
W Qmax !
is a function. Then the range of is bounded in . This is because there is a map of
the R onto Qmax .
The usual analysis of terms shows that L.R/ is a cardinal in L.R/G. By Theorem 6.74(1), !1L.R/ and !2L.R/ are cardinals in L.R/G. Therefore (1) follows.
Similarly for (3) one can reduce to the case that for some < ,
f W !1 ! :
and so (3) follows from Theorem 3.42, Lemma 6.77, and Theorem 6.74(2).
t
u
6 Pmax variations
362
t
u
The analysis of L.R/Qmax given by the assumption of the existence of a huge cardinal can now be carried out just assuming ADL.R/ .
For example we obtain the following theorem as an immediate consequence of
Theorem 6.80, Theorem 6.30, Theorem 6.31, and Theorem 6.53.
Theorem 6.81. Assume AD holds in L.R/. Suppose
G Qmax
is L.R/-generic. Then
L.R/G ZFC
and in L.R/G:
(1) P .!1 / D P .!1 /G and IG D INS ;
(2) every set of reals of cardinality !1 is of measure 0;
(3) the reals cannot be decomposed as an !1 union of meager sets;
(4) the nonstationary ideal on !1 is !1 -dense;
(5) the function fG witnesses ++ .!1<! /;
(6) AC holds.
t
u
363
t
u
t
u
Denition 6.84. Suppose A is an alphabet for a rst order language and that A contains 2 and a unary predicate U .
A formula of L.A/ is a U -restricted 2 formula if there is a 0 -formula
.x; y; z/ in L.A n U / such that
D 8x8y9z.x/ !
.x; y; z/
t
u
Theorem 6.85. Suppose there are ! many Woodin cardinals with a measurable above
them all. Suppose
F W !1 ! H.!1 /
is a function which witnesses + .!1<! / and for all p 2 Coll.!; !1 /,
< !1 j p 2 F ./
is stationary. Suppose is a F INS -restricted 2 sentence in the language for the
structure
hH.!2 /; F INS ; X; 2 W X R; X 2 L.R/i
364
6 Pmax variations
and that
hH.!2 /; F INS ; X; 2 W X R; X 2 L.R/i :
Then
Qmax
:
Proof. From the large cardinal assumptions, AD holds in L.R/. Therefore by Theorem 6.80, Qmax is nontrivial in the sense required for the analysis of Qmax .
Fix A R with A 2 L.R/.
is a F INS -restricted 2 sentence and so
D 8x8y9z.x/ !
.x; y; z/
::
365
such that:
(1.1) j.!1M / D !1 ;
(1.2) INS \ M D I ;
(1.3) j j.f /./ D F ./ contains a club in !1 .
Let f D j.f /, H D j.h/ and let B D j.b/. The sentence holds in V and so
there exists a set D 2 H.!2 / such that
hH.!2 /; A; 2i
H; B; D:
M; A; b; h; F; f ; M X:
Let .MX ; IX / be the transitive collapse of .X; I/, let .fX ; BX ; HX ; DX / be the image
of .f ; B; H; D/ under the collapsing map. Thus
fX D f jX \ !1 ;
HX D H jX \ !1 and
BX D B \ X \ !1 :
Similarly DX D D \ X \ !1 .
Let h.N ; J /; gi 2 Qmax be a condition such that MX 2 N and such that MX is
countable in N .
Choose an iteration,
k W .MX ; IX / ! .MQ X ; IQX /;
in N of length .!1 /N such that
j k.fX /./ g./ 2 J:
The iteration exists by Lemma 6.66 since g 2 .YColl .J //N .
Thus h.N ; J /; k.fX /i 2 Qmax and
h.N ; J /; k.fX /i < h.M; I /; f i:
Since X V0 ,
hH.!2 /; A \ X; 2iMX
HX ; BX ; DX ;
and so by elementarity
Q
Finally k..HX ; BX // is the image of .h; b/ under the iteration of .M; I / which
t
u
sends f to k.fX / and this contradicts the choice of h.M; I /; f i and b.
6 Pmax variations
366
< !1 j p 2 F ./
is stationary. Suppose that for each F INS -restricted 2 sentence, , in the language
for the structure
hH.!2 /; F INS ; X; 2 W X R; X 2 L.R/i
if
Qmax
:
then
hH.!2 /; F INS ; X; 2 W X R; X 2 L.R/i :
Then there exists G Qmax such that:
(1) G is L.R/-generic;
(2) P .!1 / L.R/G;
(3) fG D F .
Proof. By Theorem 6.80, Qmax is nontrivial in the sense required for the analysis of
Qmax .
Suppose
g 2 F NS :
We associate to the function g a lter
Fg Qmax
dened to be the set of h.M; I /; f i 2 Qmax such that there is an iteration
j W .M; I / ! .M ; I /
such that the set
< !1 j F ./ D j .f /./
367
t
u
The next theorem shows that the Qmax -extension satises a restricted version of the
homogeneity condition satised by the Pmax -extension (cf. Theorem 5.67).
We x some notation. Suppose t 2 R and that
g Coll.!; <!1 /
is a lter which is Lt -generic. Let
Fg W !1 ! H.!1 /
be the function dened as follows. Let
Hg W ! !1 ! !1
be the function given by g. For each < !1 ,
Fg ./ D p 2 Coll.!; / j p.n/ D Hg .n; / for all n 2 dom.p/:
Recall that if G Qmax is a semi-generic lter then fG INS is the set of all functions
f W !1 ! H.!1 /
such that
< !1 j f ./ D FG ./
contains a closed unbounded subset of !1 .
368
6 Pmax variations
A D j.a/
j W .M; I / ! .M ; I /
is the iteration such that j.f / D fG . By genericity we can suppose that in L.R/,
h.M; I /; f i Qmax ; w; ; fG INS
where is the term for j.a/.
Let t 2 R code h.M; I /; f i. Suppose g Coll.!; <!1 / is a lter which is Lt generic.
Work in Lt g and let
hM ; G ; j; W < !1 i
369
D !1 :
This uniquely species the iteration. We note that for each < !1 ,
hM ; G ; j; W < < i 2 Lt g \ Coll.!; </
M
!1 .
where D
Therefore by the genericity of g, it follows by induction on that
for each < !1 , Fg ./ is M -generic and so j; C1 exists as specied.
A key property of the iteration is that
j j0;!1 .f /./ D Fg ./
contains
!1
j < !1
370
6.2.3
6 Pmax variations
2
Qmax
2 j.s/
C 2 j.t /
j0; .s/ 2 G
if and only if
j0; .t / 2 G
C
where D .
371
is as desired.
As a corollary to Lemma 6.88 and the basic analysis of L.R/Qmax we obtain the
following lemma. The proof is essentially identical to that of Lemma 5.17, using
Lemma 6.88 in place of Lemma 5.16.
Lemma 6.89. Assume AD L.R/ . Suppose G Qmax is L.R/-generic. Then
L.R/G
AC :
t
u
Lemma 6.89 combined with Theorem 6.78 yields the following variation of Theorem 6.80.
Corollary 6.90. Assume AD holds in L.R/. Then for each set A R with A 2 L.R/,
there is a condition h.M; I /; f i 2 Qmax such that
(1) M ZFC C
AC ,
(2) A \ M 2 M,
(3) hH.!1 /M ; A \ M; 2i hH.!1 /; A; 2i,
(4) .M; I / is A-iterable.
t
u
Denition 6.91. 2 Qmax is the set of nite sequences hM; I; J; f; Y i which satisfy the
following.
(1) M ZFC C
AC .
372
6 Pmax variations
t
u
Proof. Let
373
!1M n a J:
Therefore
M a is a stationary, co-stationary, subset of !1 .
Since j1 .f / D j2 .f / it follows that j1 .a/ D j2 .a/.
The lemma follows by Lemma 5.15.
t
u
Using Corollary 6.90 we trivially obtain the nontriviality of 2 Qmax as required for
the analysis of the 2 Qmax -extension.
Lemma 6.94. Assume AD holds in L.R/. Then for each set A R with
A 2 L.R/;
there is a condition hM; I; J; f; Y i 2 2 Qmax such that
(1) jY =I j D !1 in M,
(2) A \ M 2 M,
(3) hH.!1 /M ; A \ M; 2i hH.!1 /; A; 2i,
(4) .M; I / is A-iterable.
Proof. By Corollary 6.90, there is a condition h.M0 ; I0 /; f0 i 2 Qmax such that
.M0 ; I0 / satises (2)(4). We may suppose, by modifying f0 is necessary, that for
all < !1M0 , f0 ./ is a maximal lter in Coll.!; /.
For each < !1M0 let
S D < !1M0 j < 1 C and .0; / 2 f0 ./:
Dene
374
6 Pmax variations
(1.3) I D I0 ,
(1.4) J D I0 jS0 .
It follows that hM; I; J; f; Y i 2 2 Qmax and is as required.
t
u
The iteration lemmas for 2 Qmax are proved by minor modications in the arguments used to prove the iteration lemmas for Qmax . The only difference is that the
iteration lemmas for 2 Qmax are more awkward to state.
Lemma 6.95. Suppose that
hM0 ; I0 ; J0 ; f0 ; Y0 i 2 2 Qmax ;
hM1 ; I1 ; J1 ; f1 ; Y1 i 2 2 Qmax ;
and that
hM0 ; I0 ; J0 ; f0 ; Y0 i 2 .H.!1 //M1 :
Suppose that Y Y1 , Y 2 M1 and that
M1 jY =I1 j D !1 :
Then there exists an iteration
j W .M0 ; I0 ; J0 / ! .M0 ; I0 ; J0 /
such that j 2 M1 and such that the following hold.
(1) < !1M1 j j.f0 /./ f1 ./ 2 I1 .
(2) I0 D M0 \ I1 and J0 D M0 \ J1 .
(3) j.Y0 /=I1 Y =I1 .
Proof. The key point is the following. Suppose
jQ W .M0 ; I0 ; J0 / ! .MQ 0 ; IQ0 ; JQ0 /
is a countable iteration and that
Q
g Coll.!; !1M0 /
is MQ 0 -generic.
Then
(1.1) there exists an iteration
kQ W .MQ 0 ; JQ0 / ! .MQ 1 ; JQ1 /
such that
Q
kQ jQ.f0 /.!1M0 / D g;
375
Q
kQS jQ.f0 /.!1M0 / D g
Q
!1M0 2 kQS .S /:
j.a0 / n 5Y1 2 I1
where
j W .M0 ; I0 ; J0 / ! .M1 ; I1 ; J1 /
is the .unique/ iteration such that j.f0 / D f1 .
t
u
b \ a 2 I1
t
u
376
6 Pmax variations
Using the proof of Lemma 6.95 and of its generalization to sequences of conditions,
the analysis of the 2 Qmax extension can be carried out in a manner quite similar to that
for the Qmax -extension.
The results are summarized in the next theorem where we use the following notation.
Suppose G 2 Qmax is L.R/-generic. Let
fG D [f j hM; I; J; f; Y i 2 G:
For each condition
hM; I; J; Y; f i 2 G
there is a unique iteration
j W .M; I; J / ! .M ; I ; J /
such that j.f / D fG . We let Y denote j.Y /. Let
(1) P .!1 /G D [P .!1 / \ M j hM; I; J; Y; f i 2 G,
(2) IG D [I j hM; I; J; f; Y i 2 G,
(3) JG D [J j hM; I; J; f; Y i 2 G,
(4) YG D [Y j hM; I; J; f; Y i 2 G.
Theorem 6.98. Assume ADL.R/ . Suppose G 2 Qmax is L.R/-generic. Then
L.R/G ZFC
and in L.R/G:
(1) P .!1 / D P .!1 /G ;
(2) IG D INS ;
(3) for each set A 2 JG there exists Y YG such that jY j D !1 and such that
A n 5Y 2 IG I
(4) YG is predense in .P .!1 / n IG ; /;
(5) For each S 2 P .!1 / n JG ,
j p 2 fG ./ n S 2 JG
for some p 2 Coll.!; !1 /;
(6) For each S 2 YG and for each T S such that T IG ,
j p 2 fG ./ n T 2 IG
for some p 2 Coll.!; !1 /;
(7) JG D sat.IG /.
377
t
u
t
u
378
6 Pmax variations
2 L.; x0 /:
Let be the ordertype of . We can suppose that !
D by increasing if necessary.
Let M 2 LA0 be a countable transitive set such that
(3.1) x0 2 M ,
(3.2) < !1M ,
(3.3) M ZFC C There exist measurable cardinals ,
(3.4) M is iterable (by linear iterations using the normal measures in M ).
379
Since
1
LA0
2 -Determinacy;
the transitive set M exists. It follows that there exists an iteration by the normal measures in M ,
j W M ! M ;
such that is M -generic for product Prikry forcing. Thus !1 is inaccessible in M .
However x0 2 M and so
2 M :
This proves (1.1).
Fix A and x x 2 R such that
x # L.A; x/:
Assume toward a contradiction that
A D !2 :
Then for every set B !1 ,
.A;B/ D !2
where
.A;B/ D sup.!2 /LZ j Z !1 ; .A; B/ 2 LZ; and RLAB D RLZ :
Thus we can assume that !1 D .!1 /LA and that .A;x/ D !2 .
Let 0 be an innite set of indiscernibles of Lx with 0 !2 n !1 . Let Z !1
witness that .A;x/ > sup.0 /.
Thus there exists a countable set 2 LZ such that
(4.1) !2LZ ,
(4.2) 0 ,
(4.3) x 2 L ,
(4.4) is countable in L .
By (1.1), !1 is inaccessible in L and so by Jensens Covering Lemma,
x # 2 L LZ:
This contradicts that RLZ D RLAx .
t
u
This (essentially) rules out one method for attempting to have weak Kurepa trees
in L.R/P where P is any Pmax variation we have considered so far.
Remark 6.101. (1) There are Pmax -variations which yield models in which any previously specied set of reals is !1 -borel in the simplest possible manner, given
380
6 Pmax variations
>
(2) There is an interesting open question. Suppose that INS is !2 -saturated and that
P .!1 /# exists. For each A !1 let A be as dened in Lemma 6.100. Must
t
u
A < !2 ?
To prove that there are weak Kurepa trees in L.R/Qmax , it is necessary to to nd a condition h.M; I /; f i 2 Qmax and a tree T 2 M of rank !1 in M such that if h.N ; J /; gi
is a condition in Qmax and M 2 N with M countable in N then there is an iteration
j W .M; I / ! .M ; I /
in N with the following properties.
(1) J \ M D I .
(2) j.f / D g modulo J .
(3) There is a conal branch b of j.T / such that b M .
The next lemma identies the requirements which we shall use.
Lemma 6.102 (ZFC ). Suppose that f is a function which witnesses + .!1<! /. Then
there is triple .T; g; h/ such that the following conditions hold.
(1) T is a subtree of 0;1<!1 and T has cardinality !1 .
(2) g W !1 ! T is a bijection.
(3) For all < !1 the set x 2 T j < dom.x/ is dense in T .
(4) h W !1 !1<! ! T .
(5) For each < !1 let T be the .countable/ subtree of T generated by the range
of g restricted to and let
h W !1<! ! T
be the section of h at ; h .s/ D h.; s/. Then for all limit 0 < < !1 ,
a) h is an order preserving function, h W !1<! ! T ,
b) h .!1<! / is dense in T ,
c) for all s 2 !1<! , h .t / j s t is dense below h .s/ in T .
381
t
u
The role of g in the conditions specied in Lemma 6.102 is simply to control the
sets T , for example it follows that T T whenever < .
Let T be the set of hM; I; f; .T; g; h/i such that
(1) h.M; I /; f i 2 Qmax ,
(2) f witnesses ++ .!1<! / in M,
(3) .T; g; h/ together with f satisfy in M the conditions (1)(6) indicated in the
statement of Lemma 6.102.
Lemma 6.103 (ZFC C ++ .! <! /). Suppose that f is a function which witnesses
++ .!1<! / and that
hM0 ; I0 ; f0 ; .T0 ; g0 ; h0 /i 2 T :
Then there is an iteration
k W .M0 ; I0 / ! .M0 ; I0 /
such that the following hold.
(1) k.!1M0 / D !1 .
(2) k.f0 / D f on a club in !1 .
(3) There exists a M0 -generic branch for k.T0 /.
Proof. Let p0 D hM0 ; I0 ; f0 ; .T0 ; g0 ; h0 /i. Fix a club C !1 such that for all
2 C , f ./ is L.f j; C \ ; p0 /-generic for Coll.!; /. We work in L.f; C; p0 /
and construct by induction on < !1 an iteration of .M0 ; I0 /,
h.M ; I /; G ; j; W < < !1 i
and a sequence hb W < !1 i such that for all < < !1 ,
(1.1) b is an M -generic branch of j0; .T0 /,
(1.2) b b ,
(1.3) the set j G D f . / contains a club in !1 .
382
6 Pmax variations
g D g j!1M :
t
u
From the previous lemmas and the basic analysis of Qmax it follows that there is
a weak Kurepa tree in L.R/Qmax . Recall that if G Qmax is L.R/-generic then the
associated function fG witnesses ++ .!1<! / in L.R/G.
Theorem 6.104. Suppose ADL.R/ and that G Qmax is L.R/-generic. Suppose that
.T; g; h/ 2 L.R/G and that .T; g; h/ together with fG satisfy the conditions .1/.6/
of Lemma 6.102.
(1) If D is a set of dense subsets of T and D has cardinality !1 then there is a
D-generic branch of T .
(2) T has !2 distinct branches of rank !1 .
383
6.2.5
KT
Qmax
As our next example of a variant of Qmax we dene a partial order KT Qmax . The partial
order KT Qmax is obtained from Qmax by simply changing the denition of the order. Our
goal is to produce a model in which the nonstationary ideal on !1 is !1 -dense and in
which there are no weak Kurepa trees on !1 . By Lemma 6.51, if there is an !1 -dense
ideal on !1 then there is a Suslin tree. Thus one cannot obtain a model in which there
is an !1 -dense ideal on !1 and in which there are no weak Kurepa trees, by sealing
trees.
We shall also state as Theorem 6.121, the absoluteness theorem for the KT Qmax extension which is analogous to the absoluteness theorem (Theorem 6.84) which we
proved for the Qmax -extension.
384
6 Pmax variations
Denition 6.105. Let KT Qmax be the partial order obtained from Qmax as follows:
KT
Qmax D Qmax ;
but the order on KT Qmax is the following strengthening of the order on Qmax .
A condition h.M1 ; I1 /; f1 i < h.M0 ; I0 /; f0 i if h.M1 ; I1 /; f1 i < h.M0 ; I0 /; f0 i
relative to the order on Qmax and if addition the following holds.
Let
j W .M0 ; I0 / ! .M0 ; I0 /
is the (unique) iteration such that j.f0 / D f1 . Suppose b !1M1 , b 2 M1 and
b \ 2 M0
for all < !1M1 .
Then b 2 M0 .
t
u
The iteration lemmas necessary for the analysis of KT Qmax are an immediate corollary of the following lemmas.
Lemma 6.106 (ZFC + + .!1<! / + ). Suppose
F W !1 ! H.!1 /
is a function which witnesses .!1<! /. Then for any h.M; I /; f i 2 Qmax there is an
iteration
j W .M; I / ! .M ; I /
+
385
F ./ Coll.!; !1 /
is a lter which is M.M ; h W < i/-generic then
G D F ./:
This iteration is easily constructed.
Since F witnesses + .!1<! / it follows by (1.2) that
< !1 j j0;!1 .f /./ D F ./
contains a club in !1 .
A key property of the iteration is the following one. Suppose < !1 . Then
2 M C1 or
M!1 :
This follows the genericity requirement of (1.1).
We note the following. Suppose A !1 . Let S be the set of < !1 such that
A \ D and such that F ./ is M.A \ /-generic for Coll.!; /. Then S is
stationary in !1 . This following by reection, since by ZFC for each A !1 , there
exists A < !2 such that !1 < A and such that LA A is admissible.
Now suppose b !1 and that b \ 2 M!1 for all < !1 . Let A !1 code
.b; h.M ; I /; G ; j; W < !1 i/:
Thus for a stationary set of < !1 the following hold.
(2.1) A \ D .
(2.2) F ./ is M.A \ /-generic for Coll.!; /.
(2.3) j0; .!1M0 / D and
.b; h.M ; I /; G ; j; W < i/ 2 LA \ :
By (1.2), for each such , G D F ./ and so b \ 2 M or b \ M C1 . But
if b \ M C1 then
b \ M!1
which is a contradiction. Hence b \ 2 M .
Thus for a stationary set of < !1 ,
b \ 2 M
and so b 2 M!1 .
Therefore the iteration is as desired.
A similar argument proves the following generalization of Lemma 6.106.
t
u
386
6 Pmax variations
(iii) !1
MkC1
D !1
(iv) IkC1 \ Mk D Ik ,
(v) fkC1 D fk ,
(vi) if C 2 Mk is closed and unbounded in !1M0 then there exists D 2 MkC1 such
that D C , D is closed and unbounded in C and such that
D 2 Lx
for some x 2 R \ MkC1 .
Then there is an iteration
j W h.Mk ; Ik / W k < !i ! h.Mk ; Ik / W k < !i
of length !1 such that:
(1) F D j.f0 / on a club in !1 ;
(2) if b !1 is a set such that b \ 2 [Mk for all < !1 , then
b 2 [Mk :
t
u
KT
387
L.R/G ZFC
and in L.R/G:
(1) P .!1 /G D P .!1 /;
(2) IG is a normal !1 -dense ideal on !1 ;
(3) IG is the nonstationary ideal;
(4) there are no weak Kurepa trees;
(5) for every A !1 there exists B !1 such that A 2 LB and such that for all
S !1 if
S \ 2 LB
for all < !1 then S 2 LB.
Proof. By Lemma 6.47, for every set A R with A 2 L.R/ there is a condition
h.M; I /; f i 2 Qmax such that
(1.1) A \ M 2 M,
(1.2) hH.!1 /M ; A \ M; 2i hH.!1 /; A; 2i,
(1.3) .M; I / is A-iterable,
(1.4) holds in M,
(1.5) f witnesses ++ .!1<! / in M.
The proof that
L.R/G !1 -DC
requires dovetailing the proof of Lemma 6.107 with the construction of an !1 decreasing sequence of conditions in Qmax . This being done in the model associated to a
condition in Qmax which satises (1.1)(1.5) above relative to a suitable choice for A.
As in the case of Qmax this is straightforward. Similarly the proof that AC (or AC )
holds in L.R/G is a routine adaptation of the proof of the corresponding claim for
Qmax .
The remaining claims, except (4) and (5), follow from arguments which follow
quite closely the arguments for Qmax .
The claim (4) is an immediate consequence of (5). By (2) (or (1)), for every set
B !1 , B # exists and so
jP .!1 / \ LBj D !1 :
We sketch the proof of (5). We begin with the following claim.
6 Pmax variations
388
!1M0
is the unique iteration such that j.f0 / D fG . We work in L.R/ and we assume that
the condition h.M0 ; I0 /; a0 i forces that X0 D j.b0 / is a counterexample to (5).
Let z 2 R be any real such that M0 2 Lz and M0 is countable in Lz. For each
i !, Let i be the i th Silver indiscernible of Lz. Let
k W L! z ! L! z
be the canonical embedding such that cp(k) = 0 and let
L! z D k.f /.0 / j f 2 L! z:
Let U be the L! z-ultralter on 0 given by k,
U D A 0 j A 2 L! z; 0 2 k.A/:
Thus
L! z Ult.L! z; U /
and k is the associated embedding. For each X P .0 / \ L! z if X 2 L! z and
jX j 0 in L! z then U \ X 2 L! z. Therefore .L! z; U / is naturally iterable.
389
p .k; / D p.k/:
390
6 Pmax variations
391
!1M0
and if
a \ 2 M0
!1M0 ,
b \ 2 M0
M
t
u
392
6 Pmax variations
t
u
The analysis of Qmax generalizes to KT Qmax using the proof of Lemma 6.111 and
using Theorem 6.113 to obtain the necessary conditions.
One obtains Theorem 6.113 by modifying the proof of Theorem 6.64. (6) is the
key requirement, the other requirements are automatically satised by the condition
produced in the proof of Theorem 6.64. The modication of the proof of Theorem 6.64
involves proving the following strengthening of Lemma 6.63.
Lemma 6.112 (For all x 2 R, x # exists). Suppose N is a transitive model of ZFC of
height !1 and A !1 is a conal set such that
A\ 2N
for all < !1 . Suppose B !1 , B A and that .N; A; B/ is constructible from a
real. Let z 2 R be such that
.N; A; B/ 2 Lz
and such that .N; A; B/ is denable from .!1V ; z/ in Lz. Then there exist x 2 R and
a function
f W H.!1 / ! H.!1 /
such that f is 11 .x/ and such that the following hold.
(1) For all 2 B, if .g; h/ 2 H.!1 / and if
a) g is N -generic for Coll .!; A \ /,
b) h is N g-generic for Coll .!; /,
then f .g; h/ is an N hg-generic lter for Coll .!; S / where
S D A \ j < <
and is the least element of B above .
393
be any
denable function which satises (1). These exist by Lemma 6.63.
It follows that f must satisfy (2).
t
u
Using Lemma 6.112 in the proof of Theorem 6.64 yields the requisite strengthening
of Theorem 6.65.
Theorem 6.113. Assume AD holds in L.R/. Then for every set A R with A 2 L.R/
there is a condition .hMk W k < !i; f / 2 Qmax such that the following hold.
(1) A \ M0 2 M0 .
(2) hH.!1 /M0 ; A \ M0 ; 2i hH.!1 /; A; 2i.
(3) hMk W k < !i is A-iterable.
(4) holds in M0 .
(5) f witnesses + .!1<! / in M0 .
(6) Suppose b !1M0 , b 2 [Mk j k < !, and
b \ 2 M0
for all < !1M0 . Then b 2 M0 .
t
u
6 Pmax variations
394
!1N0
j.f / D g
and such that if b !1N0 is a set in N0 satisfying
b \ 2 M0
b 2 [Nk j k < !;
b \ 2 M0
Theorem 6.114. Assume ADL.R/ . Then KT Qmax is !-closed and homogeneous. Suppose G KT Qmax is L.R/-generic. Then
L.R/G ZFC
and in L.R/G:
395
t
u
This sufces for the consistency result. With just a little more work one can easily prove the following lemmas which are the relevant versions of Lemma 6.77 and
Lemma 6.79. This in turn leads to absoluteness theorems for the KT Qmax -extension of
L.R/.
Lemma 6.115. Assume ADL.R/ and suppose
G KT Qmax
is L.R/-generic. Then in L.R/G, for every set A 2 P .R/ \ L.R/ the set
X hH.!2 /; A; 2i j MX is A-iterable and X is countable
t
u
The proof of Lemma 6.116 follows that of Theorem 6.78 using Lemma 6.115 in
place of Lemma 6.77.
Lemma 6.116. Assume AD holds in L.R/. Suppose G
Then in L.R/G the following holds.
Suppose > !2 ,
L .R/G ZFC ;
KT
Qmax is L.R/-generic.
and that
L .R/ 1 L.R/:
Suppose
X L .R/G
is a countable elementary substructure with G 2 X . Let MX be the transitive collapse
of Y and let
IX D .INS /MX :
Then for each A R such that A 2 X \ L.R/, .MX ; IX / is A-iterable.
t
u
396
6 Pmax variations
Theorem 6.117. Assume AD holds in L.R/. Then for each set A R with A 2 L.R/,
there is a condition h.M; I /; f i 2 Qmax such that the following hold.
(1) M ZFC .
(2) I D .INS /M .
(3) A \ M 2 M.
(4) hH.!1 /M ; A \ M; 2i hH.!1 /; A; 2i.
(5) .M; I / is A-iterable.
(6) f witnesses ++ .!1<! / in M.
(7) For each a !1M with a 2 M there exists b !1M such that
a) b 2 M and a 2 Lb,
b) for all d 2 M, if d !1M and if d \ 2 Lb for all < !1M then
d 2 Lb.
t
u
Theorem 6.117 provides the conditions in Qmax which are sufcient to carry
out the proof of Theorem 6.108. The key point is that the principle specied in
Theorem 6.117(7) successfully substitutes for in the proofs of Lemma 6.106,
Lemma 6.107, and in proving the generalizations which arise in the proof of Theorem 6.108. We illustrate this claim by proving Lemma 6.106 and for this we recall the
principle, + :
There is a sequence
ha W < !1 i
397
S \ 2 LB
j0; .!1M0 / D
398
6 Pmax variations
(3.1) A \ 2 a .
(3.2) F ./ is M.A \ /-generic for Coll.!; /.
(3.3) j0; .!1M0 / D and
.b \ ; h.M ; I /; G ; j; W < i/ 2 M.A \ /:
By (2.1) and (2.2), for each such , G D F ./ and so b\ 2 M or b\ M C1 .
But if b \ M C1 then
b \ M!1
which is a contradiction. Hence b \ 2 M .
Thus for a stationary set of < !1 ,
b \ 2 M
and so b 2 M!1 .
In summary we have proved that if b !1 , b 2 LB and if
b \ 2 M!1
for all < !1 , then b 2 M!1 .
Now suppose b !1 , b 2 V , and that
b \ 2 M!1
for all < !1 . Then
b \ 2 LB
for all < !1 since M!1 2 LB. Therefore b 2 LB by the key property of B, and
so b 2 M!1 .
Therefore the iteration has the desired properties in V .
t
u
There are absoluteness theorems corresponding to
the details to the reader.
Theorem 6.121 corresponds to Theorem 6.85.
KT
< !1 j Y \ 2 a
t
u
399
(i) M ZFC,
(ii) I 2 M and I is the tower of ideals I< as computed in M where is a Woodin
cardinal in M .
Suppose f 2 M , f witnesses + .!1<! / in M and for all p 2 .!1<! /M the set
< !1M j p 2 f ./
is stationary within M .
Then there is an iteration
j W .M; I/ ! .M ; I /
of length !1 such that:
(1) F D j.f / on a club in !1 ;
(2) if b !1 is a set such that b \ 2 M for all < !1 , then
b 2 M :
t
u
KT Q
max
:
t
u
400
6 Pmax variations
We end this section with a sketch of the proof of Theorem 5.75. For this it is
convenient to make the following denition. A tree
T 0;1<!1
is weakly special if for all countable elementary substructures
X hH.!2 /; T; 2i;
if
b W !1 \ X ! 0;1
a conal branch of TX such that b MX , then there is a bijection
W ! ! !1MX
which is denable in the structure
hMX ; TX ; b; 2i
where hMX ; TX ; 2i is the transitive collapse of X .
We shall require the following lemma which is a reformulation of a theorem of
Baumgartner. Recall that if P is a proper partial order and if
GP
is V -generic then
V
D .INS /V G \ V:
INS
Lemma 6.122 (Baumgartner). Suppose that
T 0;1<!1
is a tree with rank !1 . Then there is a proper partial order P such that if
GP
is V -generic then in V G;
(1) the tree T is weakly special,
(2) suppose that b 2 0;1!1 is a conal branch of T , then b 2 V .
t
u
It follows, by absoluteness and reection, that the set of weakly special trees of
cardinality !1 is 1 denable in the structure
hH.!2 /; 2i
using !1 as a parameter.
This leads to a strengthening of the sentence .
Denition 6.123. C
: For all A !1 there exists B !1 such that
(1) A 2 LB,
(2) the tree TB is weakly special where
TB D 0;1<!1 \ LB;
(3) suppose b 2 0;1!1 is a branch of TB , then b 2 LB.
C
t
u
401
!1M
then d 2 Lb.
j0 W .M0 ; I0 / ! .M0 ; I0 /
6 Pmax variations
402
(3.1) N ZFC,
(3.2) x 2 N ,
(3.3) is a Woodin cardinal in N .
Let g0 be N -generic for the partial order
.Coll.!1 ; R//N
and let g1 be N g0 -generic for Coll.!; <!1N /.
Thus
N g0
and
N g0 g1 C ++ .!1<! /:
By Lemma 6.106, there exists an iteration
j W .M; I / ! .M ; I /
in N such that:
(4.1) I D .INS /N \ M ;
(4.2) if b 2 P .!1 /N is a set such that b \ 2 M for all < !1 , then
b 2 M :
Thus
T D 0;1<!1 \ Lb1 :
By Lemma 6.122 there exists a partial order P 2 N g0 g1 such that if g P is
N g0 g1 -generic then
(6.1) .INS /N g0 g1 D N g0 g1 \ .INS /N g0 g1 g ,
(6.2) T is weakly special in N g0 g1 g,
403
t
u
404
6 Pmax variations
We x some notation. We let A denote the following partial order. This is Amoebaforcing scaled by 1=2. Conditions are perfect sets X 0;1 such that
.X / > 1=2
and such that
.X \ O/ > 0 for all open sets O 0;1 with X \ O ;. The latter
condition serves to make A separative. The order on A is by set inclusion. Suppose
G A is V -generic and in V G let
X D \P
V G
j P 2 G
V G
where P
denotes the closure of P computed in V G. This is P as computed in
V G. Then X has measure 1=2 and every member of X is random over V .
Suppose I is a uniform, countably complete, ideal on !1 and
F W !1 ! P .0;1/:
Let YA .F; I / be the set of all pairs .S; P / such that the following hold.
(1) S !1 and S I .
(2) P 0;1 and P 2 A.
(3) Suppose hPk W k < !i is a maximal antichain in A below P . Then
2 S j F ./ 6 Pk for all k < ! 2 I:
(4) If Q P is a perfect set of measure > 1=2 then
2 S j F ./ Q I:
(5) For all 2 S ,
.F .// D 1=2.
Suppose I is a uniform normal ideal on !1 and that F is a function such that
YA .F; I / is nonempty. Suppose .S1 ; P1 / 2 YA .F; I /, .S2 ; P2 / 2 YA .F; I / and that
S1 S2 . Then P1 P2 . Therefore if
G P .!1 / n I
is a lter in .P .!1 / n I; / then
HG D P 2 A j .S; P / 2 YA .F; I / for some S 2 G
generates a lter in A.
Lemma 6.125 (ZFC ). Suppose I is a uniform normal ideal on !1 and
F W !1 ! P .0;1/
is a function such that YA .F; I / is nonempty. Suppose .S1 ; P1 / 2 YA .F; I /.
(1) Suppose P2 is a perfect subset of P1 and P2 2 A. Let
S2 D 2 S1 j F ./ P2 :
Then .S2 ; P2 / 2 YA .F; I /.
405
(2) Suppose S2 S1 and S2 I . Then there exists .S3 ; P3 / 2 YA .F; I / such that
S3 S 2 .
Proof. We rst prove (1). To show that .S2 ; P2 / 2 YA .F; I / we have only to prove
that condition (3) in the denition of YA .F; I / holds for .S2 ; P2 /. The other clauses
are an immediate consequence of the fact that .S1 ; P1 / 2 YA .F; I /.
We may assume that
.P2 / <
.P1 / for otherwise there is nothing to prove. Let
hXi W i < !i be a maximal antichain in A below P2 . Let hZi W i < !i be a maximal
antichain in A of conditions below P1 which are incompatible with P2 . The key point
is that we may assume that for each i < !,
.Zi \ P2 / < 1=2; if
.Z \ P2 / D 1=2
then there exists a condition W 2 A such that
(1.1) W < Z,
(1.2)
.W \ P2 / < 1=2.
Clearly
Xi j i < ! [ Zi j i < !
is a maximal antichain below P1 . Since .S1 ; P1 / 2 YA .F; I /, for I -almost all 2 S1 ,
there exists i < ! such that either F ./ Xi or F ./ Zi . For every 2 S2 and
for all i < !,
.F .// D 1=2, F ./ P2 and
.P2 \ Zi / < 1=2. Therefore for
I -almost all 2 S2 , F ./ Xi for some i < !. Therefore condition (3) holds for
.S2 ; P2 / and so
.S2 ; P2 / 2 YA .F; I /:
This proves (1).
We prove (2). Suppose G P .!1 / n I is V -generic for .P .S1 / n I; /. Let
j W V ! .M; E/
be the associated generic elementary embedding. Since the ideal I is normal it follows
that !1 belongs to the wellfounded part of .M; E/. Since .S1 ; P1 / 2 YA .F; I / it
follows that HG is V -generic for A where
HG D Q 2 A j j.F /.!1 / Q:
By part (1) of the lemma this induces a complete boolean embedding
W RO.AjP1 / ! RO..P .S1 / n I; //
where AjP1 denotes the suborder of A obtained by restricting to the conditions below
P1 . Let
b D ^c 2 RO.AjP1 / j S2 .c/
and let hXi W i < !i be a maximal antichain below b of conditions in A. For each
i < ! let
Ti D 2 S2 j F ./ Xi :
For each i < !, if Ti I then .Ti ; Xi / 2 YA .F; I /. Therefore it sufces to show that
for some i < !, Ti I .
Note that if Q 2 AjP1 and T S are such that T .Q/ then
2 T j F ./ 6 Q 2 I:
t
u
This follows from the denition of HG . Hence Ti I for all i < !.
6 Pmax variations
406
407
Let Q 0;1 be a perfect (nowhere dense) set of positive measure. Since Q has
positive measure, Q is of measure 1.
Fix < !1 such that P Q . Q is an F and so there exist closed (proper)
intervals
I J 0;1
with rational endpoints such that P \ I ; and such that
P \ I D QJ \ P \ I D Q \ P \ I:
This implies that
J .P \ I / D Q \ J .P \ I /
and so
J .P \ J / \ J .I / Q
and
.J .P \ J / \ J .I // > 0. There are only !1 many sets of the form
J .P \ J / \ J .I /
where I J 0;1 are closed subintervals with rational endpoints, < !1 and
I \ P ;. . Therefore these sets collectively witness (2).
Finally we show that (2) implies (3). Let hP W < !1 i be a sequence of perfect
subsets of 0;1 each of positive measure such that the sequence witnesses (2).
Suppose Q 0;1 is a perfect set of positive measure. For each < !1 let
X D [P j < and P Q:
We claim that for all sufciently large ,
.Q n X / D 0. This is immediate. Suppose
< !1 and
.Q n X / > 0. Then there exists < !1 such that
P Q n X
and so
.X / <
.X / for some < !1 . The claim follows.
Let
hQ W < !1 i
enumerate the perfect subsets of 0;1 which can be expressed as a nite union of the
P s. Thus hQ W < !1 i witnesses (3).
t
u
Lemma 6.127 (ZFC ). Assume + .!1<! /. The following are equivalent.
(1) There is a function
F W !1 ! P .0;1/
such that YA .F; I / ; for some normal uniform ideal I .
(2) For any normal uniform ideal I on !1 there is a function
F W !1 ! P .0;1/
such that .!1 ; 0;1/ 2 YA .F; I /.
(3) There is a sequence hB W < !1 i of borel subsets of 0;1 such that each B is
of measure 1 and such that if B 0;1 is of measure 1 then B B for some
< !1 .
6 Pmax variations
408
Proof. We rst prove that (1) implies (3). Fix F and I . It follows immediately from
the denition of YA .F; I / that if B 0;1 is a set of measure 1 then F ./ B for
some < !1 . The point is that the set
Q 2 A j Q B
is dense in A. Therefore by Lemma 6.126, (3) holds.
We nish by proving that (3) implies (2). Let I be a normal ideal on !1 . Let
f W !1 ! H.!1 /
be a function which witnesses + .!1<! /. By modifying the function f if necessary we
may assume that for all p 2 Coll.!; !1 /, the set
2 !1 j p 2 f ./
is I -positive.
By Lemma 6.126, there is a dense set in A of cardinality !1 and so there is a
complete boolean embedding
W RO.A/ ! RO.Coll .!; !1 //:
Dene a function F W !1 ! P .0;1/ by
F ./ D \Q 2 A j .Q/ > p for some p 2 f ./:
Thus on a club in !1 , F ./ is a perfect set of measure 1=2. It is straightforward to
u
t
verify that .!1 ; 0;1/ 2 YA .F; I /.
Denition 6.128. M Qmax consists of nite sequences h.M; I /; f; F; Y i such that:
(1) h.M; I /; f i 2 Qmax ;
(2) M ZFC ;
(3) f witnesses ++ .!1<! / in M;
(4) F 2 M and
F W !1M ! P .0;1/I
409
t
u
We prove the basic iteration lemmas for M Qmax . There are two iteration lemmas,
one for models and one for sequences of models. The latter is necessary to show that
M
Qmax is !-closed. As usual its proof is an intrinsic part of the analysis of M Qmax .
We need a preliminary lemma.
Lemma 6.129 (ZFC ). Suppose h.M; I /; f; F; Y i 2 M Qmax and Q 0;1 is a perfect set with measure greater than 1=2.
Suppose .S; P / 2 Y and
.Q \ P / > 1=2:
Suppose that A 2 M,
A .P .!1 / n I /M ;
With this lemma the main iterations lemmas are easily proved. As usual it is really
the proofs of these iteration lemmas which are the key to the analysis of M Qmax .
Lemma 6.130 (ZFC C + .!1<! /). Assume there exists a sequence,
hB W < !1 i;
of borel subsets of 0;1 such that each B is of measure 1 and such that if B 0;1
is of measure 1 then B B for some < !1 .
410
6 Pmax variations
Suppose
g W !1 ! H.!1 /
is a function which witnesses + .!1<! / and that J is a uniform normal ideal on !1
such that for all p 2 Coll.!; !1 /,
j p 2 g./ J:
Suppose h.M; I /; f; F; Y i 2 M Qmax . Then there is an iteration
j W .M; I / ! .M ; I /
such that
(1) j j.f /./ D g./ contains a club in !1 ,
(2) j.Y / D YA .j.F /; J / \ M .
Proof. By Lemma 6.126, A has a dense set of size !1 . Let hP W < !1 i be a
sequence of conditions in A which is dense.
Let
h.M ; I /; G ; j; W < !1 i
be an iteration of .M; I / such that the following hold.
(1.1) For all < !1 if D !1M and if g./ is M -generic for Coll.!; / then G
is the corresponding generic lter.
(1.2) For all < !1 ,
MC1
/ P
411
if S 2 GC1 then
.P \ P / > 1=2. The lter GC1 is MC1 -generic and so
G D P 2 AM C1 j .S; P / 2 YC1 for some S 2 GC1
is a lter in AM C1 which is MC1 -generic. (Clearly G generates a generic lter
which is all we require. By Lemma 6.125, G literally is the lter it generates since
.!1 ; 0;1/M C1 2 YC1 .)
However for each P 2 G,
.P \ P / > 1=2:
It follows that
\P j P 2 G P :
This is an elementary property of the generic for Amoeba forcing. Let
XG D \P j P 2 G:
Then
.XG / D 1=2 and
.X \P / D 1=2. But if O 0;1 is open and O \XG ;
then
.XG \ O/ 0. Therefore
XG D XG \ P :
Finally
M C1
/ XG
M C1
/ P :
This veries that condition (1.2) can be met at every relevant stage. We consider
the effect of condition (1.1). Since g witnesses + .!1<! /, the set of < !1 such that
D !1M and g./ is M -generic for Coll.!; / contains a club in !1 . Therefore
< !1 j j0;!1 .f /./ D g./
contains a club in !1 . The situation here is similar to that in the proof of Lemma 6.23.
We nish by proving that
j0;!1 .Y / D YA .j0;!1 .F /; J / \ M!1 :
This is straightforward. We rst show that
YA .j0;!1 .F /; J / \ M!1 j0;!1 .Y /:
This is immediate by absoluteness and the fact that
J \ M!1 D I!1 :
Therefore we have only to show that
j0;!1 .Y / YA .j0;!1 .F /; J / \ M!1 :
Suppose .S; P / 2 j0;!1 .Y /. We must show that
.S; P / 2 YA .j0;!1 .F /; J /:
For this it sufces to show the following.
412
6 Pmax variations
413
(iii) !1
MkC1
D !1
(iv) Fk D F0 and fk D f0 ,
(v) IkC1 \ Mk D Ik ,
(vi) Yk D YkC1 \ Mk ,
(vii) if C 2 Mk is closed and unbounded in !1M0 then there exists D 2 MkC1 such
that D C , D is closed and unbounded in C and such that
D 2 Lx
for some x 2 R \ MkC1 .
Then there is an iteration
j W h.Mk ; Ik / W k < !i ! h.Mk ; Ik / W k < !i
such that
(1) j j.f0 /./ D g./ contains a club in !1 ,
(2) YA .j.F0 /; J / \ Mk D j.Yk /.
Proof. By Corollary 4.20 the sequence
h.Mk ; Ik / W k < !i
is iterable.
Given this the proof of the lemma is essentially identical to the proof of
Lemma 6.130.
Let hP W < !1 i be a sequence of conditions in A which is dense.
Let
hh.Mk ; Ik / W k < !i; G ; j; W < !1 i
be an iteration of h.Mk ; Ik / W k < !i such that the following hold.
414
6 Pmax variations
M
(1.1) For all < !1 if D !1 0 and if g./ is [Mk j k < !-generic for
Coll.!; / then G is the corresponding generic lter.
(1.2) For all < !1 ,
M0C1
/ P :
M0
j; C1 .F /.!1
/Q
415
t
u
416
6 Pmax variations
The basic analysis of M Qmax follows from these lemmas in a by now familiar fashion. The results of this we give in the following theorem. The analysis requires that
M
Qmax is suitably nontrivial. More precisely one needs that for each set A R with
A 2 L.R/ there exists
h.M; I /; f; F; Y i 2 M Qmax
such that
hH.!1 /M ; A \ H.!1 /M ; 2i hH.!1 /; A; 2i
and such that .M; I / is A-iterable.
By Lemma 6.47 and Lemma 6.127, this follows from the existence of a huge cardinal.
Theorem 6.132. Assume that for each set A R with A 2 L.R/ there exists
h.M; I /; f; F; Y i 2 M Qmax
such that
(i) hH.!1 /M ; A \ H.!1 /M ; 2i hH.!1 /; A; 2i,
(ii) .M; I / is A-iterable.
Then M Qmax is !-closed and homogeneous. Suppose
G M Qmax
is L.R/-generic. Then
L.R/G !1 -DC
and in L.R/G:
(1) P .!1 /G D P .!1 /;
(2) IG is a normal !1 -dense ideal on !1 ;
(3) IG is the nonstationary ideal;
(4) YG D YA .FG ; INS /;
(5) fG witnesses ++ .!1<! /;
(6) suppose B !1 is a set of measure 1. Then
j FG ./ B
contains a club in !1 ;
(7) the sentence AC holds.
417
Proof. The proof that M Qmax is !-closed follows closely the proof that Qmax is !closed.
Suppose hpk W k < !i is a strictly decreasing sequence of conditions in M Qmax and
that for each k < !,
pk D h.Mk ; Ik /; fk ; Fk ; Yk i:
Let
f D [fk j k < !1
and let
F D [Fk j k < !1 :
For each k < ! let
and let
In a similar fashion the other claims are proved by just adapting the proofs of the
corresponding claims for Qmax .
Because of the requirement (2) in the denition of M Qmax , (5) is immediate
from (1).
(4) is an immediate consequence of (1) and the denition of the order on M Qmax .
(6) follows from (4) by the denition of YA .FG ; INS /.
t
u
418
6 Pmax variations
There is a version of M Qmax analogous to Qmax for which the analysis can be carried
out just assuming ADL.R/ . This version is a little tedious to dene and we leave the
details to the reader. The net effect of this is the following theorem that M Qmax is
suitably nontrivial just assuming ADL.R/ . This is analogous to Theorem 6.80.
Theorem 6.133. Assume ADL.R/ . Then for each set A R with A 2 L.R/ there
exists
h.M; I /; f; F; Y i 2 M Qmax
such that
(1) hH.!1 /M ; A \ H.!1 /M ; 2i hH.!1 /; A; 2i,
t
u
(3) there is a sequence hB W < !1 i of borel subsets of 0;1 such that each B
is of measure 1 and such that if B 0;1 is set of measure 1 then B B for
t
u
some < !1 .
There are absoluteness theorems for M Qmax analogous to the absoluteness theorems
for Qmax . These require the following preliminary lemmas. With these lemmas in hand
the proof of the absoluteness theorem, Theorem 6.139, is an easy variation of the proof
of the corresponding theorem for Qmax , Theorem 6.85. We leave the details as an
exercise.
We generalize the denition of YA .F; INS / to the setting of the stationary tower.
Suppose that is strongly inaccessible and that
F W !1 ! P .0;1/:
Let YA .F; / be the set of all pairs .S; P / such that the following hold.
(1) S 2 Q< and !1 [S .
(2) P 0;1 and P 2 A.
419
t
u
The next two lemmas, Lemma 6.136 and Lemma 6.137, are used to prove the
iteration lemma, Lemma 6.138, just as Lemma 6.125 and Lemma 6.129 are used to
prove the basic iteration lemmas for M Qmax , Lemma 6.130.
Lemma 6.136. Suppose is strongly inaccessible
F W !1 ! P .0;1/
is a function such that YA .F; / is nonempty. Suppose .S1 ; P1 / 2 YA .F; /.
(1) Suppose P2 is a perfect subset of P1 and P2 2 A. Let
S2 D a 2 S1 j F .a \ !1 / P2 :
Then .S2 ; P2 / 2 YA .F; /.
(2) Suppose that S2 S1 in Q< . Then there exists .S3 ; P3 / 2 YA .F; / such that
S3 S 2 .
Proof. This is the analog of Lemma 6.125. The proof is similar. For (2) one uses the
generic ultrapower associated to Q< in place of the generic ultrapower associated to
t
u
P .!1 /=I .
Lemma 6.137. Suppose .M; I/ is a countable iterable structure such that
(i) M ZFC,
(ii) I 2 M and I is the tower of ideals I< as computed in M where is a Woodin
cardinal in M .
420
6 Pmax variations
A .Q< /M ;
t
u
Using Lemma 6.137 the basic iteration lemma is easily proved. The proof follows
that of Lemma 6.130 using Lemma 6.137 in place of Lemma 6.129.
Lemma 6.138. Suppose
F W !1 ! P .0;1/
is a function such that .!1 ; 0;1/ 2 YA .F; INS /. Suppose
H W !1 ! H.!1 /
is a function which witnesses .!1<! /.
Suppose .M; I/ is a countable iterable structure such that
+
(i) M ZFC,
(ii) I 2 M and I is the tower of ideals I< as computed in M where is a Woodin
cardinal in M .
421
t
u
Theorem 6.139 is an absoluteness theorem for M Qmax . Again the proof is an easy
adaptation of earlier arguments and stronger absoluteness theorems can be proved.
For this theorem one uses the iteration lemma, Lemma 6.138, modifying the proof
of the corresponding absoluteness theorem for Qmax , Theorem 6.85. The situation here
is simpler since there are no restricted 2 sentences to deal with.
Theorem 6.139 (+ .!1<! /). Suppose that there are ! many Woodin cardinals with a
measurable above them all and that there is a sequence hB W < !1 i of borel subsets
of 0;1 such that each B is of measure 1 and such that if B 0;1 is set of measure 1
then B B for some < !1 .
Suppose is a 2 sentence in the language for the structure
hH.!2 /; X; 2 W X R; X 2 L.R/i
and that
hH.!2 /; X; 2 W X R; X 2 L.R/i :
Then
MQ
hH.!2 /; X; 2 W X R; X 2 L.R/iL.R/
6.3
max
:
t
u
Nonregular ultralters on !1
We consider ultralters on !1 .
Denition 6.140. Suppose that U is a uniform ultralter on !1 .
(1) The ultralter U is nonregular if for each set W U of cardinality !1 there
exists an innite set Z W such that
\Z ;:
422
6 Pmax variations
t
u
We begin with the basic relationship between the existence nonregular ultralters
on !1 and the existence of weakly normal ultralters on !1 . This relationship is summarized in the following theorem of Taylor .1979/. This theorem is the analog for !1
of the theorem that if is measurable then there is a normal measure on .
Theorem 6.141 (Taylor). Suppose that U is a uniform ultralter on !1 .
(1) Suppose that U is weakly normal. Then U is nonregular.
(2) Suppose that U is nonregular. Then there exists a function
f W !1 ! !1
such that U is weakly normal where
U D A !1 j f 1 A 2 U :
t
u
The relative consistency of the existence of nonregular ultralters on !1 rst established by Laver. Laver proved that if there exists an !1 -dense uniform ideal on !1 and
holds, then there exists a nonregular ultralter on !1 .
Huberich improved Lavers theorem proving the theorem without assuming .
Thus in L.R/Qmax there is a nonregular ultralter on !1 .
The basic method for producing nonregular ultralters on !1 is to produce them
from suitably saturated normal ideals on !1 . The approach is due to Laver and involves
the construction of indecomposable ultralters on the quotient algebra,
P .!1 /=I:
Denition 6.142. Suppose that B is a countably complete boolean algebra. An ultralter
U B
is indecomposable if for all X B,
_X 2 U
if and only if
_Y 2 U
for some countable set Y X .
t
u
The fundamental connection between normal ideals on !1 and nonregular ultralters on !1 is given in the following lemma due to Laver.
423
t
u
The following theorem was rst proved by Laver assuming and then by Huberich,
.Huberich 1996/, without any additional assumptions.
Theorem 6.144 (Huberich). Let
B D RO.Coll.!; !1 //:
Then there is an ultralter U on B which is indecomposable.
t
u
424
6 Pmax variations
425
t
u
The following theorem, in conjunction with Theorem 6.148, completes the analysis of
the consistency strength of the assertion that there exists an !1 -dense ideal on !1 . The
proof of this theorem involves the core model induction which is also the method used
to prove Theorem 5.111 and as noted is beyond the scope of this book.
Theorem 6.149. Suppose that I is a normal, uniform, ideal on !1 such that I is !1 dense. Then
L.R/ AD:
t
u
Corollary 6.150. The following are equiconsistent:
(1) ZF C AD.
(2) ZFC C INS is !1 -dense.
(3) ZFC C There is a normal, uniform, !1 -dense ideal on !1 .
t
u
Chapter 7
Conditional variations
In this chapter we dene two conditional variations of Pmax . The models obtained are
in essence simply conditional versions of the Pmax -extension, i. e. the models maximize
the collection of 2 sentences which can hold in the structure
hH.!2 /; INS ; 2i
given that some specied sentence holds.
The Qmax -extension is an example of such a variation. It conditions the extension
on the assertion that the nonstationary ideal is !1 -dense.
There is an analogy for these conditional variations with variations of Sacks forcing. Suppose is a 13 sentence which is true in V and that there is a function
f W ! ! ! which eventually dominates all those functions which are constructible.
Then is true in LP where P is Laver forcing. This can be proved by a modication of Manselds argument. Thus the Laver extension of L realizes all possible 13
sentences conditioned on the existence of fast functions.
These variations of Pmax also yield models in which conditional forms of Martins
Maximum hold. For example we shall dene a variation Bmax such that in L.R/Bmax
the Borel Conjecture holds together with a large fragment of Martins Axiom.
7.1
Suslin trees
427
Denition 7.2. C
S : For every set X !1 there exists Y !1 such that X 2 LY
and such that every tree T 2 LY which is a Suslin tree in LY is a Suslin tree in
V.
t
u
This strong version of S seems quite subtle in the context of large cardinals. For
example assuming for all A !1 , A# exists; it is not obvious that it can even hold. We
prove that if for all A !1 , A# exists then C
S implies :CH.
Lemma 7.3. Suppose that A !1 is a set such that R LA and such that A#
exists. Then there is a tree T 2 LA such that T is a Suslin tree in LA and such that
T has a conal branch.
Proof. We naturally view any .!1 ; !1 / tree as an order on !1 ! such that for each
2 !1 , ! is the set of nodes in T on the th level. We restrict our considerations
to trees with only innite levels and which are splitting.
We may suppose that for all < !1 ,
!1LA\ :
Let
T D .!1 !; <T /
be such a tree such that T 2 LA and such that T satises the following condition.
For each < !1 we let T< denote the restriction of T to the rst levels. Thus
T< D . !; <T j . !//:
The condition is:
(1.1) For all < !1 , the set of conal branches of T< with upper bounds in T is
exactly the set of branches of T< which are LA \ ; T< -generic for T<
and which belong to L.A \ ; T< /# .
We prove that such a tree exists in LA. The point is that since R LA, and
since A# exists, for all x 2 R \ LA, x # 2 LA. The tree T is easily constructed in
LA by induction on the levels of T provided we verify the following claim.
Suppose that T< is given and that (1.1) holds for all < . Then for all p 2 T<
there is an LA \ ; T< -generic branch of T< such that p 2 b and such that
b 2 L.A \ ; T< /# :
There are 2 cases. First suppose that is countable in L.A \ ; T< /# . Then it
follows that
P ./ \ LA \ ; T<
is countable in L.A \ ; T< /# . The existence of the generic branch b is immediate.
The second case is that
#
D .!1 /L.A\;T< / :
428
7 Conditional variations
It follows that T< is a Suslin tree in LA \ ; T< and that there is a conal branch
b T< such that p 2 b and such that
b 2 L.A \ ; T< /# :
However T< is a Suslin tree in LA \ ; T< and so the branch b is necessarily
LA \ ; T< -generic.
This veries the claim and so it follows that the tree T exists. Necessarily T is a
Suslin tree in LA and repeating the argument for the claim, T has a conal branch in
t
u
LA# .
We now dene the partial order Smax . It in essence is just Pmax with a more restrictive ordering though we modify the fragment of MA which is to hold in the models
occurring in the conditions. Recall that a partial order P is -centered if P is the
countable union of sets, S P , with the property that if a S is nite then there
exists q 2 P such that q p for all p 2 a. For example if P the union of countably
many lters then P is -centered. MA!1 . -centered/ is the variant of Martins Axiom
which asserts that if P is -centered and if D is a collection of dense subsets of P with
jDj !1 then there is a lter F P which is D-generic.
We note that Lemma 4.35 holds for countable transitive models M such that
M ZFC C MA!1 . -centered/:
Thus if .M; I / is iterable and if a !1M is such that a 2 M and such that
!1M D .!1 /L a
where D M \ Ord, then iterations of .M; I / are uniquely determined by the image
of a.
Denition 7.4. Let Smax be the set of pairs h.M; I /; ai such that:
(1) M is a countable transitive model of ZFC C MA!1 . -centered/;
(2) I 2 M and M I is a normal uniform ideal on !1 ;
(3) .M; I / is iterable;
(4) a !1M ;
(5) a 2 M and M !1 D !1Lax for some real x.
Dene a partial order on Smax as follows.
h.M1 ; I1 /; a1 i < h.M0 ; I0 /; a0 i
if M0 2 M1 ; M0 is countable in M1 and there exists an iteration
j W .M0 ; I0 / ! .M0 ; I0 /
such that:
(1) j.a0 / D a1 ;
(2) M0 2 M1 and j 2 M1 ;
(3) I1 \ M0 D I0 ;
(4) For any T 2 M0 , if T is a Suslin tree in M0 then T is a Suslin tree in M1 .
t
u
429
Lemma 7.7 is the iteration lemma that essentially allows the proofs for Pmax to
generalize in a straightforward fashion. As usual it is really the proof of the lemma
that is important.
Before proving Lemma 7.7 we prove two useful technical lemmas.
Lemma 7.5. Suppose that I is a normal uniform ideal on !1 , S !1 is I -positive
and that T is a Suslin tree. Suppose that
f WS !T
is a function such that for all for all 2 S ,
f ./ 2 T ;
where T denotes the
with a < b,
th
is I -positive.
Proof. For each b 2 T let
Sb D 2 S j b < f ./;
and let
D D b 2 T j Sb 2 I :
For each < !1 let T< be the subtree of T obtained by restricting T to the rst
many levels of T ; i. e.
T< D [T j < :
Let
S D 2 S j Sb for all b 2 D \ T :
Since the ideal I is normal it follows that
S n S 2 I:
Assume toward a contradiction that D is dense in T .
Suppose < !1 and that a 2 T . Let ba be the conal branch of T< dened
by a.
Since T is a Suslin tree it follows that
< !1 j for all a 2 T ; D \ ba ;
contains a club in !1 .
Therefore there exists 2 S such that bf ./ \ D ; which is a contradiction.
Therefore D is not dense and this proves the lemma.
t
u
430
7 Conditional variations
431
Fix a sequence hAk; W k < !; < !1 i of J positive sets which are pairwise
disjoint and disjoint from S . The ideal J is normal hence each Ak; is stationary in
!1 .
Fix a function
f W ! !1M ! M
such that
(1.1) f is onto,
(1.2) for all k < !, f jk !1M 2 M,
(1.3) for all A 2 M if A has cardinality !1M in M then A ran.f jk !1M / for
some k < !.
The function f is simply used to anticipate elements in the nal model.
Suppose
j W .M; I / ! .M ; I /
is an iteration. Then we dene
j.f / D [j.f jk !1M / j k < !
and it is easily veried that M is the range of j.f /. This follows from (1.3).
We construct an iteration of M of length !1 using the function f to provide a
book-keeping device for all of the subsets of !1 which belong to the nal model and
do not belong to the image of I in the nal model. Implicit in what follows is that for
2 C if
j W .M; I / ! .M ; I /
is an iteration of length then j.!1M / D . This is a consequence of Lemma 4.6(1).
More precisely construct an iteration h.M ; I /; G ; j; W < !1 i such that
for each < !1 , if
(2.1) 2 Ak; \ C ,
(2.2) < ,
(2.3) j0; .f /.k; / ,
(2.4) j0; .f /.k; / I ,
then j0; .f /.k; / 2 G .
These requirements place no constraint on the choice of G for 2 S \ C . For
2 C \ S choose G such that if codes .k; ; T; / where
(3.1) < ,
(3.2) T is a Suslin tree in M ,
(3.3) j0; .f /.k; / D T ,
(3.4) is dense in T ,
then is predense in j; C1 .T /. Lemma 7.6 shows G exists.
432
7 Conditional variations
433
The analysis of the Smax -extension requires the generalization of Corollary 7.8 to
sequences of models. This in turn requires the generalization of Lemma 7.6 to sequences of models.
Lemma 7.9 (ZFC). Suppose h.Mk ; Ik / W k < !i is an iterable sequence of countable
structures such that for all k < !;
(i) Mk is a countable transitive model of ZFC,
(ii) Mk 2 MkC1 ,
(iii) jMk jMkC1 D .!1 /MkC1 ,
Mk
(iv) !1
MkC1
D !1
(v) IkC1 \ Mk D Ik ,
(vi) suppose Ak 2 Mk ,
Ak P .!1 k / \ Mk n Ik
and that Ak is predense, then Ak is predense in
.P .!1 / n IkC1 ; /MkC1 :
Suppose that
T 2 [Mk j k < !
is a Suslin tree in [Mk j k < !, and that
T
is a dense subset of T . Then there is an iteration
j W h.Mk ; Ik / W k < !i ! h.Mk ; Ik / W k < !i
of length 1 such that is predense in j.T /.
Proof. We suppose that T 2 M0 . For each
< !1M0
let T denote the th level of T .
Let
hxk W k < !i
be an enumeration of [Mk j k < ! such that for all k < !, xk 2 Mk . By successive
applications of Lemma 7.5 there exists a sequence
hsk W k < !i
such that for each k < !,
(1.1) skC1 sk ,
434
7 Conditional variations
(1.2) sk !1M0 ,
(1.3) sk 2 Mk n Ik ,
(1.4) if xk !1M0 then sk D xk or sk D !1M0 n xk ,
(1.5) if xk is a predense subset of
then sk s for some s 2 xk ,
(1.6) if xk is a function
f W !1M0 ! T
.P .!1 /Mk n Ik ; /
f ./ 2 T ;
Let
G D sk j k < !:
By (1.4) and (1.5), for each k < !, G \ Mk is Mk -generic for .P .!1 / n Ik /Mk .
Since the sequence, h.Mk ; Ik / W k < !i, is iterable, G denes an iteration
j W h.Mk ; Ik / W k < !i ! h.Mk ; Ik / W k < !i
of length 1 where for each k < !, Mk is transitive.
By (1.6), is predense in j.T /.
t
u
Lemma 7.10 (ZFC). Assume S . Suppose J is a normal uniform ideal on !1 . Suppose that h.M; Ik / W k < !i is a sequence such that for each k < !, Mk is a countable
transitive model of ZFC, Ik 2 Mk and such that in Mk , Ik is a uniform normal ideal
M
on !1 k . Suppose that for all k < !,
(i) Mk 2 MkC1 ,
(ii) jMk jMkC1 D .!1 /MkC1 ,
Mk
(iii) !1
MkC1
D !1
(iv) IkC1 \ Mk D Ik ,
(v) .Mk ; Ik / is iterable,
(vi) suppose Ak 2 Mk ,
Ak P .!1 k / \ Mk n Ik
and that Ak is predense, then Ak is predense in
.P .!1 / n IkC1 ; /MkC1 ;
(vii) if C 2 Mk is closed and unbounded in !1M0 then there exists D 2 MkC1 such
that D C , D is closed and unbounded in C and such that
D 2 Lx
for some x 2 R \ MkC1 .
435
J \ Mk D Ik :
(3) Suppose that T 2 [Mk j k < ! and that for all k < !, if T 2 Mk then T is
a Suslin tree in Mk . Then T is a Suslin tree.
Proof. By Corollary 4.20 the sequence
h.Mk ; Ik / W k < !i
is iterable.
The lemma follows by a construction essentially the same as that given in the
proof of Corollary 7.8. For this proof the construction uses Lemma 7.9 in place of
Lemma 7.6.
t
u
As we have indicated, the nontriviality of Smax is an immediate corollary to nontriviality of Pmax . However we shall need a slightly stronger version of this. The reason is
simply that the iteration lemmas for Smax require additional assumptions. The situation
here, though much simpler, is reminiscent of that in Section 6.2.5 where we analyzed
KT
Qmax . Smax is a renement of Pmax which is analogous to KT Qmax as a renement of
Qmax .
Lemma 7.11. Assume AD holds in L.R/. Suppose X R and that
X 2 L.R/:
Then there is a condition h.M; I /; ai 2 Smax such that
(1) M S ,
(2) X \ M 2 M,
(3) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(4) .M; I / is X -iterable,
and further the set of such conditions is dense in Smax .
Proof. The proof is identical to the proof of Lemma 4.40 with one slight change.
We use the notation from that proof. The modication concerns the choice of the
partial order Q 2 N . For this proof one chooses Q 2 N such that
N Q ZFC C MA!1 . -centered/:
436
7 Conditional variations
437
!1M0
is the unique iteration such that j.a0 / D AG . We work in L.R/ and we assume that
the condition h.M0 ; I0 /; a0 i forces that X0 D j.b0 / is a counterexample to (5).
Let z 2 R be any real such that M0 2 Lz and M0 is countable in Lz. For each
i ! let i be the i th Silver indiscernible of Lz. Let
k W L! z ! L! z
be the canonical embedding such that cp(k) = 0 and let
L! z D k.f /.0 / j f 2 L! z:
Let U be the L! z-ultralter on 0 given by k,
U D A 0 j A 2 L! z; 0 2 k.A/:
Thus
L! z Ult.L! z; U /
and k is the associated embedding. For each X P .0 / \ L! z if X 2 L! z and
jX j 0 in L! z then U \ X 2 L! z. Therefore .L! z; U / is naturally iterable.
Let g Coll.!; <0 / be L! z-generic. Let N D L! zg. Therefore
0 D !1N and the ultralter U denes an ideal I on !1N with I N . Further for
each X 2 N if jX j !1N in N then I \ X 2 N .
438
7 Conditional variations
439
is a function such that f 2 N and such that for all for all 2 S ,
f ./ 2 T ;
where T denotes the
b 2 T with a < b,
th
is I -positive.
Thus the proof Lemma 7.7 can be applied to .N; I /.
Let h.M2 ; I2 /; ai be any condition in Smax such that N 2 M2 , N is countable in
M2 and such that holds in M2 . Then there is an iteration
k W .N; I / ! .N ; I /
in M2 such that
(3.1) k .!1N / D .!1 /M2 ,
(3.2) I2 \ N D I ,
(3.3) if T 2 N and if T is a Suslin tree in N then T is a Suslin tree in M2 .
Let a2 D k .a1 /, b2 D k .b1 / and let y2 D k .y1 /. Thus
(4.1) h.M2 ; I2 /; a2 i < h.M0 ; I0 /; a0 i,
(4.2) b2 D j.b0 / where j is the embedding given by the iteration of .M0 ; I0 / which
sends a0 to a2 ,
(4.3) b2 .!1 /M2 ,
(4.4) y2 .!1 /M2 ,
(4.5) b2 2 .Ly2 /M2 ,
(4.6) if T 2 Ly2 is a Suslin tree in Ly2 then T is a Suslin tree in M2 .
440
7 Conditional variations
Smax
:
t
u
As in the case of Pmax the converse also holds in the sense of Theorem 4.76. The
proof requires the version of Lemma 4.74 for conditions in Smax .
Theorem 7.15. Assume ADL.R/ . Suppose that for each 2 sentence in the language
for the structure
hH.!2 /; 2; INS ; T ; X I X 2 L.R/; X Ri
if
Smax
441
then
hH.!2 /; 2; INS ; T ; X I X 2 L.R/; X Ri :
Then L.P .!1 // D L.R/G for some G Smax which is L.R/-generic.
7.2
t
u
AC .I /:
!1 n A 2 I
AC .I /
AC .I /
AC
t
u
to I . A key difference
t
u
442
7 Conditional variations
The following theorem is a slight variation of Theorem 2.65. We shall use Theorem 7.18 to construct conditions in Bmax .
Theorem 7.18. Suppose that is a Woodin cardinal and that
G Coll.!1 ; </
is V -generic. Then in V G there is a normal, uniform, ideal I on !1 such that
(1) I \ V D .INS /V ,
(2) I is !2 -saturated in V G,
(3) V G
AC .I /.
Proof. The proof is quite similar to that of Theorem 2.65. We sketch the argument.
As before the ideal I is rather easy to dene. Suppose that
G Coll.!1 ; </
is V -generic. For each < let
G D G \ Coll.!1 ; </:
Let mc.V / be the set of < such that is a measurable cardinal in V . For each
2 mc.V / we dene a normal ideal
I 2 V GC1
as follows by induction on . Fix a wellordering, in V , of V .
There are two cases. We rst suppose that 2 mc.V / and that is not a successor
element of mc.V /.
In this case we dene
J 2 V GC1
to be the set of A !1 such that for some
f W !1 ! P .!1 / n INS ;
(1.1) A D < !1 j f ./ for all < ,
(1.2) if A D f ./ j < !1 then
A 2 V G ;
and A is semiproper in V G .
I is the normal ideal in V GC1 generated by
J [ I<
where
I< D [I j 2 \ mc.V /:
The second case is that 2 mc.V / and that
mc.V / \
443
AC .I /
AC .I /
and let
f W !1 !
be a surjection with f 2 V GC1 . The pair .S; T / is chosen using the wellordering
of V G induced by the chosen wellordering of V .
Let A D A0 [ A1 where
A0 D 2 !1 j 2 S and ordertype.f / T
and
A1 D 2 !1 j S and ordertype.f / 2 T :
We dene I to be the normal ideal in V GC1 generated by
I [ A:
Let I be the normal ideal generated in V G by
[I j 2 mc.V /:
Here as in the proof of Theorem 2.65, the only difculty is to verify that I is a proper
ideal. If I is a proper ideal then arguing as in the proof of Theorem 2.65, I is a
saturated ideal and further it follows easily that
V G
AC .I /:
444
7 Conditional variations
AC .I /
445
a) if
\ mc.V /
has no maximum element then (2.2) is satised by X at , otherwise,
b) (2.3) is satised by X at .
For the construction of this elementary chain we note that the requirements corresponding to the desired properties, (3.5(a)) and (3.5(b)), do not conict. The requirement
which yields (3.5(a)) is easily handled using the denition of a semiproper subset of
P .!1 / n INS . The requirement for (3.5(b)) is handled using the following observation.
Suppose
Y M
is a countable elementary substructure and that 2 Y \ is a measurable cardinal.
Then there exists a closed unbounded set
C !1
such that for each 2 C there exists
Y M
such that Y Y , such that Y \ is an initial segment of Y \ and such that
Y \
has ordertype . Now suppose that
2 V Coll.!1 ;< C1/
is a term for a stationary, co-stationary subset of !1 and that
q 2 Coll.!1 ; < C 1/:
Then for conally many 2 C there is a condition q < q such that
q 2
and for conally many 2 C there is a condition q < q such that
q :
With this observation (3.5(b)) is easily achieved.
Now suppose
G Coll.!1 ; </
is V -generic and that p 2 G .
Since
q 2 X \ Coll.!1 ; <0 C 1/ j p < q
446
7 Conditional variations
We prove that AC .I / does allow one to recover iterations from iterates. The proof
is quite similar to that of Lemma 5.15.
We state the lemma only in the special form that we shall need.
Lemma 7.19. Suppose M is a countable transitive set such that
M ZFC C
where I 2 M is in M, a normal ideal on
!1M .
AC .I /
Suppose a 2 M,
a !1M ;
and for some x 2 M \ R,
M !1 D !1La;x :
Suppose
j1 W .M; I / ! .M1 ; I1 /
and
j2 W .M; I / ! .M2 ; I2 /
are iterations of M such that M1 is transitive, M2 is transitive and such that
j1 .a/ D j2 .a/:
Then M1 D M2 and j1 D j2 .
Proof. Fix a and x. Clearly we may suppose that either j1 or j2 is not the identity.
Therefore since M contains precipitous ideals,
M For every set A !1 , A# exists:
Therefore
447
sup.a/ D !1M
Thus N 2 M,
!1M D !1N
and
j1 .N / D j2 .N /:
Let b
!1M
X .0;1/ j X is Z-small
448
7 Conditional variations
is h0 -small.
The denition of the partial order Bmax is motivated by the following reformulation
of the Borel Conjecture:
Suppose X .0;1/ is uncountable, then there exists h W ! ! ! and there
exists a function f W !1 ! X such that if O .0;1/ is open and h-small then
j f ./ O is stationary.
That this is a reformulation of the Borel Conjecture is a corollary to the following
theorem due to J. Zapletal.
Theorem 7.23 (Zapletal). Suppose that
I P .0;1/
is a -ideal and that X 0;1 is a set of cardinality @1 such that X I .
Then there exists a function
f W !1 ! X
such that for all Y 2 I
2 !1 j f ./ Y
is stationary.
Proof. Fix a surjection
W !1 ! X
and dene
W !1 ! P!1 .X /
by
./ D ./ j :
Thus since X I , for all Y 2 I ,
2 !1 j ./ Y
is countable.
449
Choose functions
fi W !1 ! X
for each i < !, such that for all < !1 ,
./ D fi ./ j i < !:
We claim that one of the functions fi is as desired. Otherwise for each i < ! there
exist Yi 2 I and a club Ci !1 such that
Ci < !1 j fi ./ 2 Yi :
Let
C D \Ci j i < !;
and let
Y D [Yi j i < !:
Since I is a -ideal, Y 2 I . However for each 2 C , ./ Y . Therefore
X Y
t
u
which is a contradiction.
A similar argument proves the following theorem.
I D Y 0;1 j Y is Z-small
is a -ideal. The set Z exists by Lemma 7.22.
Thus since X I , for all Y 2 I ,
2 !1 j ./ Y
is countable.
450
7 Conditional variations
Choose functions
fi W !1 ! X
for each i < !, such that for all < !1 ,
./ D fi ./ j i < !:
Let hhj W j < !i enumerate Z.
We claim that for some i; j 2 ! the pair .fi ; hj / is as desired. Otherwise for each
.i; j / 2 ! ! there exist Yi;j 0;1 and a club Ci;j !1 such that
Ci;j < !1 j fi ./ 2 Yi;j
and such that Yi;j is hj -small.
For each i < ! let
Ci D \Ci;j j j < !;
and let
Yi D \Yi;j j j < !:
Thus for each i < !, Yi is Z-small and
Ci < !1 j fi ./ 2 Yi :
Finally let
C D \Ci j i < !;
and let
Y D [Yi j i < !:
Since I is a -ideal, Y 2 I and so Y is Z-small. However for each 2 C ,
./ Y . Therefore
X Y
which is a contradiction since X is not Z-small.
t
u
Remark 7.25. We originally proved Theorem 7.24 with the additional hypothesis that
t
u
the nonstationary ideal on !1 is !2 -saturated.
Suppose h 2 ! ! . Let hE be the set of functions f 2 ! ! such that for some
i; j 2 !,
h.i C k/ D f .j C k/
for all k 2 !. Thus X 0;1 is hE -small if and only if X is h.m/ j m < !-small
where for each m < !,
h.m/ .k/ D h.m C k/
for k < !.
Denition 7.26. Suppose I is a uniform normal ideal on !1 . ZBC .I / is the set of all
pairs
.h.fi ; hi / W i < ni; S /
such that the following hold.
451
t
u
where:
a) p.0/ D 0 and p.k/ < p.k C 1/ for all k < !;
b) for some m 2 !, m > 1 and p.k/ D k m for all sufciently large k < !;
c) for all i < n and for all j < !,
hi .j / D hi .k/
where k < ! is such that p.k/ j < p.k C 1/.
t
u
452
7 Conditional variations
where:
(1) p.0/ D 0 and p.k/ < p.k C 1/ for all k < !;
(2) for some m 2 !, m > 1 and p.k/ D k m for all sufciently large k < !;
(3) for all i < n and for all j < !,
hi .j / D hi .k/
where k < ! is such that p.k/ j < p.k C 1/.
t
u
Then
.h.fi ; hi / W i < ni; S / 2 YBC .I /:
Proof. Dene for each ordinal , a subset
ZBC
.I /
as follows.
0
ZBC
.I / D ZBC .I / and if is a limit ordinal then
.I / D \ZBC
.I / j < :
ZBC
C1
Finally for each ordinal , ZBC
.I / is the set of
t
u
453
such that:
.I / and such
(1.1) for each i < nO there exists g 2 ! ! such that .h.fOi ; g/i; SO / 2 ZBC
that for sufciently large k 2 !,
g.j / hO i .k/
for all j < 5k ;
(1.2) for some p 2 ! ! ,
O SO / 2 ZBC
.I /
.h.fOi ; hO i / W i < ni;
where:
a) p.0/ D 0 and p.k/ < p.k C 1/ for all k < !;
b) for some m 2 !, m > 1 and
p.k/ D k m
for all sufciently large k < !;
c) for all i < nO and for all j < !,
hO i .j / D hO i .k/
where k < ! is such that p.k/ j < p.k C 1/.
Thus for for sufciently large ,
.I / D YBC .I /:
ZBC
Fix
.h.fi ; hi / W i < ni; S / 2 ZBC .I /
satisfying the conditions of the lemma and assume toward a contradiction that
.h.fi ; hi / W i < ni; S / YBC .I /:
Thus for some ordinal ,
.I /:
.h.fi ; hi / W i < ni; S / ZBC
We may suppose that the choice of .h.fi ; hi / W i < ni; S / minimizes . Thus is
a successor ordinal. Let 0 be such that D 0 C 1.
Let p 2 ! ! be a function such that
.I /
.h.fi ; hi / W i < ni; S / 2 YBC
where:
(2.1) p.0/ D 0 and p.k/ < p.k C 1/ for all k < !;
(2.2) for some m 2 !, m > 1 and
for all sufciently large k < !;
p.k/ D k m
454
7 Conditional variations
.I /
.h.fi ; hi /i; S / 2 YBC
t
u
a contradiction.
Denition 7.31. Bmax consists of nite sequences h.M; I /; a; Y i such that the following hold.
(1) M is a countable transitive model of ZFC.
(2) M CH.
(3) I 2 M and M I is a normal uniform ideal on !1 .
(4) M
AC .I /.
M1 2 H.!1 /M2
455
t
u
Remark 7.32. (1) The only reason for requiring that the models occurring in the
Bmax conditions actually be models of ZFC instead of ZFC is so that (6) in the
denition of Bmax is unambiguous. The trivial point is that ZFC does not prove
that YBC .I / exists.
(2) There is actually a parameterized family of generalizations of Bmax . Fix a function h 2 ! ! . Let Bmax .h/ be the suborder of Bmax consisting of those conditions
h.M; I /; a; Y i such that if g 2 ! ! and g occurs in Y then for sufciently large
i 2 !, h.i / < g.i /.
t
u
We prove the basic iteration lemmas for Bmax . There are two iteration lemmas, one
for models and one for sequences of models. The latter is necessary to show that Bmax
is !-closed and its proof is an intrinsic part of the analysis of Bmax just as in the case
of Pmax .
We need several preliminary lemmas. For all m 2 ! and for all h 2 ! ! , let h.m/
be the function obtained by shifting h,
h.m/ .k/ D h.m C k/
for k 2 !. Thus a set X 0;1 is hE -small if and only if X is h.m/ j m < !-small.
We note that if
.h.fi ; hi / W i < ni; S / 2 ZBC .I /
then for all m 2 !,
.h.fi ; h.m/
i / W i < ni; S / 2 ZBC .I /:
.h.fi ; h.m/
i / W i < ni; S / 2 YBC .I /:
This is easily veried noting the following. Suppose g 2 ! ! and h 2 ! ! are such that
for all k m0 ,
g.j / h.k/
456
7 Conditional variations
457
Fix n 2 !.
It sufces to show that
Ijm j m > n; j 2 !
witnesses that Xn is g .n/ -small.
We note that for each a 2 ! with a > n,
X
2k D
kCmDa;am>n
a.nC1/
X
2k D 2an
1 2aC1
n
kD0
.I /:
.h.f; g/i; S / 2 YBC
t
u
458
7 Conditional variations
s.k/ if k < m,
s h.k/ D
h.k/ if k m,
where m D dom.s/.
Suppose g0 2 ! ! and that
.h.f0 ; g0 /i; S / 2 YBC .I /:
459
<!
such that
h.f0 ; s .g .m0 / //; T i 2 YBC
.I /:
460
7 Conditional variations
t
u
.h.fi ; h.m/
i /i; T / 2 YBC .I /:
Assume the lemma fails.
Thus by Lemma 7.30, for each m 2 !,
.h.fi ; h.m/
i / W i < n C 1i; T / YBC .I /:
.I / as a renement of ZBC .I /,
Therefore as a consequence of the denition of YBC
for each m 2 ! there exists a sequence
hOim W i < n C 1i
of open sets such that
(1.1) T n 2 T j fi ./ 2 Oim for some i < n C 1 2 I ,
(1.2) for all i < n C 1, Oim is Hi;m -small.
The ideal I is countably complete and so there is a set T1 T such that T nT1 2 I
and such that for all m 2 ! and for all 2 T1 ,
fi ./ 2 Oim
for some i < n C 1.
461
t
u
462
7 Conditional variations
(1.1) For all i < n, ti D h.rki ; ski / W k < mi where for all k < m,
a) 0 rki < ski 1,
b) rki 2 Q,
c) ski 2 Q,
d) .ski
rki / < 1=.hi .k/ C 1/.
(1.2) For all T S with T 2 A, either
.h.fi ; h.m/
i / W i < n C 1i; T / Y;
or for some i < n and for some 2 T ,
fi ./ 2 [.rki ; ski / j k < m:
The ordering on T is by (pointwise) extension,
hsi W i < ni hti W i < ni
if ti si for all i < n.
Clearly T 2 M.
Suppose T has an innite branch. Then by absoluteness, T has an innite branch
in M. We work in M and assume toward a contradiction that T has an innite branch.
Any such branch yields for each i < n a sequence
h.rki ; ski / W k < !i
of open intervals in .0;1/ with rational endpoints such that for all i < n and for all
k < !,
jski
rki j < 1=.hi .k/ C 1/:
These sequences have the additional property that for all T 2 A such that T S and
for all m < ! either
.h.fi ; h.m/
i / W i < n C 1i; T / Y;
or for some i < n and for some 2 T ,
fi ./ 2 [.rki ; ski / j k < m:
For each i < n let
OQ i D [.rki ; ski / j k < !:
Thus for each i < n, OQ i is hi -small.
Let
T0 D 2 S j fi ./ OQ i for all i < n:
Since .h.fi ; hi / W i < ni; S / 2 Y ,
T0 I:
463
The iteration lemmas are proved using the following lemmas which in turn follow
rather easily from the previous lemmas.
Lemma 7.38. Suppose h.M; I /; a; Y i 2 Bmax . Suppose .h.fi ; hi / W i < ni; S / is an
element of Y and suppose
h.fi ; hi ; Si / W i < !i
is a sequence extending h.fi ; hi ; S / W i < ni such that for each i < ! if i n then
.h.fi ; hi /i; Si / 2 Y . Suppose
hBi W i < !i
is a sequence of borel sets such that each i < !, if i < n then Bi is hi -small and if
i n then Bi is hi E -small. Then there is an iteration
j W .M; I / ! .M ; I /
of length 1 such that
(1) for all i < !, if !1M 2 j.Si / then j.fi /.!1M / Bi ,
(2) !1M 2 j.S /.
Proof. Let hAi W n i < !i enumerate the sets in M which are dense in
.P .!1 / n I /M .
Using Lemma 7.37 it is straightforward to build sequences
hTi W i < !i;
and
hNi W i < !i
(1.6) .h.fj ; hj
/ W j 2 Z \ i i; Ti / 2 Y .
464
7 Conditional variations
We rst construct
hTi W i ni;
and
hNi W i ni:
and
hNi W i mi
i
are given. For each i < m and k < ! let Jki D IkCN
. Therefore
m
m/
/ W i 2 Z \ mi; Tm / 2 Y
.h.fi ; h.N
i
and for each i < m, the sequence hJki W k < !i witnesses that Oi is .hi.Nm / /-small
where
Oi D [Jki j k < !:
By Lemma 7.37, there exist L0 2 ! and T 0 2 AmC1 such that
(3.1) T 0 SmC1 or T 0 \ SmC1 D ;,
(3.2) T 0 Tm ,
0
m / .L /
/ / W i 2 Z \ mi; T 0 / 2 Y ,
(3.3) .h.fi ; .h.N
i
465
By Lemma 7.37 again, there exist L00 2 ! and T 00 2 AmC1 such that
(4.1) T 00 T 0 ,
00
m / .L /
/
/ W i 2 Z \ m C 1i; T 00 / 2 Y ,
(4.2) .h.fi ; .h.N
i
(4.3) for all k < L00 , for all i 2 Z \ m, fi ./ Jki for all 2 T 00 .
Of course if T 0 \ SmC1 D ; then one can simply let T 00 D T 0 and L00 D L0 .
Set TmC1 D T 00 and NmC1 D Nm C L00 . Choose a sequence hJk W k < !i such
.N
/
that hJk W k < !i witnesses that Bm is hm mC1 -small. The sequence exists since Bm
is hm E -small. For each k < ! set Ikm D Jk .
Therefore by induction the sequences exist.
Let G be the lter generated by Ti j i < !. Thus G is M-generic. Let
j W .M; I / ! .M ; I /
be the associated iteration of length 1. It follows from (1.5), (1.7), and (1.8) that for all
t
u
i < !, if !1M 2 j.Si / then j.fi /.!1M / Bi .
There is an analogous version of the previous lemma for sequences of models. We
shall apply this lemma only to sequences which are iterable. However the lemma holds
for sequences which are not necessarily iterable and it is this more general version
which we shall prove, (for no particular reason).
Lemma 7.39. Suppose that h.Mk ; Ik / W k < !i is a sequence such that for each
k < !, Mk is a countable transitive model of ZFC, Ik 2 Mk and such that in Mk , Ik
M
is a uniform normal ideal on !1 k .
For each k < ! let
Yk D .YBC .Ik //Mk :
Suppose that for all k < !,
(i) Mk 2 MkC1 ,
(ii) jMk jMkC1 D .!1 /MkC1 ,
Mk
(iii) !1
MkC1
D !1
(iv) IkC1 \ Mk D Ik ,
(v) Yk D YkC1 \ Mk ,
M
466
7 Conditional variations
A .P .!1 / n Ik /Mk
and A is predense in
.P .!1 / n Ik ; /Mk :
By (vi) in the hypothesis of the lemma we can suppose that for each i < !,
Ai 2 Mi :
Following the proof of Lemma 7.38 it is straightforward, using Lemma 7.37
and (v), to build sequences hTi W i < !i, hIji W i; j < !i and hNi W i < !i such
that for all i < ! the following hold. Let
Z D i < ! j Si \ Ti ;:
(1.1) Ni D 0 and Ti D S for i < n.
(1.2) If i n then Ti 2 Ai and either Ti Si or Ti \ Si D ;.
(1.3) If i n then TiC1 Ti S , Ni 2 ! and Ni < NiC1 .
(1.4) Iji is an open interval with rational endpoints and
.Iji / < 1=.hi .Ni C j / C 1/:
(1.5) Bi [Iji j j < !.
.Ni /
(1.6) .h.fj ; hj
/ W j 2 Z \ i i; Ti / 2 Yi .
467
and
hNi W i mi
i
are given. For each i < m and k < ! let Jki D IkCN
.
m
Therefore
m/
.h.fi ; h.N
/ W i 2 Z \ mi; Tm / 2 Ym
i
and for each i < m, the sequence hJki W k < !i witnesses that Oi is .hi.Nm / /-small
where
Oi D [Jki j k < !:
By (v),
Ym D YmC1 \ Mm
and so
m/
.h.fi ; h.N
/ W i 2 Z \ mi; Sm / 2 YmC1 :
i
468
7 Conditional variations
m / .L /
(3.4) .h.fi ; .h.N
/ / W i 2 Z \ mi; T 0 / 2 Ym ,
i
m / .L /
/
/ W i 2 Z \ m C 1i; T 00 / 2 YmC1 ,
(4.2) .h.fi ; .h.N
i
(4.3) for all k < L00 , for all i 2 Z \ m, fi ./ Jki for all 2 T 00 .
Of course, as in the proof of Lemma 7.38, if T 0 \ SmC1 D ; then one can simply let
T 00 D T 0 and L00 D L0 .
Set TmC1 D T 00 and NmC1 D Nm C L00 . Choose a sequence hJk W k < !i such
.N
/
that hJk W k < !i witnesses that Bm is hm mC1 -small. The sequence exists since Bm
is hm E -small. For each k < ! set Ikm D Jk .
Therefore by induction the sequences exist.
Let G be the lter generated by Ti j i < !. Thus G is [Mi j i < !-generic.
Let
j W h.Mk ; Ik / W k < !i ! h.Mk ; Ik / W k < !i
be the associated iteration of length 1. It follows from (1.5), (1.7) and (1.8) that for all
t
u
i < !, if !1M0 2 j.Si / then j.fi /.!1M0 / Bi .
With these lemmas the main iterations lemmas are easily proved. As usual it is
really the proofs of these iteration lemmas which are the key to the analysis of Bmax .
Lemma 7.40 (CH). Suppose h.M; I /; a; Y i 2 Bmax and that J is a normal uniform
ideal on !1 . Then there is an iteration
j W .M; I / ! .M ; I /
such that
(1) J \ M D I ,
(2) j.Y / D YBC .J / \ M .
469
Proof. Let hSk; W k < !; < !1 i be a sequence of pairwise disjoint J -positive sets
such that
!1 D [Sk; j k < !; < !1 :
Let hs W < !1 i be an enumeration (with repetition) of all nite sequences of open
subsets of .0;1/ such that for each nite sequence s of open subsets of .0;1/, and for
each .k; / 2 ! !1 ,
2 Sk; j s D s
is a set which is J -positive.
Let hB W < !1 i be an enumeration of all the borel subsets of .0;1/.
Let x be a real which codes M and let
C !1
be a closed unbounded set of ordinals which are admissible relative to x.
Fix a function
F W ! !1M ! M
such that
(1.1) F is onto,
(1.2) for all k < !, F jk !1M 2 M,
(1.3) for all A 2 M if A has cardinality !1M in M then A ran.F jk !1M / for
some k < !.
The function F is simply used to anticipate elements in the nal model. Our situation is similar to that in the proof of Lemma 7.7.
Suppose
j W .M; I / ! .M ; I /
is an iteration. Then we dene
j.F / D [j.F jk !1M / j k < !
and it is easily veried that M is the range of j.F /. This follows from (1.3).
Implicit in what follows is that for 2 C if
j W .M; I / ! .M ; I /
is an iteration of length then j.!1M / D . This is a consequence of Lemma 4.6(1).
We construct an iteration
h.M ; I /; G ; j; W < !1 i
of M of length !1 using the function F to provide a book-keeping device for dealing
with elements of
j0;!1 ..P .!1 / n I /M /
and for dealing with elements of j0;!1 .Y /.
470
7 Conditional variations
(2.2) !1
2 j.S /.
471
!1
then
2 j.S 0 /
M
j.f 0 /.!1 / B
for all < such that B is h0 -small. Let j;C1 D j and let G be the associated
M -generic lter. If
j0; .F /.k; / 2 .P .!1 / n I /M ;
then choose j such that in addition to the requirement above,
M
!1
2 j.S /
t
u
The analysis of the Bmax -extension requires the generalization of Lemma 7.40 to
sequences of models. We state this lemma only for the sequences that arise, specically
those sequences of structures coming from descending sequences of conditions in Bmax .
Suppose that hpk W k < !i is a sequence of conditions in Bmax such that for all
k < !,
pkC1 < pk :
We let hpk W k < !i be the associated sequence of conditions which is dened as
follows. For each k < ! let
h.Mk ; Ik /; ak ; Yk i D pk
and let
472
7 Conditional variations
be the iteration obtained by combining the iterations given by the conditions pi for
i > k. Thus jk is uniquely specied by the requirement that
jk .ak / D [ai j i < !:
For each k < !,
pk D h.M ; Ik /; jk .ak /; jk .Yk /i:
We note that by Corollary 4.20, the sequence
h.Mk ; Ik / W k < !i
is iterable (in the sense of Denition 4.8).
Lemma 7.41 (CH). Suppose hpk W k < !i is a sequence of conditions in Bmax such
that for each k < !
pkC1 < pk :
Let hpk W k < !i be the associated sequence of Bmax conditions and for each k < !
let
h.Mk ; Ik /; ak ; Yk i D pk :
Suppose that J is a normal uniform ideal on !1 . Then there is an iteration
j W h.Mk ; Ik / W k < !i ! h.Mk ; Ik / W k < !i
such that for all k < !;
(1) J \ Mk D Ik ,
(2) YBC .J / \ Mk D j.Yk /.
Proof. By Corollary 4.20, the sequence
h.Mk ; Ik / W k < !i
is iterable.
The lemma follows by a routine modication of the proof of Lemma 7.40 using
Lemma 7.39 in place of Lemma 7.38.
t
u
Theorem 7.43 establishes the nontriviality of Bmax in the sense required for the
analysis of L.R/Bmax . The proof requires Theorem 7.18, Theorem 7.42 and the transfer
principle supplied by Theorem 5.36.
473
Theorem 7.42. Suppose that is a Woodin cardinal. Suppose A R and that every
set of reals which is projective in A is C -weakly homogeneously Suslin. Then there is
an iterable structure .M; I / such that
M ZFC C CH
and such that
(1) M
AC .I /,
(2) A \ M 2 M,
(3) hH.!1 /M ; A \ Mi hH.!1 /; Ai,
(4) .M; I / is A-iterable.
Proof. Suppose that
G Coll.!1 ; </
is V -generic. By Theorem 7.18, in V G there exists a normal uniform saturated ideal
IG on !1 such that
IG \ V D .INS /V
and such that
V G
AC .IG /:
V G
t
u
474
7 Conditional variations
AC .I /,
(2.4) A \ M 2 M and
hH.!1 /M ; A \ M; 2i hH.!1 /M ; A \ M; 2i;
(2.5) .M; I / is A \ M -iterable in M .
Thus
Y D .YBC /M
475
The analysis of Bmax is now a straightforward generalization of that of Pmax . Suppose G Bmax is L.R/-generic. Then for each h.M; I /; a; Y i 2 G there corresponds
a unique iteration
j W .M; I / ! .M ; I /
such that j .!1M / D !1 . This iteration is constructed by combining the countable
iterations of .M; I / given by conditions p 2 G such that
p < h.M; I /; a; Y i:
Let
(1) YG D [j .Y / j h.M; I /; a; Y i 2 G,
(2) IG D [j .I / j h.M; I /; a; Y i 2 G,
(3) AG D [j .a/ j h.M; I /; a; Y i 2 G,
476
7 Conditional variations
t
u
Lemma 4.52 and Theorem 4.53 also generalize to the Bmax -extension.
Theorem 7.46. Assume AD L.R/ . Suppose G Bmax is L.R/-generic. Then in
L.R/G the following hold.
(1) Suppose B R and B 2 L.R/. Then the set
X hH.!2 /; B; 2i j MX is B-iterable and X is countable
contains a club, where MX is the transitive collapse of X .
(2) Suppose S !1 is stationary and f W S ! Ord. Then there is a function
g 2 L.R/ such that
2 S j g./ D f ./
is stationary.
Proof. (1) follows by an argument essentially identical to the proof of Lemma 4.52.
The application of the argument requires
P .!1 /G D P .!1 /
which is true by Theorem 7.44.
(2) follows from (1) and Theorem 3.42 using a chain condition argument to reduce
to the case that
f WS !
for some < L.R/ , cf. the proof of Theorem 4.53.
t
u
To prove that the Borel Conjecture holds in L.R/Bmax we use the following lemmas.
Suppose U is a free ultralter on !. Recall that the ultralter U is selective if for
all partitions hk W k < !i of ! either k 2 U for some k < ! or there exists 2 U
such that
j \ k j 1
for all k < !.
Let PU denote the partial order dened as follows. This is Prikry forcing adapted
to U .
PU consists of pairs .s; / such that s is a nite subset of ! and 2 U . The order
is dened by,
.s1 ; 1 / .s0 ; 0 /
if s0 s1 , 1 0 and
s1 n s0 0 :
477
is a function where !<! denotes the set of nite subsets of !. There exists a set
2 U such that for all k 2 , n .k C 1/ .a/ for all a k C 1.
Given this the usual analysis of Prikry applies. We give the results of this in the
next lemma. The Prikry Property is due essentially to Prikry and the Geometric Condition is due to Mathias. Prikry was concerned with the standard formulation of Prikry
forcing which is dened from a normal measure, however the proofs in that case immediately generalize to this case. A good reference for generalizations of Prikry forcing to
more general ultralters is Blass .1988/ to which we also refer the reader for historical
remarks.
Lemma 7.47. Suppose U is a selective ultralter on !.
(1) (Prikry property) Suppose .s; / 2 PU and b 2 RO.PU /. Then there exists
2 U such that .s; / b or such that .s; / b 0 .
(2) (Geometric condition) Suppose a ! is an innite set such that a n is nite
for all 2 U . Let
Ga D .a \ k; / 2 PU j k < ! and a .a \ k/ [ :
Then Ga is a V -generic lter in PU .
t
u
478
7 Conditional variations
s .NkC1 \ / [ akC1 :
Let W D [J.
We claim that W is h -small and that
.;; / 0 WG :
We rst prove that
Suppose
.;; / 0 WG :
.s; / .;; /
k 2 j k > [s:
.;; / 0 WG :
479
480
7 Conditional variations
jJm j 2kC2
Nk m < NkC1 :
be the collection of intervals
For each m 2 !, let Jm
interval .a; b/ 2 Jm , Jm contains the intervals,
.a; .a C b/=2/;
Let J D
[Jm
.a C .b
a/=4; b
.b
a/=4/;
and
j m < !. Thus
W D [J :
Jm
t
u
481
for some n 2 !.
Proof. Suppose F 2 V G is a function,
F W !1 ! !1
such that F ./ < for all > 0.
Suppose A !1 , A 2 V G and A I.G/ . We may suppose 0 A.
We must show that there exists B such that B A, B I.G/ and such that F jB is
constant.
Let F 2 V PU be a term for the function F and let A 2 V P be a term for the set
A.
Fix a condition .s0 ; 0 / 2 G. We may suppose that
.s0 ; 0 / A I.G/ :
We work in V . Let A be the set of < !1 such that there exists a condition
.s; / < .s0 ; 0 / with the property that
.s; / 2 A :
Since
.s0 ; 0 / A I.G/ ;
it follows that A I .
For each 2 !1 choose a condition .s ; / < .s0 ; 0 / and an ordinal <
such that
.s ; / ./ D ;
and such that if 2 A then
.s ; / 2 A ;
and if A then
.s ; / A :
482
7 Conditional variations
.s; n n/ ./ D :
Then for each n 2 ! there exist an open set On and a set An 2 I.G/ such that On is
h.n/
G -small and such that
< !1 j f ./ On An :
I.G/ is the ideal generated by I and PU is ccc. Therefore there must exist A 2 I
such that for all n < ! and for all 2 !1 n A, f ./ 2 On .
Let X D f ./ j 2 A. Thus X 2 V , jX j D !1 and X is hG E -small in V G.
This is a contradiction.
Therefore for some n 2 !,
V G
:
.h.f; h.n/
G /i; !1 / 2 .ZBC .I.G/ //
t
u
483
for some n 2 !.
Proof. (1) is an immediate consequence of the fact that PU is ccc.
We prove (2). Suppose B .0;1/ is a borel set in V . Let BG be the interpretation
of B in V G.
It sufces to prove that for all
.h.fi ; hi / W i < ni; S / 2 .YBC .I //V ;
.h.fi ; hi / W i < ni; S / 2 .ZBC .I.G/ //V G :
Granting this (2) follows from the denition of
.YBC .I.G/ //V G
as a subset of .ZBC .I.G/ //V G .
The claim that
.YBC .I //V .ZBC .I.G/ //V G
follows from Lemma 7.48. To illustrate how we suppose
.h.f; h/i; S / 2 .YBC .I //V
and prove that
.h.f; h/i; S / 2 .ZBC .I.G/ //V G :
Fix p0 2 ! ! such that
.h.f; h /i; S / 2 .YBC .I //V
where:
(1.1) p0 .0/ D 0 and p0 .k/ < p0 .k C 1/ for all k < !;
(1.2) for some m 2 !, m > 1 and
p0 .k/ D k m
for all sufciently large k < !;
484
7 Conditional variations
Therefore
V G
.h.f; h.n/
G /i; !1 / 2 .YBC .I.G/ //
t
u
485
f0 W !1M0 ! .0;1/
AC .I1 /:
Let < !1 be such that < and such that N ZFC. Finally let
M1 D N G0 G1 ;
486
7 Conditional variations
let,
Y1 D .YBC .I1 //N G0 G1 ;
and let
a1 D j.a0 /:
We claim that h.M1 ; I1 /; a1 ; Y1 i 2 Bmax and is as desired. By Lemma 4.4 and
Lemma 4.5, .M1 ; I1 / is iterable.
By Lemma 7.50,
.YBC .INS //N D .YBC .INS //N G0 \ N;
and for some n < !,
N
N G0
.h.j.f0 /; h.n/
:
G0 /i; !1 / 2 .YBC .INS //
Since
.Coll.!1 ; <//N G0
is !-closed in N G0 , and since
I1 \ N G0 D .INS /N G0 ;
it follows that
.YBC .INS //N G0 D .YBC .I1 //N G0 G1 \ N G0 :
Thus
and for some h 2 M1 ,
j.Y / D Y1 \ M0
.h.j.f0 /; h/i; !1M1 / 2 Y1 :
t
u
We shall use the following lemma to show that in L.R/Bmax , every set X R of
cardinality !1 has Lebesgue measure 0. We thank A. Miller for revealing the lemma
to us.
Lemma 7.52. Suppose that U is a selective ultralter on ! and that
G PU
is V -generic. Then in V G,
V \R
has Lebesgue measure 0.
Proof. We work in V G and prove that
V \ 2!
has Lebesgue measure 0. For each set a ! let
fa W ! ! 0;1
487
t
u
Combining Theorem 7.44, Lemma 7.50, and Lemma 7.52 we obtain the following
additional corollary. The proof is quite similar to that of Theorem 7.51.
Theorem 7.53. Assume ADL.R/ . Then
L.R/Bmax Every set of reals of cardinality @1 has measure 0:
t
u
t
u
488
7 Conditional variations
Theorem 7.56. Assume BC holds and that there are ! many Woodin cardinals with
a measurable above. Suppose is a 2 sentence in the language for the structure
hH.!2 /; 2; INS ; YBC .INS /; X I X 2 L.R/; X Ri:
Suppose that
hH.!2 /; 2; INS ; YBC .INS /; X I X 2 L.R/; X Ri :
Then
Bmax
:
t
u
(1) Let BCFA denote the forcing axiom: Suppose P is a poset such
P
t
u
t
u
489
Theorem 7.59 shows that BCFACC .!1 / holds in L.R/Bmax . An amusing question
is whether there are any partial orders in L.R/Bmax satisfying (in L.R/Bmax ) the preservation requirements of BCFA and of cardinality !1 . A natural conjecture is that there
are none. It seems likely that in L.R/Bmax , any nontrivial partial order of cardinality !1
adds a Cohen real. Nevertheless the proof of Theorem 7.59 does generalize to prove
a non-vacuous version of the theorem, cf. Theorem 9.42. The proof also quite easily
adapts to prove, from suitable assumptions, that
.c/
L.R/Bmax BCFACC
ZF
where BCFAZF .c/ is the version of BCFA.c/ analogous to Martins MaximumZF .c/, see
Denition 2.51.
Theorem 7.59. Assume ADL.R/ . Suppose G Bmax is L.R/-generic.
Then
L.R/G ZFC C BC C BCFACC .!1 /:
Proof. Suppose that
P 2 L.R/G
is a partial order satisfying the requirements for BCFA.!1 /. We view P D .!1 ; P /.
It sufces to prove the following: Suppose that
!1 !1
denes a term for a subset of !1 which codes a subset of YBC .INS / as computed in
L.R/GP . For each < !1 let
A D < !1 j .; / 2 :
Suppose that
!1 !1
is such that for each < !1 , D is dense in P where
D D < !1 j .; / 2 :
Then there is a lter F in P such that for all < !1 ,
F \ D ;
and such that X codes a subset of YBC .INS / where
X D 2 !1 j A \ F ;:
By Theorem 7.44, there exists a condition
h.M0 ; I0 /; a0 ; Y0 i 2 G
such that
where
P ; ; 2 M0
j W .M0 ; I0 / ! .M0 ; I0 /
490
7 Conditional variations
M1j1 .P0 / ;
< !1M1 j .; / 2 j1 .0 /
is dense in j1 .P0 /;
where j1 is the (unique) iteration
j1 W .M0 ; I0 / ! .M0 ; I0 /
such that j1 .a0 / D a1 .
Fix h.M0 ; I0 /; a0 ; Y0 i, P0 and .0 ; 0 /. We work in L.R/.
As in the proof of Theorem 7.51, let z 2 R code M0 and let N be a transitive inner
model of ZFC C CH such that Ord N , z 2 N and such that for some < !1 ,
N is a Woodin cardinal:
We also require that !1 is strongly inaccessible in N . Since AD holds in L.R/, N
exists by Theorem 5.34.
Let < !1 be such that < and such that
N ZFC:
By Lemma 7.40, there exists an iteration
j0 W .M0 ; I0 / ! .M0 ; I0 /
such that j0 2 N and such that
(2.1) .INS /N \ M0 D I0 ,
(2.2) j0 .Y / D .YBC .INS //N \ M0 .
Thus j0 .P0 / is a partial order on !1N .
We claim the following hold in N .
491
AC .I1 /:
492
7 Conditional variations
Thus
h.M1 ; I1 /; a1 ; Y1 i 2 Bmax :
Clearly
.YBC .INS //N h1 D N h1 \ .YBC .INS //N h1 g1 ;
and so by (3.2) and the choice of j0 ,
j0 .Y0 / D M0 \ .YBC .INS //N h1 g1 :
Thus Yb Y1 and
h.M1 ; I1 /; a1 ; Y1 i < h.M0 ; I0 /; a0 ; Y0 i:
By genericity we can suppose that
h.M1 ; I1 /; a1 ; Y1 i 2 G:
Let
t
u
Chapter 8
| principles for !1
For our next example of a Pmax variation, we consider versions of the principle |. Our
purpose, in part, is to illustrate degrees of freedom in the analysis of a Pmax variation
which we have not yet had to exploit.
We x some notation. Suppose are ordinals. Then
denotes the set of all subsets of of ordertype and
<
denotes the set of all subsets of of ordertype less than .
Denition 8.1 (Ostaszewski). |: There is a sequence
h W < !1 i
satisfying the following conditions.
(1) For each limit ordinal 0 < < !1 , 2 ! and is conal in .
(2) For each conal set A !1 , the set < !1 j A is stationary.
t
u
The principle | was introduced by Ostaszewski .1975/ as a weakening of . Assuming CH it is equivalent to and it easily veried that | implies that the nonstationary ideal is not saturated. One natural question, which plausibly can be answered
by the techniques of this section, is the following:
Assume |. Can INS be semi-saturated?
The underlying question is whether | can be obtained in a variation of a Pmax extension.
We dene two variations of |.
Denition 8.2. |0NS : There is a sequence
h W < !1 i
satisfying the following conditions.
(1) For each limit ordinal 0 < < !1 , 2 ! and is conal in .
(2) For each closed, conal set C !1 , the set
< !1 j n C is nite
contains a closed, conal subset of !1 .
t
u
494
8 | principles for !1
The principle |0NS weakens | in that only closed, conal, subsets of !1 are guessed,
and the anticipation is not as strong, being modulo a nite set. However this must
happen on a closed unbounded subset of !1 rather than just on a stationary subset of
!1 . This requires weakening how sets are guessed.
The proof of the forthcoming Lemma 8.25 can easily be modied to prove the
following lemma.
Lemma 8.3. Suppose that h W < !1 i is a sequence witnessing |0NS . Then there
exists a co-stationary set S !1 such that
< !1 j n S is nite
t
u
is stationary in !1 .
t
u
8 | principles for !1
495
It is easily checked that |NS holds in L, though unlike |, |NS is not implied
by . |NS seems more closely related to C though we do not know if it is implied
by C . Building nontrivial models in which |0NS holds seems difcult using the standard methods of iterated forcing, see .Shelah 1998/ for related results and additional
references. In particular a natural question is whether either |0NS or |NS implies that
INS is not saturated.
|NS
and we shall prove that if
We shall dene a partial Pmax
L.R/ AD
|NS
Pmax
|NS
Pmax
then
2 L.R/,
is L.R/-generic then
NS
NS
is !-closed and Pmax
is homogeneous. Further if G Pmax
NS
ness theorem for the Pmax
-extension. The absoluteness theorem we prove is somewhat
technical and very likely more elegant versions are possible.
|NS
-extension will require, as usual, proving several iteration
The analysis of the Pmax
lemmas. We shall prove these by working in L-like models, i. e. models in which very
strong condensation principles hold. This degree of freedom has always been available
but until now it has not been particularly useful. One purpose of this chapter is simply
to illustrate this approach.
There is another potential feature of the Pmax variations which is illustrated by the
|NS
|NS
which we give. It is only after the initial analysis of Pmax
that we
analysis of Pmax
|NS
NS
are able to prove that INS is !2 -saturated in L.R/Pmax . More precisely if G Pmax
is L.R/-generic then as a result of the initial analysis we obtain (assuming ADL.R/ );
L.R/G ZFC,
P .!1 /L.R/G D P .!1 /G ,
where P .!1 /G is dened from the lter G in the usual fashion.
|NS
|NS
, of Pmax
, obtaining a new family
We then extend this analysis to a variant, Umax
of iterable structures, which are generated from countable elementary substructures of
|NS
is L.R/-generic and
L .R/G where G Umax
L .R/G ZFC C ZC C 1 -Replacement:
|
NS
By considering iterations of these structures we are then able to prove that if G Pmax
L.R/
is L.R/-generic then (assuming AD
)
496
8 | principles for !1
8.1
Condensation Principles
We briey discuss the condensation principle we shall use to prove the iteration lemmas
|NS
required for the analysis of the Pmax
-extension.
We begin with the denition of a generalized condensation axiom.
Denition 8.5. Suppose that A and that
F W <! !
where is an uncountable ordinal. The function F witnesses condensation for the set
A if for all X such that X is set generic over V , if F X <! X then
hX ; AX ; 2i 2 V
where hX ; AX ; 2i is the transitive collapse of the structure hX; A; 2i.
t
u
We say that condensation holds for a set A Ord if there is a function which
witnesses condensation for A. The Axiom of Condensation asserts that for every set of
ordinals there exists a function witnessing condensation for the set.
We give in the next three theorems some of the elementary consequences of the Axiom of Condensation. The rst shows that in testing whether a function F is a witness
for condensation one need only consider elementary substructures which lie in a simple Cohen extension of V . The second of these theorems gives the key absoluteness
results relating to condensation and the third shows that the Axiom of Condensation
implies GCH.
Theorem 8.6. Suppose that A and that
F W <! !
where is an uncountable ordinal. Suppose that M is a transitive model of ZC such
that
(1) A; F M and R n M ;,
(2) for all X , if F X <! X then
hX ; AX ; 2i 2 M
where hX ; AX ; 2i is the transitive collapse of the structure hX; A; 2i.
Then F witnesses condensation for A.
t
u
t
u
497
t
u
Then 2
D jj .
t
u
Corollary 8.11. Assume the Axiom of Condensation holds in V . Then GCH holds. u
t
D. Law has improved Corollary 8.11, proving that follows from the Axiom of
Condensation, Law .1994/. The proof yields a different proof of Corollary 8.11; the
original proof used Namba forcing.
Theorem 8.12 (Law). Assume the Axiom of Condensation holds in V . Then holds.
Proof. Suppose that
j WM !N
is an elementary embedding such that
(1.1) M and N are transitive,
(1.2) M ZC C 1 -Replacement,
(1.3) N D j.f /.!1M / j f 2 M ,
(1.4) .H.!2 //M 2 N .
498
8 | principles for !1
Then
M :
To see this x
f0 W !1M ! M
such that f0 2 M and such that
h.H.!2 //M ; <0 i D j.f0 /.!1M /
where <0 is a wellordering of .H.!2 //M such that <0 2 N .
Working in M , one can dene, in the usual fashion using f0 , a sequence.
Now x 2 Ord such that
V ZC C 1 -Replacement;
and such that cof./ > !1 . Let
X V
be a countable elementary substructure. Fix < and A such that
(2.1) 2 X ,
(2.2) A 2 X ,
(2.3) H.!2 / 2 L A.
By elementarity there exists a function
W <! !
such that 2 X and such that witnesses condensation for A.
Let
Y D f .X \ !1 / j f 2 X :
Hence
(3.1) X Y V ,
(3.2) Y \ is closed under .
Finally let M be the transitive collapse of X , let N be the transitive collapse of Y and
let
j WM !N
be the elementary embedding, given by the image of the inclusion map from X to Y .
Thus
(4.1) M and N are transitive,
(4.2) M ZC C 1 -Replacement,
(4.3) N D j.f /.!1M / j f 2 M .
499
Let .AX ; X ; X / be the image of .A; ; / under the transitive collapse of X . The key
point is that j.X / witnesses condensation in N for j.AX / and so by absoluteness,
AX 2 N;
since j./ j < X is closed under j.X /. But this implies that
.H.!2 //M 2 N
since .H.!2 //M 2 LX AX .
Thus .N; M; j / satises (1.1)(1.4) and so
M
which implies that holds in V .
t
u
Remark 8.13. The proof of Theorem 8.12 easily adapts to prove directly that the Axiom of Condensation implies that for any (uncountable) regular cardinal , holds at
on any stationary subset of .
It is open whether the Axiom of Condensation implies C or whether it implies
t
u
principles such as !1 .
Natural models in which the Axiom of Condensation holds are provided by AD.
Theorem 8.14. Assume AD holds in L.R/ and let
M D H.!1 / \ .HODx/L.R/
where x 2 R. Then
M ZFC C Axiom of Condensation:
t
u
t
u
500
8 | principles for !1
Remark 8.16. (1) The denition of strong condensation imposes some unnecessary requirements on M . A slightly more general denition could be given by
specifying as a witness, a wellordering of M .
(2) We shall essentially only be concerned with strong condensation for transitive
sets of the form H./ where is an uncountable cardinal (actually !3 in most
cases).
t
u
We note that in the denition of a witness for strong condensation it is necessary
only to consider elementary substructures which lie in V as opposed to the case of
witnesses for condensation where it is necessary to consider elementary substructures
which are generic over V . This is veried in the following theorem.
Theorem 8.17. Suppose that M is a transitive set closed under the Godel operations
and that
F W Ord \ M ! M
is a bijection. Suppose that N is a transitive inner model such that
(1) N ZC C 1 -Replacement,
(2) M; F N ,
(3) F witnesses strong condensation for M in N .
Then F witnesses strong condensation for M .
t
u
NS
8.2 Pmax
501
Remark 8.20. (1) Theorem 8.19 generalizes to other inner models of AD, satisfying V D L.P .R//, provided that a particular form of AD is assumed, see
Theorem 9.9.
(2) Suppose that the Axiom of Condensation holds. Does strong condensation hold
for H.!2 /?
(3) Suppose that A Ord and that for each uncountable cardinal of LA, strong
condensation holds in LA for H. /LA . Suppose that A# exists. Then there
exists < !1 and a set A such that
LA D LA :
(4) Does condensation or strong condensation capture the combinatorial essence of
inner models like L? One test question is the following.
Suppose that N is a transitive inner model of ZFC containing the ordinals
such that for each uncountable cardinal of N , strong condensation holds
in N for H. /N . Suppose that covering fails for N in V . Must there exist
a real x such that
N Lx
Note that if N D LA for some A Ord, then by (3) and Jensens Covering
Lemma, the answer is yes.
t
u
NS
8.2 Pmax
|
NS
We dene Pmax
as a variation of Pmax
. For this denition and the subsequent analysis we shall use a generalization of the partial orders PU to the case where U is an
ultralter on !1 , cf. the discussion preceding Lemma 7.47.
t
u
502
8 | principles for !1
NS
8.2 Pmax
503
t
u
504
8 | principles for !1
t
u
These lemmas however do not rule out the following. There is a set Y of triples
.U; I; B/ such that
(1) U is a uniform ultralter on !1 which extends the club lter,
(2) I is a normal uniform saturated ideal on !1 ,
(3) B is an !2 -complete boolean subalgebra of P .!1 /=I ,
(4) RO.PU / B,
NS
8.2 Pmax
505
NS
This is how we shall obtain |NS in the Pmax
-extension except the ultralters U
will be generic over the model, see Theorem 8.84. In fact there will exist an .!1 ; 1/distributive partial order PF (dened from F ) for adding U such that
t
u
506
8 | principles for !1
Suppose that F and U are as in Denition 8.26. In general IU;F is not a proper
ideal. Suppose that IU;F is a proper ideal and that
F P .!1 / n IU;F
is a V -normal ultralter (occurring in a set generic extension of V ).
Let .M; E/ D Ult.V; F / and let
j W V ! .M; E/
be the corresponding elementary embedding.
Since IU;F is a normal ideal, !2V OrdM ; i. e. !2V is contained in the wellfounded
part of M . Thus
j.F /.!1V / 2 !1V ! :
The key point is that by Lemma 8.23, it follows that j.F /.!1V / is V -generic for PU .
Let GF denote the V -generic lter
GF PU
determined by
j.F /.!1V /.
Thus
GF D p 2 PU j Zp;F 2 F :
t
u
NS
8.2 Pmax
507
FO P .!1 / n IU;F
t
u
Zq;F \ S IU;F :
Proof. The lemma easily follows from the denitions and Lemma 8.23 which gives the
geometric condition which characterizes when a conal ! sequence in !1V is V -generic
for PU .
If .S; p/ 2 RU;F then it is immediate that for all q p,
Zq;F \ S IU;F :
Now suppose that .S; p/ RU;F . Then by the denability of forcing, there must exist
q0 p such that if
G PU
is a V -generic lter, with q0 2 G, then G GF for any V -normal ultralter, F , such
that
(1.1) F .P .!1 / n IU;F /V ,
(1.2) S 2 F ,
(1.3) F is set generic over V .
It follows that in V , Zq0 ;F \ S 2 IU;F .
|
t
u
NS
We dene Pmax
. The denition is closely related to that of Pmax
which is given as
Denition 5.41.
508
8 | principles for !1
|
NS
Denition 8.30. Pmax
is the set of pairs
MkC1
(2) Mk 2 MkC1 ; !1
D !1
(3) [Mk j k 2 !
AC .
F W !1M0 ! !1M0 ! :
Mk
in Mk ,
M
.!1 k ; p/
2 .RU;F /Mk
Fk D \U j U 2 Yk :
Then
\.IU;F /Mk j U 2 Yk Ik :
NS
8.2 Pmax
509
NS
is dened as follows. A condition
The ordering on Pmax
O
O
.h.Mk ; Yk / W k < !i; FO / < .h.Mk ; Yk / W k < !i; F /
if hMk W k < !i 2 MO 0 , hMk W k < !i is hereditarily countable in MO 0 and there exists
an iteration j W hMk W k < !i ! hMk W k < !i such that:
(1) j.F / D FO ;
(2) hMk W k < !i 2 M0 and j 2 M0 ;
(3) For all k < !,
and
M
t
u
NS
.h.Mk ; Yk / W k < !i; F / 2 Pmax
:
NS
.h.Mk ; Yk / W k < !i; F / 2 Pmax
and if
NS
is suitably nontrivial assuming ADL.R/ by proving an
We shall prove that Pmax
iteration lemma for structures of the form .M; I/ where
I D .Q< /M
for some 2 M which is a Woodin cardinal in M.
For this we x some additional notation.
510
8 | principles for !1
t
u
NS
8.2 Pmax
511
NS
, are of
The structures we shall iterate in order to establish the nontriviality of Pmax
the form .M; I; a/ where for some 2 M, is a Woodin cardinal in M, where I is
the directed system, .I< /M , of ideals associated to the stationary tower .Q< /M , and
where
a 2 .Q< /M :
The iterations will be restricted so that the generic lters contain the images of a.
Denition 8.33. Suppose M is a countable transitive model of ZFC, 2 M and that
is a Woodin cardinal in M. Suppose that .M; I/ is iterable where I is the directed
system of nonstationary ideals,
.I< /M
and suppose that a 2 .Q< /M .
A sequence
h.M ; I ; a /; G ; j; W < i
is an iteration of .M; I; a/ if
(1) h.M ; I /; G ; j; W < i is an iteration of .M; I/,
(2) a0 D a,
(3) for all , j0; .a/ D a ,
(4) for all < , a 2 G .
t
u
We dene the collection of structures which are the subject of the rst iteration
lemma.
Denition 8.34. M|NS is the set of triples
h.M; I; a/; Y; F i
such that the following hold.
(1) M is a countable transitive model of ZFC.
(2) Strong condensation holds in M for M where is the least inaccessible cardinal
of M.
(3) I 2 M and I D I< as computed in M for some 2 M such that is a Woodin
cardinal in M.
(4) .M; I/ is iterable.
(5) F 2 M and
F W !1M ! !1M ! :
512
8 | principles for !1
\.IU;F /M j U 2 Y I:
I D .IU;F /M j U 2 Y :
t
u
NS
8.2 Pmax
513
Lemma 8.36. Suppose that strong condensation holds for H.!3 /. Then there is a
function
F W !1 ! !1 !
such that for every uniform ultralter, U , on !1 , the normal ideal IU;F is proper and
.!1 ; 1PU / 2 RU;F :
Proof. Fix a function
h W !3 ! H.!3 /
which witnesses strong condensation for H.!3 /.
For each < !3 let
M D h./ j <
and let
h D hj:
Let S be the set of < !3 such that
(1.1) M is transitive,
(1.2) h 2 M for all < ,
(1.3) hM ; h ; 2i ZFC n Powerset,
(1.4) .!2 /M exists and .!2 /M 2 M .
The key point is that for many constructions one can use the sequence
h.M ; h / W 2 S \ !1 i
exactly as one uses the sequence h.L ; <L / W < !1L i for an analogous construction
within L.
For each elementary substructure
X hH.!3 /; h; 2i;
let MX be the transitive collapse of X and let
X D MX \ Ord:
Since the function h witnesses strong condensation for H.!3 /, hjX is the image of
h under the collapsing map. Thus the transitive set MX is uniquely determined by the
ordinal X . If X is countable then X 2 S \ !1 and
MX D MX :
We construct the function
F W !1 ! !1 ! ;
dening F j by induction on . The construction is uniform and so since for each
2 S \ !1 ,
.S \ .!1 /M ; hj.!1 /M / 2 M ;
514
8 | principles for !1
NS
8.2 Pmax
515
where p D h.1 /. Note that by Lemma 8.29, 1 is dened if (2.2) holds. 1 is trivially
dened if (2.1) holds.
Dene F ./ D h.0 /.
This completes the denition of the function F . We verify that F has the desired
properties. If this fails then there exist
!1 < 0 < 0 < !2
such that
(4.1) 0 2 S ,
(4.2) in M 0 , h.0 / is an ultralter on !1 such that either
a) .IU;F /M0 is not a proper ideal in M 0 , or
b) .!1 ; 1PU /M0 .RU;F /M0 ,
where U D h.0 /.
We suppose that .0 ; 0 / is as small as possible (with !1 < 0 ) and we set
U D h.0 /. In either case, (4.1(a)) or (4.2(b)), there must exist
p 2 .PU /M0
such that
.Zp;F /M0 2 .IU;F /M0 I
if (4.1(a)) holds this is trivial and if (4.2(b)) holds this follows by Lemma 8.29.
Let 1 be least such that
.Zp;F /M0 2 .IU;F /M0 ;
where we set p D h.1 /.
We rst prove that .IU;F /M0 is a proper ideal in M 0 . The function h witnesses
strong condensation for H.!3 / and so it follows from the denition of F that for each
function
e W !1 <! ! U
such that e 2 M 0 ,
!1 n Ze;F 2 INS :
Therefore
.IU;F /M0 INS
and so .IU;F /M0 is a proper ideal in M 0 .
A similar argument shows that
!1 n .Zp;F /M0 2 INS
which contradicts
.Zp;F /M0 2 .IU;F /M0
since
.IU;F /M0 INS :
t
u
516
8 | principles for !1
t
u
(1) b 2 G,
(2) D j.F0 /.!1M0 /,
where j W .M0 ; I0 / ! .M0 ; I0 / is the iteration of length 1 given by G.
Suppose
a P!1 .[a/
t
u
NS
8.2 Pmax
517
and
b P!1 .[b/:
Let
X D .[a/ [ .[b/:
Then a and b are equivalent if there exists a set C P!1 .X /, closed and unbounded
in P!1 .X /, such that for each Z 2 X ,
Z \ .[a/ 2 a
if and only if
Z \ .[b/ 2 b:
Thus if a and b are stationary then a and b are equivalent if they dene the same
elements of RO.Q< / where is any ordinal such that
a; b V :
Remark 8.39. Suppose that
a P!1 .[a/;
and that [a has cardinality !1 .
(1) Suppose that a is stationary. Then there is a stationary set S !1 such that S
and a are equivalent. Further if T !1 is a stationary set which is equivalent to
a then
S M T 2 INS :
(2) Suppose that a is nonstationary. Then a is equivalent to each set T !1 such
t
u
that T 2 INS .
Lemma 8.40. Suppose that
h.M0 ; I0 ; a0 /; Y0 ; F0 i 2 M|NS
and that strong condensation holds for H.!3 /. Let 0 2 M0 be the Woodin cardinal
in M0 associated to I0 and let
Q0 D .Q<0 /M0
be the associated stationary tower.
Let
J0 D \.IU;F /M0 j U 2 Y0
and suppose that h.S ; T / W < !1M0 i 2 M0 is such that
S ; T j < !1M0 P .!1 /M0 n J0 :
Then there is an iteration
j W .M0 ; I0 ; a0 / ! .M0 ; I0 ; a0 /
of length !1 such that the following hold where F D j.F0 /.
8 | principles for !1
518
.j.0 // M0
/ ;
.p; b/ 2 .RW;F
where W D
equivalent to b.
(2) Suppose that h.S ; T / W < !1 i D j.h.S ; T / W < !1M0 i/: Let
h W < !1 i
be the increasing enumeration of the ordinals 2 !1 n .M0 \ Ord/ such that
is a cardinal in L.M0 /. Let
C D < !1 j D :
Then for all 2 C and for all < ,
2 S
if and only if
C 2 T :
(1.4) !2
M
exists and !2
2 M ,
NS
8.2 Pmax
519
2 M :
Suppose that
hG W < 0 i
is given. Let
h.M ; I ; a /; G ; j; W < 0 i
be the corresponding iteration.
We rst suppose that for all 2 S if
0 D .!1 /M ;
then the iteration
h.M ; I ; a /; G ; j; W < 0 i
satises the requirements of the lemma in M .
520
8 | principles for !1
Then
G0 D h.0 /
where 0 is least such that
h.0 / Q0 ;
a0 2 h.0 /, h.0 / is M0 -generic, and such the corresponding iteration
h.M ; I ; a /; G ; j; W < 0 C 1i
satises (2.1).
This denes G0 in this case (which we note includes the case that
0 .!1 /M
for all 2 S).
For the remaining cases let 0 2 S be least such that
0 D .!1 /M0
and such that the iteration
h.M ; I ; a /; G ; j; W < 0 i
fails to satisfy the requirements of the lemma in M 0 .
We shall extend the iteration dening G0 , attempting to eliminate the least counterexample. There are several cases depending on how the iteration
h.M ; I ; a /; G ; j; W < 0 i
fails to satisfy the requirements of the lemma within M 0 . Let 0 D j0;0 .0 /.
Because the iteration satises (2.1) necessarily requirement (2) of the lemma is
satised in M 0 . Therefore (1) must fail.
Let 0 be least such that:
(3.1) h.0 / 2 M 0 ;
(3.2) M 0 h.0 / is a uniform ultralter on !1 ;
(3.3) h.0 / \ M0 2 j0;0 .Y0 /;
(3.4) Let U D h.0 /. Either
a) .IU;F /M0 is not a proper ideal, or
b) there exists
. /
NS
8.2 Pmax
521
2 M
We prove that this iteration satises the conditions of the lemma. Clearly this
iteration satises (2) in the statement of lemma. We prove that (1) is also satised.
If not let 0 2 S be least such that
!1 D .!1 /
M
0
522
8 | principles for !1
and such that the iteration fails to satisfy the conditions of the lemma interpreted in
M 0 . There are several cases to consider depending on how the iteration fails to satisfy
the requirements of the lemma in M 0 .
Let 0 be least such that (3.1)(3.4) hold; i. e. 0 is least such that h.0 / 2 M 0
and witnesses that the iteration fails to satisfy the (1) of the lemma. Let U D h.0 /.
Suppose
X hH.!3 /; h; 2i
is a countable elementary substructure containing M0 . The iteration is denable in the
structure
hH.!1 /; hj!1 ; 2i
from M0 and so 0 ; 0 2 X . Let hMX ; hX ; 2i be the transitive collapse of X . Thus
hX D hj!1MX D hj.X \ !1 /
and MX \Ord 2 S. Let 0X D !1MX and let 0X be the image of 0 under the collapsing
map. Let UX be the image of U under the collapsing map, thus
hX .0X / D h.0X / D UX :
X
Let X
0 D MX \ Ord. Thus 0 2 S and
MX D M X :
0
X
0
2 M
MX
M
0
INS :
NS
8.2 Pmax
523
where
(7.1) q 2 .PU /M0 ,
(7.2) q p,
(7.3) .Zq;F /M0 \ Sb 2 .IU;F /M0 .
We obtain a contradiction by reection. Again let
X hH.!3 /; h; 2i
be a countable elementary substructure and let hMX ; hX ; 2i be the transitive collapse
of X . Let 0X be the image of 0 under the collapse. Thus
h.0X / D hX .0X /:
Let
0X D .!1 /MX D X \ !1
and let .UX ; qX ; bX ; qX SX / be the image of .U; p; b; q; Sb / under the collapsing map.
Arguing as above, GX is chosen using MX and so
0
(8.1) bX 2 GX ,
0
(8.2) F .0X / is MX -generic for .PUX /MX and qX belongs to the corresponding MX generic lter.
Further
jX ;!1 .bX / D b
0
since b 2 M!1 .
Therefore b is closed and unbounded in P!1 .[b/.
Similarly
M
!1 n .Zq;F / 0 2 INS
and
.Zq;F /
M
0
\ Sb 2 .IU;F /
M
0
Finally as above,
.IU;F /
M
INS :
M
0
M
0
\ Sb
t
u
524
8 | principles for !1
.j.0 // M0
.p; b/ 2 .RW;F
/ ;
NS
8.2 Pmax
525
(6) Suppose that h.S ; T / W < !1 i D j.h.S ; T / W < !1M0 i/: Let
h W < !1 i
be the increasing enumeration of the ordinals 2 !1 n .M0 \ Ord/ such that
is a cardinal in L.M0 /. Let
C D < !1 j D :
Then for all 2 C and for all < ,
2 S
if and only if
C 2 T :
C 2 T :
8 | principles for !1
526
NS
As a corollary to Lemma 8.41, if AD holds in L.R/ then Pmax
is suitably nontrivial. For this we require the following renement of Theorem 5.36.
NS
8.2 Pmax
527
(1.1) HOD N .
(1.2) There exist two Woodin cardinals in N below !1V .
(1.3) Let be the least inaccessible cardinal of N . Then
P . / \ N D P . / \ HOD:
We briey indicate how to obtain N . For each pair .x; y/ of reals with x 2 HODy
let Nx be the inner model,
0 x
HODLZ
Z0
and let Nx;y be the inner model
x y
HODN
:
Nx
HOD D LZ0 :
By the arguments given in the proof of Theorem 5.36, there exists x0 2 R such
that for all x 2 R if x0 2 HODx then there exists y0 2 R such that for all y 2 R if
y0 2 HODy then the inner model Nx;y satises (1.1)(1.3).
Let 0 be the least Woodin cardinal of N and let 1 be the next Woodin cardinal
of N .
The set A is 21 and so there exist trees
S .! 21 /<!
and
T .!
21 /<!
such that S; T HOD and such that
pS D R \ pT :
Thus by Theorem 2.32, S and T are <1 weakly homogeneous in N .
Again since A is 21 , by (1.1)
hV!C1 \ N; A \ N; 2i hV!C1 ; A; 2i:
This follows by standard arguments using the fact that every 21 set is the projection of
a denable tree; i. e. a tree in HOD.
Finally let be the least strongly inaccessible cardinal of N . By Theorem 8.19 and
(1.3), strong condensation holds for N in HOD. By Theorem 8.17, strong condensation holds for N in N . Let be the least strongly inaccessible cardinal of N above 1
and let
M D N :
Thus M witnesses that A is not a counterexample to the theorem, a contradiction.
t
u
In fact the next theorem shows that must less determinacy is required to obtain the
|NS
nontriviality of M|NS from which the nontriviality of Pmax
follows. The rst theorem
is in essence a lightface version of Theorem 8.19.
8 | principles for !1
528
1
is equivalent to
2 -Determinacy.
From Theorem 8.43 one obtains a little more than just that for every x 2 R there
exists
h.M; I; a/; Y; F i 2 M|NS
with x 2 M. One can require for example that modest large cardinals exist in M,
above the Woodin cardinal of M associated to I.
1
Theorem 8.45 (
2 -Determinacy). For each x 2 R there exists
.M; I; / 2 H.!1 /
such that
(1) x 2 M,
(2) M is transitive and M ZFC C is a Woodin cardinal,
(3) I D .I< /M ,
(4) .M; I/ is iterable,
(5) M L.M/,
(6) strong condensation holds in M for M where is the least inaccessible cardinal of M.
t
u
As a corollary to the previous lemmas we obtain the following lemma, which is a
variation of Lemma 5.23.
NS
8.2 Pmax
529
such that
(1) there exists a countable iteration
j W .M0 ; I0 ; a0 / ! .M0 ; I0 ; a0 /
such that j.F0 / D FO and such that
.M0 ; j.Y0 // D .MO 0 ; YO0 /;
(2) hMO k W k < !i is A-iterable for each set A 2 X0 such that every set of reals
which is projective in A is C -weakly homogeneously Suslin.
Proof. Let 2 X0 be the Woodin cardinal whose image under the transitive collapse
of X0 is the Woodin cardinal in M0 associated to I0 .
We dene by induction on k a sequence
hh.Mk ; Ik ; ak /; Yk ; Fk i W k < !i
|NS
of elements of M
be the associated stationary tower. Let hk W < !1 i be the increasing enumeration
of the ordinals 2 !1 n Mk such that is a cardinal in L.Mk /. Let
Ck D j k D
and let
Jk D \.IU;F /Mk j U 2 Yk :
530
8 | principles for !1
NS
8.2 Pmax
Let a D aI where
531
I D IU;F j U 2 Y :
|NS
.h.MO k ; YOk / W k < !i; FO / 2 Pmax
is guaranteed by (2.2(c)).
t
u
532
8 | principles for !1
As an immediate corollary of Lemma 8.46 and Theorem 8.42 we obtain the req|NS
. The statement of this
uisite theorem regarding the existence of conditions in Pmax
. The reason is that we have not
theorem is weaker than that of its counterpart for Pmax
|
NS
yet established the iteration lemmas for Pmax
and so we cannot conclude that the set
|NS
.
of conditions indicated in Theorem 8.47 is dense in Pmax
Theorem 8.47. Assume AD holds in L.R/. Then for each set A R with
A 2 L.R/;
there is a condition
|NS
.h.MO k ; YOk / W k < !i; FO / 2 Pmax
such that
(1) A \ MO 0 2 MO 0 ,
O
NS
8.2 Pmax
533
t
u
|
NS
The next iteration lemma we shall prove concerns conditions in Pmax
. This involves iterating sequences of models. We shall need the following lemma.
NS
:
.h.Mk ; Yk / W k < !i; F / 2 Pmax
Suppose that hUk W k < !i is a sequence such that for each k < !,
(1) Uk 2 Yk ,
(2) Uk UkC1 .
Suppose that
2 !1M0 !
and that for each k < !, is Mk -generic for .PUk /Mk . Then there exists an iteration
j W hMk W k < !i ! hMk W k < !i
of length 1 such that
D j.F /.!1M0 /:
Gk .PUk /Mk
534
8 | principles for !1
Suppose that
!1M0 .
where S1 D f 1 ./. Otherwise there must exist q0 2 Gk such that q0 < p0 and q0
M
forces that this fails; i. e. for all < !1 k ,
.f 1 ./; q1 / .RUk ;F /Mk
for any q1 q0 . However q0 < p0 and so
.S0 ; q0 / 2 .RUk ;F /Mk :
This is a contradiction; let U be a Mk -normal ultralter such that
U .P .!1 / n RUk ;Fk /Mk
and such that q0 2 gU where
gU .PUk /Mk
NS
8.2 Pmax
535
D j.F /.!1M0 /:
t
u
As an easy corollary to Lemma 8.48 and to the proof of Lemma 8.40 we obtain the
generalization of Lemma 8.40 to sequences of structures. We leave the details to the
reader.
Lemma 8.49. Suppose that strong condensation holds for H.!3 / and that
|
NS
:
.h.Mk ; Yk / W k < !i; F / 2 Pmax
t
u
|
NS
,
The next lemma when combined with Lemma 8.49 yields the !-closure of Pmax
|NS
with the appropriate assumptions on the nontriviality of Pmax .
NS
.h.Mk ; Yk / W k < !i; F / 2 Pmax
and that
536
8 | principles for !1
where W D U \ Mk .
Then there exists a set Y of uniform ultralters on !1 such that
(1) for any sequence hUk W k < !i such that for all k < !, Uk 2 j.Yk / and
Uk UkC1 , there exists U 2 Y such that
U \ Mk D Uk
for all k < !,
(2) let I be the ideal dual to the lter
F D \U j U 2 Y ;
then
\IU;j.F / j U 2 Y I;
(3) if U0 is an ultralter on !1 such that
\U j U 2 Y U0 ;
and such that for all k < !,
U0 \ Mk 2 j.Yk /;
then U0 2 Y .
Proof. Using the function j.F / the proof is essentially identical to that of Lemma 8.37.
Let Z be the set of uniform ultralters U on !1 such that for all k < !,
U \ Mk 2 j.Yk /:
We dene by induction on a normal ideal J as follows:
J0 D \IU;j.F / j U 2 Z
and for all > 0,
J D \IU;j.F / j U 2 Z and for all < , J \ U D ;:
It follows easily by induction that if 1 < 2 then
J1 J2 :
NS
8.2 Pmax
537
NS
We introduce the following notation for the constituents of a condition p 2 Pmax
:
p D .h.M.p;k/ ; Y.p;k/ / W k < !i; F.p/ /:
NS
NS
1
Corollary 8.51 (
2 -Determinacy). For each p0 2 Pmax there exists p1 2 Pmax such
that p1 < p0 and such that for each sequence
hWk W k < !i 2 M.p1 ;0/ ;
if
2 j.Y.p0 ;k/ /
Wk \ M.p
0 ;k/
for all k < !, then there exists U 2 Y.p1 ;0/ such that for all k < !,
I
Wk D U \ M.p
0 ;k/
where
W k < !i
j W hM.p0 ;k/ W k < !i ! hM.p
0 ;k/
is the .unique/ iteration such that j.F.p0 / / D F.p1 / .
Proof. Let x 2 R code p0 and let
.M; I; ; / 2 H.!1 /
538
8 | principles for !1
be such that
(1.1) x 2 M,
(1.2) M is transitive and M ZFC C is a Woodin cardinal,
(1.3) I D .I< /M ,
(1.4) .M; I/ is iterable,
(1.5) < and M M,
(1.6) strong condensation holds in M for M where is the least inaccessible cardinal of M.
The existence of .M; I; ; / follows from
12 -Determinacy, by Theorem 8.45.
|
NS M
Thus p0 2 .Pmax
/ and so by Lemma 8.49, there exists an iteration
j0 W hM.p0 ;k/ W k < !i ! hj0 .M.p0 ;k/ / W k < !i
with j0 2 M and such that the following hold in M.
j .M
/
IU;j0 .F.p0 / / \ j0 .M.p0 ;k/ / D IW;j0 .F.p0 / / 0 .p0 ;k/ ,
j .M
/
RU;j0 .F.p0 / / \ j0 .M.p0 ;k/ / D RW;j0 .F.p0 / / 0 .p0 ;k/ ,
NS
8.2 Pmax
539
NS M
NS
Thus, since p1 2 .Pmax
/ , p1 2 Pmax
. By the properties of j0 and by (5.1) and (5.2),
p1 < p0
and satises the requirements of the lemma.
t
u
540
8 | principles for !1
1
Corollary 8.52 (
2 -Determinacy). Suppose that hpk W k < !i is a sequence of ele|
NS
ments of Pmax
such that for all k < !,
pkC1 < pk :
Then there exists p 2
|NS
Pmax
AC ;
NS
is iterable then q 2 Pmax
.
k
O
The sequence hMk W k < !i satises the hypothesis of Lemma 4.17 and so it is
iterable, cf. the proof of Lemma 5.32.
|NS
such that p < q. It follows that
By Corollary 8.51 there exists p 2 Pmax
p < pk
for each k < !.
t
u
For each of the previously considered Pmax variations the proof that !1 -DC holds
in the extension has been a routine adaptation of the proof for the Pmax -extension using
|NS
this is Lemma 8.49
the appropriate analogs of Lemma 4.36 and Lemma 4.37; for Pmax
|NS
combined with Lemma 8.50. The situation for the Pmax -extension is different. Our
third iteration lemma establishes what is required to prove that !1 -DC holds in the
|NS
-extension.
Pmax
|NS
It is convenient to adapt Denition 4.44 to Pmax
.
NS
8.2 Pmax
541
NS
is semi-generic if for all < !1 there exists a
Denition 8.53. A lter G Pmax
condition
p2G
(2) IG D [M.p;0/
\ .INS /M.p;1/ j p 2 G,
t
u
NS
-extension) then the denition of YG is the natural choice.
(which will hold in the Pmax
|NS
We caution though that for an arbitrary semi-generic lter G Pmax
, the set YG is in
most cases empty. For example, we shall see that if ADL.R/ holds and
NS
G Pmax
is L.R/-generic then in L.R/G the set YG is empty.
The possibility that
YG D ;
NS
-extension is not simply a routine
is one reason that the proof of !1 -DC in the Pmax
application of our current iteration lemmas.
For the proof of Lemma 8.55 it is useful to make the following denition.
|NS
Pmax
; jk .Y.pk ;k/ // W i < !i; FO /
.h.MO k ; YOk / W k < !i; FO / D .h.M.p
k ;k/
is the condition associated to the sequence hpk W k < !i.
t
u
542
8 | principles for !1
Thus the condition associated to the sequence hpk W k < !i is precisely the condi|NS
is !-closed.
tion constructed in the proof of Corollary 8.52; i. e. that Pmax
1
Lemma 8.55 (
2 -Determinacy). Suppose that
hD W < !1 i
|
NS
NS
is a sequence of dense subsets of Pmax
and that q0 2 Pmax
. Suppose that strong
condensation holds for H.!3 /. Then there is a semi-generic lter
NS
G Pmax
G \ D ;:
(3) Suppose that .p; k/ 2 G !. For each W 2 jp;G .Y.p;k/ /, there exists U 2 YG
such that
D W:
U \ M.p;k/
(4) For each U 2 YG , the normal ideal IU;FG is proper.
(5) Suppose that .p; k/ 2 G !. For each U 2 YG ,
D .IW;FG /M.p;k/ ,
a) IU;FG \ M.p;k/
b) RU;FG \ M.p;k/
D .RW;FG /M.p;k/ ,
.
where W D U \ M.p;k/
|
NS
Proof. Since Pmax
is !-closed, we can easily build a decreasing sequence
hp W < !1 i
of conditions in
associated lter
|NS
Pmax
,
NS
G D p 2 Pmax
j p < p for some < !1
NS
is a semi-generic lter in Pmax
. The minor problem is that the set YG may be empty;
there may be no ultralters on !1 such that
U \ M.p;0/
2 jp;G .Y.p;0/ /
for all p 2 G.
NS
8.2 Pmax
543
The solution is to ensure that for each nonzero limit ordinal , the set
Y.p ;0/
is suitably large. Conditions (4) and (5) will be achieved by consideration of least
counterexamples as in the proofs of Lemma 8.40 and Lemma 8.49.
Fix a function
f W !3 ! H.!3 /
which witnesses strong condensation for H.!3 /.
Dene a function
h W !3 ! H.!3 /
as follows.
Let X be the set of t ! such that t codes a pair .; p/ where < !1 and
|NS
.
p 2 Pmax
For each < !3 let D !
. Thus
h W < !3 i
is the increasing enumeration of the limit ordinals (with 0) less than !3 .
Suppose < !3 then for each k < !,
h. C k C 1/ D f ./:
Suppose < !3 and f ./ X . Then
h. / D f ./:
Suppose < !3 and f ./ 2 X . Let .; p/ be the pair coded by f ./. Then
h. / D f . /
where is least such that f . / 2 D and such that f . / p.
Since f witnesses strong condensation for H.!3 / it follows that h also witnesses
strong condensation for H.!3 /. The verication is straightforward, note that f is
trivially denable from h in H.!3 /.
For each < !3 let
M D h./ j <
and let
h D hj:
Let S be the set of < !3 such that
(1.1) M is transitive,
(1.2) h 2 M for all < ,
(1.3) hM ; h ; 2i ZFC n Powerset,
M
(1.4) !2
M
exists and !2
2 M ,
544
8 | principles for !1
The reason for modifying f to obtain h is in order to achieve the following. Suppose 2 S. Then
M
NS M
NS M
(2.2) for each < !1 , D \ .Pmax
/ is dense in .Pmax
/ .
NS
of conditions in Pmax
below q0 such that for all < !1 , p 2 D . The lter generated
by the set p j < !1 will have the desired properties. By (2.1) and (2.2) it will
follow that for each 2 S,
(3.4) !1
M0
exists and !1
2 M 0 ,
M0
NS
NS
be the condition in Pmax
which is associated to the sequence
Let q 2 Pmax
hpk W k < !i.
NS
8.2 Pmax
545
NS
and such that the following hold where
Let 1 be least such that h.1 / 2 Pmax
p D h.1 /:
where
W k < !i
j W hM.q;k/ W k < !i ! hM.q;k/
NS M
/
j p < p for some < ;
g D p 2 .Pmax
let
Fg D [F.p/ j p 2 g
and for each p 2 g let
W k < !i
jp;g W hM.p;k/ W k < !i ! hM.p;k/
NS M
/ . This is because for each such ,
g is a semi-generic lter in .Pmax
D .!1 /M0
8 | principles for !1
546
and such that g fails to satisfy the requirements of the lemma as interpreted in M 0
relative to the sequence
hD \ M W < i:
If 0 does not exist then choose p as above. Similarly if
.Yg /M0 D ;
then again choose p as above. In fact it will follow by induction that .Yg /M0 ;
and further that g satises (3) in M 0 .
Therefore g fails to satisfy (4) or (5).
Otherwise let .0 ; 1 / be least such that
h.0 / 2 .Yg /M0 ;
h.1 / 2 g !, and such that either
(5.1) .IU;F /M0 is not a proper ideal, or
(5.2) .p; k/ 2 g ! and either
a) .IW;F /M.p;k/ .IU;F /M0 \ M.p;k/
, or
b) .RW;F /M.p;k/ .RU;F /M0 \ M.p;k/
,
where we set
F D .Fg /M0 ;
U D h.0 /;
W D h.0 / \ M.p;k/
;
and .p; k/ D h.1 /.
We shall again dene three ordinals 0 , 1 and 2 . These will depend on . Let 0
be least such that
h.0 / D hk W k < !i
where hk W k < !i is an increasing conal sequence in . Let 1 < !1 be least such
that
hpk W k < !i 2 H.!1 /M1 :
Let
q D .h.Mk ; Yk / W k < !i; F /
|
NS
be the condition in Pmax
associated to the sequence hpk W k < !i.
We shall dene an iteration of hMk W k < !i as follows. There are three cases.
Suppose rst that (5.1) holds. Let 2 be least such that h.2 / D j where
NS
8.2 Pmax
547
Next suppose that (5.1) fails. Then (5.2) holds. If (5.2(a)) holds then let 0 be least
such that
n .IW;F /M.p;k/ :
h.0 / 2 .IU;F /M0 \ M.p;k/
Let 2 be least such that h.2 / D j where
j W hMk W k < !i ! hMk W k < !i
is an iteration of length 1 such that j.F /./ is M 0 -generic for .PU /M0 and such that
!1M0 2 j.h.0 //:
If (5.2(a)) fails then (5.2(b)) holds.
Let 0 be least such that h.0 / D .S0 ; .s0 ; f0 /; .s1 ; f1 // where
where
j W hM.q ;k/ W k < !i ! hM.q
;k/ W k < !i
8 | principles for !1
548
NS
G D p 2 Pmax
j p < p for some < !1
NS
be the associated lter. Thus q0 2 G, G is a semi-generic lter in Pmax
, and for each
< !1 , G \ D ;.
We next prove that the set YG is nonempty. This is a consequence of the following
property of the sequence we have dened. For each < !1 let
W k < !i
j; W hM;k W k < !i ! hM;k
(7.4) !1
M
exists and !1
2 M ,
M
(7.6) < !1
NS
8.2 Pmax
549
Then there exists U 2 Y.p ;0/ such that for all < ,
U \ j; M.p ;0/ D j; .U /:
Using this property of the sequence it is straightforward to prove that for each
p 2 G, if
W 2 jp;G .Y.p;0/ /
then there exists U 2 YG such that
U \ jp;G .M.p;0/ / D W:
We now prove that (4) and (5). The argument is by reection and is quite similar
to that given in the proof of Lemma 8.40. We note that for all 2 S such that > !1 ,
G 2 M
and in M , G satises all of the requirements of the lemma except possibly (4) or (5).
If either (4) or (5) fail let 0 2 S be least such that
M0
!1
D !1
.IW;FG /M.p;k/ , or
(9.1) .IU;FG /M0 \ M.p;k/
.RW;FG /M.p;k/ ,
(9.2) .RU;FG /M0 \ M.p;k/
h.0 / 2 .IU;FG /M0 \ M.p;k/
n .IW;FG /M.p;k/ :
550
8 | principles for !1
NS
G Pmax
NS
8.2 Pmax
551
t
u
As a corollary to Lemma 8.55 and Lemma 8.56 we obtain the following lemma with
|NS
which the basic analysis of the Pmax
-extension is easily accomplished. Lemma 8.57 is
analogous to Lemma 4.46, though this formulation is more efcient.
8 | principles for !1
552
Lemma 8.57 (ADL.R/ ). Suppose that A R and that A 2 L.R/. Then for each
|NS
|NS
there exists p0 2 Pmax
such that p0 < q0 and such that:
q0 2 Pmax
(1) hM.p0 ;k/ W k < !i is A-iterable;
(2) hV!C1 \ M.p0 ;0/ ; A \ M.p0 ;0/ ; 2i hV!C1 ; A; 2i;
|
NS
(3) Suppose that D Pmax
is a dense set which is denable in the structure
hH.!1 /; A; 2i
M.p0 ;0/
from parameters in H.!1 /
. Then
D \ p > p0 j p 2 M.p0 ;0/ ;:
Proof. Fix A and let B be the set of x 2 R such that x codes an element of the rst
order diagram of the structure
hV!C1 ; A; 2i:
Let B be the set of x 2 R such that x codes an element of the rst order diagram of
the structure
hV!C1 ; B; q0 ; 2i:
Thus B 2 L.R/.
By Theorem 8.42 applied to B , there exist a countable transitive model M and an
ordinal 2 M such that the following hold.
(1.1) M ZFC.
(1.2) is a Woodin cardinal in M .
(1.3) B \ M 2 M and hV!C1 \ M; B \ M; 2i hV!C1 ; B ; 2i.
(1.4) B \ M is C -weakly homogeneously Suslin in M .
(1.5) Suppose is the least inaccessible cardinal of M . Then strong condensation
holds for M in M .
|
NS M
/ .
(1.6) q0 2 .Pmax
Let
hD W < !1M i
|
NS M
enumerate all the dense subsets of .Pmax
/ which are rst order denable in the structure
hH.!1 /M ; A \ M; 2i:
NS M
g .Pmax
/
such that the following hold in M where for each p 2 g,
jp;g W hM.p;k/ W k < !i ! hM.p;k/
W k < !i
is the iteration given by p.
NS
8.2 Pmax
553
(2.1) q0 2 g.
(2.2) For each < !1M ,
g \ D ;:
(2.3) Suppose that p 2 g. For each W 2 jp;g .Y.p;k/ /, there exists U 2 Yg such that
U \ M.p;k/
D W:
a) IU;Fg \ M.p;k/
D .IW;Fg /M.p;k/ ,
D .RW;Fg /M.p;k/ ,
b) RU;Fg \ M.p;k/
.
where W D U \ M.p;k/
such that in M ,
554
8 | principles for !1
from parameters in H.!1 /M0 . Then again by (4.1) it follows that there exists p 2 j.g/
such that p 2 D.
Therefore p0 is as desired.
t
u
|
NS
The basic analysis of Pmax
-extension follows easily from the iteration lemmas by
the usual arguments.
NS
Theorem 8.58. Assume AD L.R/ . Suppose G Pmax
is L.R/-generic. Then
L.R/G !1 -DC
and in L.R/G:
(1) P .!1 /G D P .!1 /;
(2) the sentence
AC
holds;
NS
8.2 Pmax
555
NS
is !-closed. By Theorem 8.47, for each x 2 R, there
Proof. By Corollary 8.52, Pmax
|NS
exists p 2 Pmax such that
x 2 M.p;0/ ;
NS
by Corollary 8.51, these conditions are dense in Pmax
.
Fix q0 2 G.
|NS
NS
(1.3) Suppose that D Pmax
is a dense set which is denable in the structure
hH.!1 /; A; 2i
from parameters in H.!1 /
M.p0 ;0/
. Then
NS
\ M.p0 ;0/ j p0 < p 2 M.p0 ;1/
p 2 Pmax
M.p0 ;0/
B \ !1
Let
2 M.p0 ;1/ :
jp0 ;G W hM.p0 ;k/ W k < !i ! hM.p
W k < !i
0 ;k/
NS
p 2 Pmax
\ M.p0 ;0/ j p0 < p 2 M.p0 ;1/
it follows that
B D jp0 ;G .b/
M.p
;0/
where b D B \ !1 0 .
This proves (1). A similar argument shows that
L.R/G !1 -DC:
The remaining claims, (2) and (3), are immediate consequences of (1) and the deni|NS
tion of the order on Pmax
.
t
u
556
8 | principles for !1
|
NS
We now begin the analysis of the nonstationary ideal on !1 in the Pmax
-extension.
Our goal is to show that the ideal is saturated. We begin with the following lemma
|NS
-extension.
which is the analog of Lemma 6.77 for the Pmax
NS
G Pmax
is L.R/-generic. Then in L.R/G, for every set A 2 P .R/ \ L.R/ the set
X hH.!2 /; A; 2i j MX is A-iterable and X is countable
contains a club, where MX is the transitive collapse of X .
Proof. The proof is identical to that of Lemma 6.77 using the basic analysis provided
by Theorem 8.58 (i. e. using Theorem 8.58 in place of Theorem 6.74) and using the
|NS
is !-closed (Corollary 8.52 in place of Theorem 6.73).
t
u
fact that Pmax
Remark 8.60. An immediate corollary of Lemma 8.59 is the following. Assume AD
|NS
is L.R/-generic. Then in L.R/G, INS is semiholds in L.R/ and that G Pmax
saturated. The verication is a routine application of Lemma 4.24.
t
u
Assume AD holds in L.R/ and that
|
NS
G Pmax
is L.R/-generic. Then it is not difcult to show that in L.R/G, the set YG is empty.
However one can force over L.R/G to make YG nonempty. The resulting model is
|NS
. We shall dene and briey
itself a generic extension of L.R/ for a variant of Pmax
|NS
analyze this variant which we denote Umax .
|NS
The basic property of Umax
is the following. Suppose that AD holds in L.R/ and
that
|NS
G Umax
is L.R/-generic. Then
L.R/G D L.R/gY
where in L.R/G;
|
NS
is L.R/-generic,
(1) g Pmax
NS
.
We now dene Umax
NS
8.2 Pmax
557
NS
is the set of pairs .p; f / such that
Denition 8.61. Umax
NS
,
(1) p 2 Pmax
NS
The ordering on Umax
is dened as follows:
NS
if p1 < p0 in Pmax
and for all 2 dom.f0 /,
j.f0 .// D f1 .j.// \ M.p
0 ;0/
where
j W hM.p0 ;k/ W k < !i ! hM.p0 ;k/ W k < !i
is the unique iteration such that j.F.p0 / / D F.p1 / .
t
u
NS
NS
Suppose that G Umax
is a lter. Then G projects to dene a lter FG Pmax
.
|NS
The lter G is semi-generic if the projection FG is a semi-generic lter in Pmax . We
|NS
is a semi-generic lter. Let
x some more notation. Suppose that G Umax
fG D jp;FG .f / j .p; f / 2 G:
Thus fG is a function with domain,
dom.fG / D sup!2LA j A 2 P .!1 /FG :
For each 2 dom.fG /,
fG ./ P .!1 /FG
NS
-extension of L.R/ is a routine generalization of the analThe analysis of the Umax
|NS
ysis of the Pmax -extension of L.R/. We summarize the basic results in the next theorem the proof of which we leave as an exercise for the dedicated reader.
NS
Theorem 8.62. Assume AD L.R/ . Suppose G Umax
is L.R/-generic.
F D FFG and let
Y D .YFG /L.R/G :
Then in L.R/G:
|
NS
;
(1) FG is L.R/-generic for Pmax
Let
558
8 | principles for !1
t
u
|
NS
Suppose that for each x 2 R there exists p 2 Pmax
such that
x 2 M.p;0/ :
We x some more notation. Suppose that
|
NS
G Pmax
IG D [M.p;0/
\ .INS /M.p;1/ j p 2 G
NS
Pmax
H.!1 /
denes a term for a dense subset of .P .!1 /G n IG ; / if the following conditions are
satised.
(1) is a set of pairs .p; b/ such that
M.p;0/
b !1
and such that
NS
(2) For each .p0 ; b0 / 2 Pmax
H.!1 / such that
M.p0 ;0/
b0 !1
and such that
there exists .p1 ; b1 / 2 such that p1 < p0 and such that b1 j.b0 / where
j W hM.p0 ;k/ W k < !i ! hM.p
W k < !i
0 ;k/
NS
is a semi-generic lter. Then
Suppose G Pmax
NS
8.2 Pmax
559
NS
1
Lemma 8.63 (
2 -Determinacy). Suppose that Pmax H.!1 / denes a term
NS
for a dense subset of .P .!1 /G n IG ; / and that q0 2 Pmax
. Suppose that strong
condensation holds for H.!3 /. Then there is a semi-generic lter
NS
G Pmax
and a set
Y0 YG
such that the following hold where for each p 2 G,
W k < !i
jp;G W hM.p;k/ W k < !i ! hM.p;k/
G n IU;FG
D .IW;FG /M.p;k/ ,
a) IU;FG \ M.p;k/
b) RU;FG \ M.p;k/
D .RW;FG /M.p;k/ ,
where W D U \ M.p;k/
.
8 | principles for !1
560
NS
enumerate all the dense subsets of Pmax
which are rst order denable in the structure
hH.!1 /; ; 2i:
We require that for each limit ordinal ,
D j <
contains all the dense sets which are denable with parameters from
f ./ j < :
Let X be the set of t ! such that t codes a pair .; p/ where < !1 and
|NS
.
p 2 Pmax
For each < !3 let D !
. Thus
h W < !3 i
is the increasing enumeration of the limit ordinals (with 0) less than !3 .
Suppose < !3 then for each k < !
h. C k C 2/ D f ./:
Suppose < !3 and f ./ X . Then
h. / D f ./:
Suppose < !3 and f ./ 2 X . Let .; p/ be the pair coded by f ./. Then
h. / D f . /
where is least such that f . / 2 D and such that f . / p.
Finally for each < !3 ,
1 if f ./ 2 ;
h. C 1/ D
0 otherwise.
Just as in the proof of Lemma 8.55, h witnesses strong condensation for H.!3 /.
The additional feature we have obtained here is that (using the notation from the proof
of Lemma 8.55) for each 2 S,
\ M 2 M :
Let
hp W < !1 i
be as constructed in the proof of Lemma 8.55 using the function h and the sequence
hD W < !1 i:
Let G
|NS
Pmax
be the lter,
|
NS
j for some < !1 ; p < p:
G D p 2 Pmax
NS
8.2 Pmax
D .IW;FG /M.p;k/ ,
a) IU;FG \ M.p;k/
D .RW;FG /M.p;k/ ,
b) RU;FG \ M.p;k/
.
where W D U \ M.p;k/
M D h./ j <
and let
h D hj:
Let S be the set of < !3 such that
(2.1) M is transitive,
(2.2) h 2 M for all < ,
(2.3) hM ; h ; 2i ZFC n Powerset,
M
(2.4) !2
M
exists and !2
2 M ,
561
562
8 | principles for !1
Suppose that
F Q
is a lter which is M -generic and let U D UF . We shall prove that
G n IU;FG
is dense in .P .!1 /G n IU;FG ; /.
We rst prove that the relevant lters exists. More precisely suppose that
D P .Q/
is set of dense subsets of Q such that jDj !1 . We prove that there exists a lter
F Q
such that W 2 F and such that F is D-generic. The proof is essentially the same as
the proof that YG ;.
Fix D and let
X hH.!3 /; h; 2i
be the elementary substructure of elements which are denable in the structure with
parameters from !1 [ D. For each < !1 let
X D f .s/ j f 2 X and s 2 <! ;
let 2 S be such that
and let G
Let
|NS M
.Pmax
/
hX ; h; 2i hM ; h ; 2i;
be the lter generated by the set, p j < .
C D X \ !1 j < !1
(3.4) !1
M
exists and !1
(3.5) H.!1 /
M
M
(3.6) < !1
2 M ,
M
D h./ j < !1
.
NS
8.2 Pmax
563
Y D Y.p ;0/
. /
j; W M.p ;0/ ! M.p
;0/
W 2 Y
and j; .W / W ;
(5.2) the lter F Q generated by the set jp ;G .W / j < !1 is D-generic.
Clearly one can require that any given element of Q belong to F .
Let 0 2 S be least such that !1 < 0 . Thus G 2 M 0 . Suppose that F0 Q is a
lter which is M 0 -generic and let U0 2 YG be such that
[F0 U0 :
We prove that
G n IU0 ;FG
is dense in .P .!1 /G n IU0 ;FG ; /. This is an immediate consequence of the genericity
of F0 . To see this suppose that p 2 G, W 2 jp;G .Y.p;0/ / and that W 2 F0 .
The key point is that
INS \ P .!1 /G D .\IU;FG j U 2 YG / \ P .!1 /G :
8 | principles for !1
564
W1 D U \ P .!1 /M.p;0/ :
If W 6 U then W W1 and it follows that for each U 2 YG either W U or
.!1 n A/ 2 IU;FG .
Now suppose that B 2 P .!1 /G , B \ A D ;, and that B is stationary; i. e. that
B .\IU;FG j U 2 YG / \ P .!1 /G :
Let U 2 YG be such that B IU;FG . The ultralter U must exist since
INS \ P .!1 /G D .\IU;FG j U 2 YG / \ P .!1 /G :
Necessarily W U . Therefore there exists q 2 G such that
(6.1) q < p,
G n IU0 ;FG
is dense in .P .!1 /G n IU0 ;FG ; /. Thus Y0 satises the requirements of the lemma. u
t
Lemma 8.63 yields the following variation of Lemma 8.57.
|
NS
H.!1 / denes a term for a dense
Lemma 8.64 (ADL.R/ ). Suppose that Pmax
subset of P .!1 / n IG and that
2 L.R/:
NS
Let A be the set of x 2 R which code an element of . Then for each q0 2 Pmax
there
|NS
exists p0 2 Pmax such that p0 < q0 and such that:
NS
8.2 Pmax
565
NS
\ M.p0 ;0/ such that g0 2 M.p0 ;0/ , such that
(3) There exists a lter g0 Pmax
p0 < p
for each p 2 g0 , and such that in M.p0 ;0/ ;
a) g0 is semi-generic and F D Fg0 ,
b) for each U 2 Y.p0 ;0/ ,
. \ H.!1 /M.p0 ;0/ /g0 n IU;F
is dense in .P .!1 /g0 n IU;F ; /,
where F D F.p0 / .
Proof. The proof is in essence identical to the proof of Lemma 8.57, using Lemma 8.63
in place of Lemma 8.55.
Let B be the set of x 2 R such that x codes an element of the rst order diagram
of the structure
hV!C1 ; A; 2i:
Let B be the set of x 2 R such that x codes an element of the rst order diagram of
the structure
hV!C1 ; B; q0 ; 2i:
Thus B 2 L.R/.
By Theorem 8.42 applied to B , there exist a countable transitive model M and an
ordinal 2 M such that the following hold.
(1.1) M ZFC.
(1.2) is a Woodin cardinal in M .
(1.3) B \ M 2 M and hV!C1 \ M; B \ M; 2i hV!C1 ; B ; 2i.
(1.4) B \ M is C -weakly homogeneously Suslin in M .
(1.5) Suppose is the least inaccessible cardinal of M . Then strong condensation
holds for M in M .
|
NS M
/ .
(1.6) q0 2 .Pmax
NS M
/
g .Pmax
and a set Y0 .Yg /M such that the following hold where for each p 2 g,
W k < !i
jp;g W hM.p;k/ W k < !i ! hM.p;k/
566
8 | principles for !1
(2.1) q0 2 g.
(2.2) Suppose that p 2 g. For each W 2 jp;g .Y.p;k/ /, there exists U 2 Y0 such that
D W:
U \ M.p;k/
../M /g n IU;Fg
D .IW;Fg /M.p;k/ ,
a) IU;Fg \ M.p;k/
D .RW;Fg /M.p;k/ ,
b) RU;Fg \ M.p;k/
.
where W D U \ M.p;k/
NS
8.2 Pmax
567
such that in M ,
(4.1) there exists a countable iteration
j W .M0 ; I0 ; a0 / ! .M0 ; I0 ; a0 /
such that j.F0 / D FO and such that
.M0 ; j.Y0 // D .MO 0 ; YO0 /;
(4.2) hMO k W k < !i is B \ M -iterable.
By (1.3), (1.4) and (4.1),
hV!C1 \ MO 0 ; A \ MO 0 ; 2i hV!C1 \ M; A \ M; 2i:
Therefore since
hV!C1 \ M; B \ M; 2i hV!C1 ; B; 2i:
it follows that
(5.1) hV!C1 \ MO 0 ; A \ MO 0 ; 2i hV!C1 ; A; 2i,
|NS
,
(5.2) .h.MO k ; YOk / W k < !i; FO / 2 Pmax
t
u
As a corollary to Lemma 8.64 we obtain the following theorem which we shall use
|NS
to prove that the nonstationary ideal is !2 -saturated in the Pmax
-extension of L.R/.
568
8 | principles for !1
|NS
Umax
NS
is L.R/-generic. Let FG be the induced lter on Pmax
and let
F D FFG :
Suppose that D 2 L.R/FG is dense in
.P .!1 / n INS ; /:
Then for each ultralter
U 2 YFG ;
the set D n IU;FG is dense in
.P .!1 / n IU;FG ; /:
Proof. This is immediate. Let
|
NS
H.!1 /
Pmax
be a set in L.R/ which denes a term for D. Let A be the set of x 2 R which code an
element of .
Fix U 2 YFG and x a set
S 2 P .!1 / n IU;FG :
By Theorem 8.62, there exists a condition .p; f / 2 G such that for some s 2 M.p;0/
and for some u 2 Y.p;0/ ,
S D jp;FG .s/
and
where
jp;FG .u/ D U \ M.p;0/
;
jp;FG W hM.p;k/ W k < !i ! hM.p;k/
W k < !i
NS
(1.3) There exists a lter g0 Pmax
\ M.p;0/ such that g0 2 M.p;0/ , such that
p<q
for each q 2 g0 , and such that in M.p;0/ ;
a) g0 is semi-generic,
b) F0 D Fg0 ,
NS
8.2 Pmax
569
.Iu ;F /M.p;0/ D IU;F \ M.p;0/
jp;FG .b/ A:
t
u
To apply Theorem 8.65 we need the following lemma which is an immediate corollary
of Lemma 8.59.
Lemma 8.66. Assume ADL.R/ and suppose
|
NS
G Umax
is L.R/-generic. Then in L.R/G, for every set A 2 P .R/ \ L.R/ the set
X hH.!2 /; A; 2i j MX is A-iterable and X is countable
contains a club, where MX is the transitive collapse of X .
|
NS
Proof. Let FG Pmax
be the L.R/-generic lter given by G. By Theorem 8.62(2),
t
u
570
8 | principles for !1
NS
(3) g Umax
\ M, g is L.R/M -generic and
M D L.R/M g:
(4) F D .Fg /M .
(5) Y D .Yg /M .
(6) I D .IU;F /M j U 2 Y .
(7) .M; I/ is iterable.
t
u
|
h.M0 ; I0 /; g0 ; Y0 ; F0 i 2 M0 NS
and that strong condensation holds for H.!3 /. Then there is an iteration
j W .M0 ; I0 / ! .M0 ; I0 /
of length !1 and a set
Y U P .!1 / j U is a uniform ultralter on !1
such that the following hold where F D j0 .F0 /.
(1) For each U 2 Y , U \ M0 2 j.Y0 /.
(2) For each U 2 Y , the ideal IU;F is proper,
where W D M0 \ U .
(3) Let I be the ideal on !1 which is dual to the lter,
F D \U j U 2 Y ;
then
\IU;F j U 2 Y I:
(4) Suppose that U0 is a uniform ultralter on !1 such that
U0 \ M0 2 j.Y0 /:
a) There exists U1 2 Y such that
U0 \ M0 D U1 \ M0 :
b) Suppose that
\U j U 2 Y U0 :
Then U0 2 Y .
t
u
NS
8.2 Pmax
571
Lemma 8.68 combined with Lemma 8.46 easily yields the following version of
|
Lemma 8.46 for M0 NS .
Theorem 8.69. Assume AD holds in L.R/ and that A R is a set in L.R/. Suppose
that
|
h.M0 ; I0 /; g0 ; Y0 ; F0 i 2 M0 NS :
Then there is a condition
|NS
.h.MO k ; YOk / W k < !i; FO / 2 Pmax
and an iteration
j W .M0 ; I0 / ! .M0 ; I0 /
such that
(1) hMO k W k < !i is A-iterable,
(2) j 2 MO 0 ,
(3) j.F0 / D FO ,
(4) j.Y0 / D U \ M0 j U 2 YO0 ,
(5) for each U 2 YO0 ,
and
where W D M0 \ U .
Lemma 8.66 yields the following strengthening of Lemma 8.64. The difference is
in the statement of (3b).
|
NS
Lemma 8.70 (ADL.R/ ). Suppose that Pmax
H.!1 / denes a term for a dense
subset of P .!1 / n IG . Let A be the set of x 2 R which code an element of . Then for
|NS
|NS
there exists p0 2 Pmax
such that p0 < q0 and such that:
each q0 2 Pmax
NS
(3) There exists a lter g0 Pmax
\ M.p0 ;0/ such that g0 2 M.p0 ;0/ , such that
p0 < p
for each p 2 g0 , and such that in M.p0 ;0/ ;
572
8 | principles for !1
|
NS
NS
Proof. Fix q0 2 Pmax
. Suppose G Umax
is L.R/-generic such that
q0 2 FG
|NS
Pmax
where FG
is the induced L.R/-generic lter.
We work in L.R/G. Let be least such that
L .R/G 2 L.R/G
and let N D L .R/G.
We claim that by Lemma 8.66 and Lemma 4.24, the set
X N j X is countable and NX is strongly iterable
is stationary in P!1 .N /. Here NX is the transitive collapse of X . To see this note that
in L.R/G,
L.R/ D !3 :
Therefore by Lemma 4.24, if M is a transitive set of cardinality !2 such that
M ZFC ;
and such that H.!2 / M , then the set
X M j X is countable and MX is strongly iterable
contains a club in P!1 .M / where for each X M , MX is the transitive collapse of
X.
Now x a function
H W N <! ! N
and let
ZN
be an elementary substructure of cardinality !2 such that
H Z <! Z
and such that H.!2 / Z.
Let M be the transitive collapse of Z. Let
HZ W M <! ! M
be the image of H under the collapsing map.
Thus there exists a countable elementary substructure
X M
such that MX is strongly iterable and such that
HZ X <! X:
Let X be the preimage of X under the transitive collapse of Z. Thus
X N;
H .X /<! X and
MX D MX
NS
8.2 Pmax
573
D n IU;FG
8 | principles for !1
574
Thus:
(2.1) INS D \I j I 2 J.
(2.2) Suppose that I0 2 J, I1 2 J and that for some A !1 ,
a) I0 B !1 j B \ A 2 I1 ,
b) !1 n A I1 .
Then I0 D I1 .
Let G0 ; F0 ; F0 ; J0 ; Y0 be the image of G; FG ; FG ; J; YFG under the transitive collapse of X0 .
|NS
The partial order Umax
is !-closed and so
N0 ; F0 ; G0 ; F0 ; D0 ; Y0 2 L.R/:
Further
h.N0 ; J0 /; G0 ; Y0 ; F0 i 2 M0 NS :
We now work in L.R/.
By Theorem 8.69 there exists a condition
|
NS
.h.MO k ; YOk / W k < !i; FO / 2 Pmax
and an iteration
such that
(3.1) hMO k W k < !i is A-iterable,
(3.2) j 2 MO 0 ,
(3.3) j.F0 / D FO ,
(3.4) j0 .Y0 / D U \ N0 j U 2 YO0 ,
(3.5) for each U 2 YO0 ,
and
where W D N0 \ U .
Let
NS
8.2 Pmax
575
h.N0 ; J0 /; G0 ; Y0 ; F0 i 2 M0 NS ;
it follows that
NS
is L.R/-generic. Then in
Theorem 8.71. Assume AD L.R/ . Suppose G Pmax
L.R/G,
(1) IG is !2 -saturated,
(2) INS D IG .
Proof. By Theorem 8.58,
P .!1 / D P .!1 /G
and IG D INS . Therefore it sufces to show that if
D P .!1 /G n IG
is dense then there exists a set
D0 D
such that D0 is predense in .P .!1 /G n IG ; / and such that jD0 j !1 .
|NS
Let Pmax
H.!1 / be a set in L.R/ which denes a term for D. Let A R
be the set of x 2 R such that x codes an element of .
8 | principles for !1
576
NS
(1.3) There exists a lter g0 Pmax
\ M.p0 ;0/ such that g0 2 M.p0 ;0/ , such that for
each p 2 g0 ,
p0 < p;
Thus D0 D and
jD0 j !1 :
5D0
contains the critical sequence of the iteration dening jp0 ;G and so D0 is necessarily
predense in .P .!1 /G n IG ; / since
.P .!1 /G n IG ; / D .P .!1 / n INS ; /:
t
u
CC
577
NS
Corollary 8.72. Assume AD L.R/ . Suppose G Pmax
is L.R/-generic.
Then in L.R/G,
YG D ;:
Proof. We note that the following must hold in L.R/G. Suppose that S !1 is
stationary. Then there exists a set A !1 such that both
2 S j A n F ./ is nite
and
2 S j A \ F ./ is nite
are stationary.
Suppose YG ; and let U 2 YG . Thus, since
P .!1 / D P .!1 /G ;
it follows that IU;FG is a proper ideal. But INS is !2 -saturated and so for some stationary set S !1 ,
IU;FG D INS jS D T !1 j T n S 2 INS :
t
u
NS
Lemma 8.73. Assume AD L.R/ . Suppose G Pmax
is L.R/-generic. Let F D FG .
Then in L.R/G the following holds. There exists a co-stationary set S !1 such
that for all ultralters U P .!1 /, if p 2 PU and
Zp;F INS ;
then
Zp;F \ S INS :
8.3
t
u
CC
NS
NS
The Umax
-extension of L.R/ is a generic extension of the Pmax
-extension. The relevant partial order is a product of a partial order PF which is dened in Denition 8.75.
The denition of PF is closely related to two renements of |NS one of which we
|NS
now dene. These renements in turn yield an absoluteness theorem for the Pmax
extension. It is not clear if the version we prove is optimal and as we have indicated,
more elegant versions are likely possible.
578
8 | principles for !1
t
u
t
u
The partial order PF is analogous to the partial order PNS which we dened in
Section 6.1. There is however an interesting difference. It is not difcult to show that
assuming ./, the partial order PNS is not !2 -cc. However if INS is !2 -saturated, which
|NS
is the case in L.R/Pmax , then PF is trivially !2 -cc for any function F which witnesses
|NS . More is actually true.
CC
579
X FS;F
then S n SX 2 INS .
Dene .X / D b where b 2 P .!1 /=INS is the element given by SX . The element b
is unambiguously dened.
The function induces the required isomorphism of RO.PF / with a complete
t
u
boolean subalgebra of P .!1 /=INS .
We also note the following reformulation of Corollary 8.72.
|
NS
Lemma 8.77. Assume AD L.R/ . Suppose G Pmax
is a semi-generic lter such that
G is L.R/-generic and such that
P .!1 / D P .!1 /G :
Then
RO.PF /
has no atoms where F D FG .
t
u
C
Lemma 8.76 and Lemma 8.77 suggest the following renement of |NS .
CC
t
u
580
8 | principles for !1
Remark 8.79. As we have already remarked, the most elegant manifestation of |NS
would be to have for some ultralter U on !1 ,
(1) U extends the club lter,
(2) the boolean algebra RO.PU / is isomorphic to a complete boolean subalgebra of
P .!1 /=INS .
Any function
F W !1 ! !1 !
CC
inducing the isomorphism for (2), witnesses that |NS holds. However by Lemma 8.25,
CC
(1) and (2) cannot both hold for any ultralter U . |NS in some sense gives the best
possible approximation to (1) and (2); cf. Corollary 8.88.
t
u
There is an interesting question.
Suppose that
F W !1 ! !1 !
C
is a function which witnesses that |NS holds. Can the boolean algebra,
RO.PF /;
be atomic?
|
NS
-extension easily yields,
The basic analysis of the Pmax
NS
Theorem 8.80. Assume AD L.R/ . Suppose G Pmax
is L.R/-generic. Then
CC
L.R/G |NS :
|
NS
, the function FG
Proof. By Theorem 8.58 and the denition of the order on Pmax
witnesses that
L.R/G |NS
and in L.R/G;
P .!1 / D P .!1 /G :
|
NS
that the function FG witnesses that
It follows from the denition of Pmax
C
L.R/G |NS :
By Theorem 8.71, the nonstationary ideal is !2 -saturated in L.R/G and so by
Lemma 8.76, the function FG witnesses that
CC
L.R/G |NS :
|
t
u
|
NS
NS
-extension of L.R/, identifying the Umax
We continue our analysis of the Pmax
|NS
extension of L.R/ as a generic extension of the Pmax -extension of L.R/. The relevant
partial order, as we have indicated, is simply a product of PF .
CC
581
p2 ./ p1 ./
t
u
in PF .
Lemma 8.82. Suppose that
F W !1 ! !1 !
CC
is a function which witnesses that |NS holds. Then the partial order PF is .!1 ; 1/distributive.
Proof. Suppose that g Coll.!2 ; P .!1 // is V -generic. Since g is V -generic for a
partial order which is .< !2 /-closed in V , it follows that in V g, F witnesses that
CC
|NS holds.
Therefore we may assume without loss of generality that 2@1 D @2 .
Suppose that G PF is V -generic. Then, by reorganizing G as a subset of !2 ,
V G D V A
where A
is a set such that A \ 2 V for all < !2V . This is a consequence of
C
the fact that F witnesses |NS in V G, see Denition 8.74(2). Since PF is !2 -cc in V
t
u
it follows that V is closed under !1 -sequences in V G.
!2V
NS
NS
The next four theorems detail the relationship between Pmax
and Umax
. We shall
not need these theorems, we simply state them for completeness. The proofs are not
difcult and we leave the details to the reader.
NS
is L.R/-generic. Let
Theorem 8.83. Assume AD L.R/ . Suppose that G Umax
F D FFG and let
t
u
8 | principles for !1
582
where g
is L.R/-generic. Let F D Fg be the function witnessing |NS given
by g. Then PF is .!1 ; 1/-distributive and further suppose G PF is V -generic.
Then in V G;
(1) IU;F is a proper ideal,
(2) INS is not saturated,
(3) IU;F D sat.INS /,
(4) IU;F is a saturated ideal,
where U D [G.
t
u
Theorem 8.84 combined with Theorem 8.58 yields the following theorem.
Theorem 8.85. Assume ADL.R/ and that
V D L.R/g
|NS
Pmax
where g
is L.R/-generic. Let F D Fg be the function witnessing |NS given
by g. Suppose p 2 g and let
j W hM.p;k/ W k < !i ! hM.p;k/
W k < !i
(2) .IW;F /M.p;0/ D IU;F \ M.p;0/
.
(3) .RW;F /M.p;0/ D RU;F \ M.p;0/
.
t
u
where g
is L.R/-generic. Let F D Fg be the function witnessing |NS given
by g.
Then QF is .!1 ; 1/-distributive.
Suppose G QF is V -generic and for each < !2 let
U D [p./ j 2 dom.p/ and p 2 G;
and let
Y D U j < !2 :
CC
583
Then in V G:
(1) Y D Yg ;
(2) For each U 2 Y ,
a) IU;F is a proper ideal,
b) IU;F is a saturated ideal;
(3) INS D \IU;F j U 2 Y ;
(4) For each U 2 Y , INS \ U D ;.
t
u
CC
One corollary of Lemma 8.82 is that |NS cannot hold in L. More generally, strong
CC
Corollary 8.87. Assume that strong condensation holds for H.!2 /. Then |NS fails.
Proof. The proof is a modication of the proof of Lemma 8.25. We sketch the argument under the additional hypothesis that V D L. The proof from strong condensation
for H.!2 / is essentially the same.
Suppose that G PF is V -generic and let U 2 V G be the ultralter on !1 given
by G;
U D X j X 2 G:
Since F witnesses |NS in V it follows that U is a V -ultralter on !1V . However F
CC
witnesses |NS in V and so by Lemma 8.82, PF is .!1 ; 1/-distributive in V . This
implies
.P .!1 //V D .P .!1 //V G
and so U is an ultralter on !1 in V G.
Let < !2 be least such that
(1.1) F 2 L ,
(1.2) L ZC,
CC
X L
584
8 | principles for !1
is a function which witnesses that |NS holds and that INS is !2 -saturated. Then there
exists a complete boolean subalgebra
B P .!1 /=INS
such that
RO.PF PU / B;
where U 2 V PF is the ultralter on !1 given by the generic lter for PF .
t
u
NS
We now come to the absoluteness theorem for the Pmax
-extension. We rst prove
|NS
a strong version of the homogeneity of Pmax . This is a corollary of the following
iteration lemma.
h.M0 ; I0 /; g0 ; Y0 ; F0 i 2 M0 NS ;
|
h.M1 ; I1 /; g1 ; Y1 ; F1 i 2 M0 NS ;
and that strong condensation holds for H.!3 /. Then there exist iterations
j0 W .M0 ; I0 / ! .M0 ; I0 /
and
j1 W .M1 ; I1 / ! .M1 ; I1 /
of length !1 and a bijection
W j0 .Y0 / ! j1 .Y1 /
such that:
(1) !1 n < !1 j j0 .F0 /./ D j1 .F1 /./ 2 INS .
(2) Suppose that W0 2 j0 .Y0 / and W1 D .W0 /. Then for all A0 2 W0 and for all
A1 2 W1 ,
A0 \ A1 INS :
CC
585
(3) Suppose that U P .!1 / is an ultralter such that U \ M0 2 j0 .Y0 / and such
that
U \ M1 2 j1 .Y1 /:
a) The ideal IU;F is proper.
b) For each i 2 0;1,
(1.4) !2
M
exists and !2
2 M ,
586
8 | principles for !1
0 .!1 /M ;
then let 0 be least such that h.0 / D .g0 ; g1 / and
0
.I0 /,
(2.1) for some I 2 j0;
0
g0 .P .0 / \ M00 n I /
and g0 is M00 -generic,
1
.I1 /,
(2.2) for some I 2 j0;
0
g1 .P .0 / \ M10 n I /
and g1 is M10 -generic.
Let
.G00 ; G10 / D h.0 / D .g0 ; g1 /:
Otherwise let 0 2 S be least such that
0 D .!1 /M0 :
Let 0 be least such that h.0 / D .g0 ; g1 / and
CC
587
0
(3.1) for some I 2 j0;
.I0 /,
0
g0 .P .0 / \ M00 n I /
and g0 is M00 -generic,
1
.I1 /,
(3.2) for some I 2 j0;
0
g1 .P .0 / \ M10 n I /
and g1 is M10 -generic,
0
1
.F0 /.0 / D j0;
.F1 /.0 /.
(3.3) j0;
0 C1
0 C1
0
1
Since .j0;
; j0;
; 0 / satises the requirements of the lemma interpreted in M 0 , 0
0
0
exists. Let
A0 \ A1 D ;:
Let 0 be least such that h.0 / is such a pair .W0 ; W1 / 2 0 and let 1 be least such
that h.1 / D .A0 ; A1 / with A0 2 W0 , A1 2 W1 and
A0 \ A1 D ;:
Let 0 be least such that h.0 / D .g0 ; g1 / and
(4.1) (2.1)(2.2) hold,
(4.2) A0 2 g0 and A1 2 g1 .
Let
.G00 ; G10 / D h.0 / D .g0 ; g1 /:
588
8 | principles for !1
(6.3) A0 2 g0 and A1 2 g1 .
We can ensure (6.2) holds because in M 0 , W0 [ W1 can be extended to an ultralter.
Let
.G00 ; G10 / D h.0 / D .g0 ; g1 /:
0
1
Finally we suppose that in M 0 , (3) fails for .j0;
; j0;
; 0 /.
0
0
Let 0 be least such that:
(7.1) h.0 / 2 M 0 ;
(7.2) M 0 h.0 / is a uniform ultralter on !1 ;
0
(7.3) h.0 / \ M00 2 j0;
.Y0 /;
0
1
(7.4) h.0 / \ M10 2 j0;
.Y1 /;
0
W0 D h.0 / \ M00 ,
c) there exists
W1 D h.0 / \ M10 .
CC
589
Let
(8.1) U D h.0 /,
(8.2) W0 D U \ M00 ,
0
0
(8.3) I0 D .IW0 ;F /M0 where F D j0;
.F0 /,
0
(8.4) W1 D U \ M10 ,
0
0
(8.5) I1 D .IW1 ;F /M1 where F D j0;
.F1 /.
0
There are several subcases. First suppose that (7.5(a)) holds. Let 0 be least such
that h.0 / D .g0 ; g1 / and
(9.1) g0 .P .0 / \ M00 n I0 / and g0 is M00 -generic,
(9.2) g1 .P .0 / \ M10 n I1 / and g1 is M10 -generic,
0
1
(9.3) j0;
.F0 /.0 / D j0;
.F1 /.0 /,
0 C1
0 C1
0
.F0 /.0 / is M 0 -generic for .PU /M0 .
(9.4) j0;
0 C1
Let
.G00 ; G10 / D h.0 / D .g0 ; g1 /:
Otherwise (7.5(a)) fails. Hence either (7.5(b)) holds or (7.5(c)) holds. We next suppose
that (7.5(b)) holds. Let 1 be least such that h.1 / D .p; S / witnessing (7.5(b)). Let 2
be least such that h.2 / D q and
(10.1) q 2 .PU /M0 ,
(10.2) q p,
(10.3) .Zq;F /M0 \ S 2 .IU;F /M0 ,
0
where F D j0;
.F0 /.
0
Let 0 be least such that h.0 / D .g0 ; g1 / and
Let
.G00 ; G10 / D h.0 / D .g0 ; g1 /:
The nal case is that both (7.5(a)) and (7.5(b)) fail. In which case (7.5(c)) holds.
This is essentially the same as the case that (7.5(b)) holds: Let 1 be least such that
h.1 / D .p; S / witnessing (7.5(c)). Let 2 be least such that h.2 / D q and
8 | principles for !1
590
Let
.G00 ; G10 / D h.0 / D .g0 ; g1 /:
This completes the inductive denition of h.G0 ; G1 ; / W < !1 i.
Let
.j0 ; j0 ; / D .j00;!1 ; j10;!1 ; !1 /:
We claim that .j0 ; j0 ; / satises the requirements of the lemma.
We prove (1) holds. For this we rst prove that for all .W0 ; W1 / 2 , if A0 2 W0
and if A1 2 W1 then
A0 \ A1 ;:
Suppose this fails. Let 0 be least such that h.0 / is such a pair .W0 ; W1 / 2 and
let 1 be least such that h.1 / D .A0 ; A1 / with A0 2 W0 , A1 2 W1 and
A0 \ A1 D ;:
Suppose
X hH.!3 /; h; 2i
is a countable elementary substructure with 2 X .
Let 0 D X \ !1 and let MX be the transitive collapse of X . Thus
MX D M
where D MX \ Ord and so 2 S. Let .0X ; 1X / be the image of .0 ; 1 / under the
collapsing map. Thus
h.1X / D .X \ A0 ; X \ A1 /:
It follows that .G00 ; G10 / was dened at stage 0 using .0X ; 1X / choosing 0 least
such that h.0 / D .g0 ; g1 / and
(14.1) (2.1)(2.2) hold,
(14.2) A0 \ X 2 g0 and A1 \ X 2 g1 ,
CC
591
and dening
.G00 ; G10 / D h.0 / D .g0 ; g1 /:
Thus A0 \ X 2 G00 and A1 \ X 2 G10 . But this implies 0 2 A0 \ A1 which is a
contradiction.
This proves that for all .W0 ; W1 / 2 , W0 [ W1 has the nite intersection property.
Therefore there is a closed unbounded set C !1 to which this reects; if 0 2 C
then for all .W0 ; W1 / 2 0 , W0 [ W1 has the nite intersection property. Therefore,
by inspection of the inductive construction, for all 0 2 C , if there exists 2 S such
that
0 D .!1 /M ;
then
0
1
j0;
.F0 /.0 / D j0;
.F1 /.0 /:
0 C1
0 C1
h.M0 ; I0 /; g0 ; Y0 ; F0 i 2 M0 NS ;
|
h.M1 ; I1 /; g1 ; Y1 ; F1 i 2 M0 NS ;
and that strong condensation holds for H.!3 /. Then there exist iterations
j0 W .M0 ; I0 / ! .M0 ; I0 /
and
592
8 | principles for !1
and
U \ M0 2 j0 .Y0 /
U \ M1 2 j1 .Y1 /:
where W D Mi \ U .
By Lemma 8.89, .j0 ; j1 ; / exists.
Let F D j0 .F0 /.
The desired set of ultralters Y is obtained just as in the proof of Lemma 8.41.
Let Z be the set of uniform ultralters U on !1 such that
U \ M0 2 j0 .Y0 /
CC
593
U \ M1 2 j1 .Y1 /:
j0 .Y0 / D U \ M0 j U 2 Y
j1 .Y1 / D U \ M1 j U 2 Y :
|
t
u
NS
is an immediate corollary. We isolate the relevant fact
The homogeneity of Pmax
in the following lemma.
h.M; I/; g; Y; F i 2 M0 NS
|
NS
NS
such that x 2 M. Suppose that p0 2 Pmax
and p1 2 Pmax
. There exist
NS
.h.Mk ; Yk / W k < !i; F / 2 Pmax
NS
(1) .h.Mk ; Yk / W k < !i; F0 / 2 Pmax
and .h.Mk ; Yk / W k < !i; F0 / < p0 ,
NS
(2) .h.Mk ; Yk / W k < !i; F1 / 2 Pmax
and .h.Mk ; Yk / W k < !i; F1 / < p1 ,
594
8 | principles for !1
h.M; I/; g; Y; F i 2 M0 NS
be such that x 2 M. Thus
NS M
/ :
p0 ; p1 .Pmax
NS N
/
gi .Umax
NS N
/
Fgi .Pmax
NS N
is the induced N -generic lter on .Pmax
/ .
Let
hYi ; Fi i D hYgi ; Fgi iN gi
and let
Ii D .IU;Fi /N gi j U 2 Yi :
A key point is that since .M; I/ is iterable it follows by Lemma 8.66 and Theorem 3.46,
that for each i 2 0;1, the structure .N gi ; Ii / is also iterable and so
|
h.Mi ; Ii /; gi ; Yi ; Fi i 2 M0 NS ;
where Mi D N gi . Strictly speaking Lemma 8.66 and Theorem 3.46 cannot be
applied in N g since we have only
N g ZFC C ZC C 1 -Replacement;
but both are easily seen to hold in this case.
Let y 2 R code .N ; g0 ; g1 / and let
O I;
O ; / 2 H.!1 /
.M;
be such that
O
(1.1) x 2 M,
(1.2) MO is transitive and MO ZFC C is a Woodin cardinal,
O
(1.3) IO D .I< /M ,
O I/
O is iterable,
(1.4) .M;
O
(1.5) < and MO M,
(1.6) strong condensation holds in MO for MO where is the least inaccessible cardiO
nal of M.
CC
595
O I;
O ; / follows from 1 -Determinacy, by Theorem 8.45. We
The existence of .M;
2
note that since for each x 2 R, there exists
|
h.M; I/; g; Y; F i 2 M0 NS
such that x 2 M, necessarily
12 -Determinacy holds. This follows by absoluteness.
Thus
|
where W D Mi \ U .
(2.3) Let I be the ideal on !1 which is dual to the lter,
F D \U j U 2 Y ;
then
\IU;ji .Fi / j U 2 Y I:
(2.4) j.Yi / D U \ Mi j U 2 Y .
Let F D j0 .F0 /. Thus there exists a 2 MO such that
O I; a/; Y; F i 2 M|NS :
h.M;
Let
X0 MO
8 | principles for !1
596
O
(3.1) X0 2 M,
O
(3.2) jX0 jM D !,
(3.3) ; a; Y; j0 ; j1 2 X0 ,
let MO X0 be the transitive collapse of X0 and let
IOX0 ; aX0 ; YX0 ; FX0 ; FX10
O a; Y; F; j1 .F1 / under the collapsing map. Thus
be the image of I;
O
h.MO X0 ; IOX0 ; aX0 /; YX0 ; FX0 i 2 .M|NS /M
Thus, since p 2
|NS M
O
/ ,
.Pmax
NS
p 2 Pmax
. Clearly
p < p0
Q
(since F D j.FX0 / and F D j0 .F0 /.)
Let
FQ1 D j.FX10 /:
Then
|NS
,
(5.1) .h.MQ k ; YQk / W k < !i; FQ1 / 2 Pmax
h.M0 ; I0 /; g0 ; Y0 ; F0 i 2 M0 NS
|
NS
such that x 2 M0 . Then Pmax
is homogeneous.
t
u
CC
597
Lemma 8.91 combined with Theorem 8.69 yields the following theorem.
|
NS
Theorem 8.93. Assume ADL.R/ and that V D L.R/g where g Pmax
is L.R/generic. Let F D Fg be the function witnessing |NS given by g. Suppose that
h.M0 ; I0 /; g0 ; Y0 ; F0 i 2 M0 NS :
Then there is an iteration
j W .M0 ; I0 / ! .M0 ; I0 /
of length !1 such that:
NS
.h.MO k ; YOk / W k < !i; FO / 2 Pmax
and an iteration
j0 W .M0 ; I0 / ! .j0 .M0 /; j0 .I0 //
such that
(1.1) j0 2 MO 0 ,
(1.2) j0 .F0 / D FO ,
(1.3) j0 .Y0 / D U \ j0 .M0 / j U 2 YO0 ,
(1.4) for each U 2 YO0 ,
and
.IW;FO /j0 .M0 / D IU;FO \ j0 .M0 /;
where W D j0 .M0 / \ U .
Let
NS
g0 Pmax
such that
598
8 | principles for !1
j W M ! M
a) .IW;F /M W D IU;F \ M W .
b) .RW;F /M W D RU;F \ M .
t
u
Suppose that
F W !1 ! !1 !
CC
is a function which witnesses |NS . With the following iteration lemma, several equivalent formulations for the notion that F is universal are easily identied. There may
well be fairly natural combinatorial properties of F which imply that F is universal. If
|NS
.
so this would lead to more elegant absoluteness theorems for Pmax
CC
599
j W M ! M
of length !1 and a set Y of uniform ultralters on !1 such that the following hold
where F D j.f / and where Y be the set of lters
W .P .!1 //M
such that
z 2 .Pj.f / /M j z W
is M -generic.
(1) Y D U \ M j U 2 Y .
(2) For each U 2 Y :
a) The ideal IU;F is proper.
b) .!1 ; 1PU / 2 RU;F .
c) Let W D M0 \ U . Then
.IW;F /M D IU;F \ M
and
.RW;F /M D RU;F \ M :
600
8 | principles for !1
z 2 .Pj.f / /M j z W
is M -generic.
(1.1) Suppose that U is an ultralter on !1 such that U \ M 2 Y . Then:
a) The ideal IU;F is proper.
b) .!1 ; 1PU / 2 RU;F .
c) Let W D M0 \ U . Then
.IW;F /M D IU;F \ M
and
.RW;F /M D RU;F \ M :
CC
601
(1.2) Suppose that h.S ; T / W < !1 i D j.h.S ; T / W < !1M i/: Let
h W < !1 i
be the increasing enumeration of the ordinals 2 !1 n .M \ Ord/ such that
is a cardinal in L.M/. Let
C D < !1 j D :
Then for all 2 C and for all < ,
2 S
if and only if
C 2 T :
there exists a
such that z 2 g and such that g is M -generic. To see this note that if
g0 .Pf /M
is an M-generic lter (which must exist since M is countable) then
k.t / j t 2 g0
generates an M -generic lter. Clearly we can suppose that z D k.t / for some t 2 M
by passing to a countable iterate of M if necessary. Thus we can choose g0 with t 2 g0
in which case the M -generic lter generated by the image of g0 contains z as desired.
Thus for the iteration specied above, necessarily
j.Ff / D \Y :
This combined with the usual thinning arguments, as in the proof of Lemma 8.41,
yields the set Y as required.
t
u
Lemma 8.95 combined with Lemma 8.46 yields the following lemma.
1
Lemma 8.96 (
2 -Determinacy). Suppose that
.M; f / 2 H.!1 /
and that
(i) M is transitive and
CC
602
8 | principles for !1
NS
.h.MO k ; YOk / W k < !i; FO / 2 Pmax
where W D M \ U .
t
u
Suppose that
F W !1 ! !1 !
CC
is a function witnessing |NS . We give in the next two lemmas, universality properties
of F which are each equivalent to the property that F is universal.
Lemma 8.97 (ADL.R/ ). Suppose that
F W !1 ! !1 !
CC
h.M0 ; I0 ; a0 /; Y0 ; F0 i 2 M|NS :
CC
603
.j.0 // M0
/ ;
.p; b/ 2 .RW;F
NS
.h.MO k ; YOk / W k < !i; FO / 2 Pmax
such that
(1.1) there exists a countable iteration
j W .M0 ; I0 ; a0 / ! .M0 ; I0 ; a0 /
such that j.F0 / D FO and such that
.M0 ; j.Y0 // D .MO 0 ; YO0 /:
Let
h.M; I/; g; Y; f i 2 M0 NS
be such that
NS M
where Fg .Pmax
/ is the associated .L.R//M -generic lter.
The existence of h.M; I/; g; Y; f i follows from the assumption of ADL.R/ . To see
this suppose that
|NS
G Umax
is L.R/-generic with
NS
where FG Pmax
is the induced L.R/-generic lter. Let be least such that
604
8 | principles for !1
By Lemma 8.66 and Lemma 4.24, the set of countable elementary substructures,
X L .R/G
such that MX is iterable where MX is the transitive collapse of X , is closed and unbounded in
P!1 .L .R/G/:
Choose such an elementary substructure with
.h.MO k ; YOk / W k < !i; FO /; G 2 X:
The transitive collapse of X yields
|
h.M; I/; g; Y; f i 2 M0 NS
as required.
Since F is universal there exists an iteration
jO W M ! M
of length !1 such that the following hold.
W
b) .RW;jO.f / /M
Finally let
W
D IU;F \ M W .
D RU;F \ M .
CC
NS
.h.MO k ; YOk / W k < !i; FO / 2 Pmax
and an iteration
j W M ! M
such that:
(4.1) j 2 MO 0 .
(4.2) j.f / D FO .
(4.3) Let Y be the set of W 2 MO 0 such that
z 2 .PFO /M j z W
is M -generic.
Then
Y D U \ M j U 2 YO0 :
where W D M0 \ U .
By Lemma 8.49 and Lemma 8.50, there exists
h.M0 ; I0 ; a0 /; Y0 ; F0 i 2 M|NS
and an iteration
such that
(5.1) jO.FO / D F0 ,
(5.2) for each k < !, jO.YOk / D U \ MO k j U 2 Y0 ,
(5.3) for each U 2 Y0 , for each k < !,
605
606
8 | principles for !1
Let
j0 W .M0 ; I0 ; a0 / ! .M0 ; I0 ; a0 /
be as given by (2). The induced iteration
j0 .jO.j // W M ! M
is of length !1 and is easily veried to witness that F is universal.
t
u
The proof of Lemma 8.97 easily adapts to prove Lemma 8.98 which gives another
CC
characterization of when a function witnessing |NS is universal. This characterization
|
NS
involves conditions in Pmax
.
NS
:
.h.Mk ; Yk / W k < !i; F0 / 2 Pmax
t
u
Suppose that
F W !1 ! !1 !
CC
CC
607
There exists a stationary set S !1 such that for all ultralters U P .!1 /, if
p 2 PU and
Zp;F INS ;
then
Zp;F \ S INS :
Suppose F1 W !1 ! !1 ! and F2 W !1 ! !1 ! . Then we dene F1 DE F2 if
j F1 ./ M F2 ./ is innite 2 INS :
Let F1 E be the equivalence class of F1 .
Theorem 8.99. Suppose that there are ! many Woodin cardinals with a measurable
above. Suppose that
F W !1 ! !1 !
is a function such that the following hold.
CC
|NS
Pmax
NS
.h.Mk ; Yk / W k < !i; f / 2 Pmax
:
:
608
8 | principles for !1
(1.3) Suppose that G PF is V -generic. Let U D [G and for each k < ! let
Wk D U \ Mk . Then for each k < !,
a) Wk 2 j.Yk /,
and an iteration
j W .M; I/ ! .M ; I /
(2.4) fO D j.FX /.
(2.5) Suppose that U 2 YO0 and let W D U \ M . Then
Chapter 9
Extensions of L.; R/
The main goal in this chapter in the basic analysis of the Pmax and Qmax extensions
of models larger than L.R/. One class of examples of models in which we shall be
interested are those of the form L.; R/ where P .R/ is a pointclass closed under
continuous preimages. If G Pmax is L.; R/-generic and if, for example,
L.; R/ ADR C is regular;
then
.L.P .!1 ///L.;R/G D L.R/G
and so L.; R/G ./. Thus by forcing with Pmax over larger models of AD we are
creating models of ./ with more subsets of !2 . In this fashion we can create models
in which ./ holds and in which P .!2 / is reasonably closed.
For a suitable choice of , the Pmax -extension of L.; R/ yields a model in which
Martins MaximumCC .c/ holds and in which !2 exhibits some interesting combinatorial features.
In Section 9.6 we shall consider the Pmax extension of even larger inner models
which are of the form,
L.S; ; R/
where S Ord and where, as above, P .R/. Applications include producing
extensions in which ./ holds and in which Strong Changs Conjecture holds. One
reason we consider the problem of obtaining extensions in which Strong Changs Conjecture holds is that since Strong Changs Conjecture is not generally expressible in
L.P .!2 //, it is not immediately obvious that such extensions can even exist.
In the second section of the next chapter, Section 10.2, we shall dene several more
variations of Pmax and Qmax , and consider the induced extensions of L.S; ; R/. One
application will be to show that
Martins Maximum CC .c/ C Strong Changs Conjecture
together with all the 2 sentences true in
Pmax
9 Extensions of L.; R/
610
Theorem 9.1. Suppose that A R and that every set B 2 P .R/ \ L.A; R/ is weakly
homogeneously Suslin. Then
t
u
L.A; R/Pmax ZFC C :
The following is an interesting open question.
Suppose that A R and that every set B 2 P .R/ \ L.A; R/ is weakly homogeneously Suslin. Does
L.A; R/ ADC
9.1
ADC
A D y 2 R j L S; y S; y:
There are many equivalent denitions of the
formula , for each set S Ord, the set
t
u
t
u
Suppose
2 Ord and that A
! . The set A is determined if there exists a
winning strategy for Player I or for Player II in the game on
corresponding to the set
A.
The following lemma is an easy exercise.
Lemma 9.4 (ZF). There exists a set A !1! such that A is not determined.
t
u
9.1 ADC
611
Suppose T is a tree on !
. We use the notation from Section 2.1 and dene
a set AT ! ! as follows. x 2 AT if Player I has a winning strategy in the game
corresponding to Bx
! where
Bx D Tx D f 2
! j .x; f / 2 T :
The set AT is easily veried to be 1 -borel. If the set
T ! !
!
is clopen in the product space, ! !
! , we shall say that T is an 1 -borel code of
AT . Note that in the case that
D !, if T is clopen then AT is borel. Without the
1
requirement that T be clopen, one can only deduce that AT is
1 . It is not difcult
1
1
to show that every -borel set has an -borel code.
One important feature of the 1 -borel sets is that assuming AD the property of being
1
-borel is a local property. One manifestation of this is given in the following lemma.
Suppose P .R/ is a pointclass which contains the borel sets, such that is closed
under continuous preimages, nite unions and complements. Recall from Section 2.1
that we have associated to two transitive sets, M and N , see Denition 2.18.
Lemma 9.5 (ZF + AD + DCR ). Suppose A R and that A is 1 -borel. Let be the
pointclass of sets which are projective in A. Then there exists
T 2 M
such that T is an
t
u
Assuming AD many ordinal games are determined and this is closely related to the
existence of Suslin representations for sets of reals.
We now give the denition of ADC . Recall that is the least ordinal which is not
the range of a function with domain R. The Axiom of Choice implies D c C . Using
the notation above, is the least ordinal such that M where D P .R/.
Denition 9.6 (ZF + DCR ). ADC :
(1) Suppose A R. Then A is 1 -borel.
(2) Suppose
< and W
! ! ! ! is a continuous function. Then for each
t
u
A R the set 1 A is determined.
The next theorem shows that assuming ADC C V D L.P .R// the basic analysis
of L.R/ generalizes.
Theorem 9.7 (ZF + DCR ). Assume ADC C V D L.P .R//. Then:
(1) The pointclass 21 has the scale property.
(2) Suppose A R is 21 . Then A D pT for some tree T 2 HOD.
(3) M2 1 L.P .R//.
1
t
u
9 Extensions of L.; R/
612
Remark 9.8. (1) We note that Theorem 9.7(1) follows from Theorem 9.7(2) and
Theorem 9.7(3) just assuming AD C DCR .
(2) Over the base theory of AD C DCR , ADC is equivalent to the assumption that
Theorem 9.7(2) and Theorem 9.7(3) both hold.
t
u
Also Theorem 8.19 generalizes.
Theorem 9.9. Assume ADC C V D L.P .R//. Suppose that x 2 R and let
N D HODL.P .R// x:
Suppose that is an uncountable cardinal of N which is below the least weakly compact cardinal of N .
t
u
Then strong condensation holds for .H. //N in N .
One important feature of ADC is that it is downward absolute.
Theorem 9.10 (ZF + DCR ). Assume ADC and that M is a transitive inner model of
ZF such that R M . Then
M ADC :
Proof. Suppose < M . Then by the Moschovakis Coding Lemma,
P ./ M:
t
u
t
u
9.1 ADC
613
t
u
t
u
t
u
L.R# / ADC :
t
u
Remark 9.16. The consequences of ADC given in Theorem 9.7 are abstractly what
is needed to generalize the analysis of L.R/Pmax to the analysis of L.; R/Pmax where
P .R/ is a pointclass, closed under continuous preimages, such that
L.; R/ ADC :
t
u
t
u
The Suslin cardinals play an important role in descriptive set theory. We note the
following two theorems.
614
9 Extensions of L.; R/
t
u
t
u
The strengthening of Theorem 9.19 over Theorem 9.18 is exactly the difference
between AD and ADC .
Theorem 9.20. The following are equivalent.
(1) ZF C DCR C ADC .
(2) ZF C DCR C AD C The set of Suslin cardinals is closed below .
t
u
Remark 9.21. Assume ZF C DCR C AD. Then it is easily veried that the following
are equivalent,
(1) Every set is Suslin.
(2) The Suslin cardinals are conal in .
Thus the essential content of Theorem 9.20 is simply that if, assuming
ZF C DCR C AD
there is a largest Suslin cardinal below , then ADC .
t
u
t
u
9.1 ADC
615
(1) 0 is the supremum of the ordinals for which there exists map
WR!
which is onto and ordinal denable.
(2) C1 is the supremum of the ordinals for which there exists map
W P . / !
which is onto and ordinal denable.
(3) If is a nonzero limit ordinal then D sup. j < /.
(4) D .
t
u
Within the theory of ADC the ordinal and the sequence h W < i are quite
important. One example is provided by the next theorem.
Theorem 9.24. Assume ADC and that V D L.P .R//. Then ADR holds if and only if
is a limit ordinal and > 0.
t
u
The next theorem, Theorem 9.27, is the original motivation for the denition of
h W < i:
It is due to Solovay. Recall that HODX is the class of sets which are hereditarily ordinal
denable from parameters in X [ X .
It is convenient, but not really necessary, to state Theorem 9.27 using the Wadge
prewellordering on P .R/. The denition of h W < i in the context of AD (as
opposed to ADC ) uses the Wadge prewellordering. This is Solovays original denition
of h W < i.
Denition 9.25. Assume AD.
(1) (Wadge) Suppose that A ! ! and that B ! ! . Then A <w B if:
A and ! ! n A are each continuous preimages of B.
B is not a continuous preimage of A.
(2) (Martin) Suppose A ! ! . The Wadge rank of A, denoted w.A/, is the ordinal
rank of the relation
t
u
.C R j C <w A; <w /:
It is a theorem of Wadge that, assuming AD, A <w B if and only if A is a continuous preimage of B and B is not a continuous preimage of A.
Dene for sets of reals, A and B, A w B, if
A is a continuous preimage of B,
B is a continuous preimage of A.
616
9 Extensions of L.; R/
t
u
Theorem 9.27. Assume ADC and that V D L.P .R//. Let h W < i be the
-sequence of L.P .R//. Suppose that < and let
D A ! ! j w.A/ < :
Then
(1) L. ;R/ D ,
(2) D P .R/ \ HOD .
t
u
617
Remark 9.28. One route to dening strong forms of ADC is through assertions about
. The base theory is
ZF C ADC C V D L.P .R//:
Some examples in increasing (consistency) strength:
D and that is regular.
This is equivalent to the assertion that ADR holds and that is a regular
cardinal.
D and is Mahlo in HOD.
is a limit of regular cardinals such that D .
D and is Mahlo.
D C 1 and is the largest Suslin cardinal.
In this case it is necessarily the case that D .
t
u
By the results of .Sargsyan 2009/ all of these strengthenings are very likely fairly
weak as measured by the large cardinal hierarchy. This is sharp contrast to what was
previously believed.
t
u
9.2
618
9 Extensions of L.; R/
t
u
Theorem 9.30. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC :
Suppose A R and A 2 L.; R/. Then for each n 2 ! there exist a countable
transitive model M and an ordinal 2 M such that the following hold.
(1) M ZFC.
(2) is the nth Woodin cardinal in M .
(3) A \ M 2 M and hV!C1 \ M; A \ M; 2i hV!C1 ; A; 2i.
(4) A \ M is C -weakly homogeneously Suslin in M .
t
u
Theorem 9.31. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC :
Then for each set X R such that
X 2 L.; R/
there is a condition
h.M; I /; ai 2 Pmax
such that
(1) X \ M 2 M,
(2) hH.!1 /M ; X \ Mi hH.!1 /; X i,
(3) .M; I / is X -iterable.
619
t
u
The basic analysis of Pmax now easily generalizes to produce the following theorem.
Theorem 9.32. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC :
Suppose G Pmax is L.; R/-generic. Then in L.; R/G:
(1) P .!1 /G D P .!1 /;
(2) P .!1 / L.R/G;
(3) IG is the nonstationary ideal;
(4) for every set B 2 P .R/ \ L.; R/ the set
X hH.!2 /; B; 2i j MX is B-iterable and X is countable
contains a club, where MX is the transitive collapse of X .
Proof. (2) and (3) are immediate consequences of (1), Theorem 4.50, and the denitions. (1) follows from an analysis of terms using the technical lemma, Lemma 4.46.
The proof of (1) is identical to the proof of Theorem 4.49(1) using Theorem 9.31 to
obtain the necessary conditions in G. By (1) it follows that in L.; R/G,
P .!2 / [L.A; R/G j A 2 :
Thus it sufces to prove that for each A 2 , (4) holds in L.A; R/G.
Fix A 2 . By (2),
L.A; R/G ZFC:
The proof that (4) holds in L.A; R/G is identical to the proof of Lemma 4.52.
t
u
Theorem 9.32 generalizes to the all of the variations of Pmax that we have discussed.
We state the appropriate version for Bmax . We shall consider the Qmax extensions in
Section 9.3.
Theorem 9.33. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC :
Suppose G Bmax is L.; R/-generic. Then in L.; R/G:
(1) P .!1 / L.R/G;
(2) for every set B 2 P .R/ \ L.; R/, the set
X hH.!2 /; B; 2i j MX is B-iterable and X is countable
contains a club, where MX is the transitive collapse of X .
t
u
620
9 Extensions of L.; R/
621
Proof. If
P .R/ \ L.; R/
then by Wadge determinacy,
L.; R/ D L.A; R/
for some A 2 P .R/ \ L.; R/.
In this case L.; R/G ZFC.
Therefore we may suppose that
D P .R/ \ L.; R/;
and so
L.; R/ V D L.P .R// :
For each set A 2 let w.A/ denote the Wadge rank of A. Let denote as
computed in L.; R/.
Suppose R 2 L.; R/G is a binary relation. Let 2 L.; R/ be a term for R.
Fix an ordinal such that 2 L .; R/ and such that
L .; R/ Powerset C 1 -Replacement:
For each < let Z be the set of a 2 L .; R/ such that a is 1 denable in
L .; R/ from .; B/ for some set B 2 with w.B/ < .
Since
L.; R/ ADC + is regular;
there exists < such that
Z \ D
Z G 1 L .; R/G:
Let N be the transitive collapse of Z . Let N be the image of under the collapsing map. Let RN be the interpretation of N . Therefore N G is the transitive collapse
of Z G and RN is the image of R under the transitive collapse of Z G.
Fix A 2 n Z . Thus w.A/ and it follows that N 2 L.A; R/.
By Theorem 9.34(3), has conality !2 in L.A; R/G. It follows that
N G!1 N G
in L.A; R/G.
Let
W N G ! L .; R/G
be the inverse of the collapsing map. is a 1 elementary embedding with
cp./ D D ./N D .!3 /N :
622
9 Extensions of L.; R/
L.A; R/G ZFC and so either for some !1 there is an increasing sequence
ha W < i of elements of RN which is not bounded above or there is an increasing
sequence ha W < !2 i.
In the rst case, ha W < i 2 N G and so .ha W < i/ is a increasing
sequence of elements of R which is not bounded above. In the second case,
h.a / W < !2 i
t
u
is an !2 increasing sequence of elements of R.
The proof of Theorem 9.35 also yields a proof of the following generalization
which we shall require.
Theorem 9.36. Suppose P .R/ is a pointclass closed under continuous preimages, S Ord, and that
L.S; ; R/ ADC C is regular:
Suppose G Pmax is L.S; ; R/-generic. Then
t
u
L.S; ; R/G !2 -DC:
623
Theorem 9.38. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC :
Suppose X Ord, X 2 L.; R/, and that a !1 is a countable set. Then in L.; R/:
HODX;a D HODX a:
t
u
The forcing axiom Martins MaximumCC .c/ is dened in Denition 2.47. It is the
restriction of Martins MaximumCC to partial orders of cardinality c.
Theorem 9.39. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ AD R C is regular:
Suppose G0 Pmax is L.; R/-generic. Suppose
H0 Coll.!3 ; H.!3 //L.;R/G0
is L.; R/G0 -generic. Then
L.; R/G0 H0 ZFC C Martins MaximumCC .c/:
Proof. By Theorem 9.13,
L.; R/ ADC
624
9 Extensions of L.; R/
625
Let
(3.1) P be the set of x such that x codes .p; ; / and
a) p 2 Pmax ,
b) .; / 2 !2 !2 ,
c) p .; / 2 P ;
(3.2) S be the set of x such that x codes .p; ; ; / and
a) .; ; / 2 !1 !1 !2 ,
b) p 2 Pmax ,
c) p .; ; / 2 S ;
(3.3) let D be the set of x such that x codes .p; ; / and
a) .; / 2 !1 !2 ,
b) p 2 Pmax ,
c) p .; / 2 D .
Suppose that M is a transitive model of ZF. Suppose that
z 2 P \ M
and z codes .p; ; /. Then by a simple absoluteness argument M decodes z as a triple
.p; ; / where ; < !2M . If !1M D !1 then D and D .
Similarly suppose that
z 2 S \ M
and z codes .p; ; ; /. Then M decodes z as a 4-tuple .p; ; ; / where < !1M ,
< !1M and < !2M . Again if !1M D !1 then D .
Now suppose M is a transitive model of ZF containing !1 so that
Pmax \ .H.!1 //M D .Pmax /M :
Assume that for all x 2 M \ R, x 2 M . Thus Pmax is nontrivial in M . Suppose that
.P S D / \ M 2 M:
Then P \ M denes in M a term PM for a subset of !2M !2M . Similarly D \ M
denes in M a term for a subset of !1M !2M and S \ M denes in M a term for a
subset of !1M !1M !2M . Let SM be the term given by S \ M and let DM be the
term given by D \ M .
These are terms in the forcing language dened in M for Pmax \ M . If
G Pmax \ M is a lter (not necessarily generic) then PM denes a subset of
!2M !2M . Similarly DM denes from G a subset of !1M !2M and SM denes
from G a subset of !1M !1M !2M . We shall say these are the sets dened by P \ M
and G, dened by D \ M and G, and dened by S \ M and G, respectively.
626
9 Extensions of L.; R/
Let T0 be a tree whose projection is the set of reals which code elements of
P S D :
Let T1 be a tree which projects to the complement of the projection of T0 . We shall use
the following. Suppose M is a transitive model of ZF and that T0 ; T1 2 M . Suppose
j W M ! M is an elementary embedding of M into a transitive model M . Then
the trees T0 and j.T0 / have the same projection in V .
By Theorem 9.37 there exists a set of ordinals S such that:
(4.1) .T0 ; T1 / 2 LS ;
(4.2) if t is a countable sequence of reals then there is a transitive model N such that,
a) N ZFC,
b) LS; t N ,
c) N D LS; t \ V where is the least (strongly) inaccessible cardinal of
LS; t ,
d) there is a countable ordinal which is a Woodin cardinal in N .
Let
be the club measure on P!1 .R/. AD R implies
is a measure. The normality
condition satised by
is the following. Suppose
F W P!1 .R/ ! P!1 .R/
and
j F . / and F . / ; 2
:
Then there exists x 2 R such that j x 2 F . / 2
.
Let S be the ultrapower of S by
, let T0 be the ultrapower of T0 by
and let
T1 be the ultrapower of T1 by
. By the remarks above the trees T0 ; T0 have the same
projection. Further T0 2 LS and so
P ; S ; D 2 L.S ; R/:
Suppose G Pmax is L.S ; R/-generic. Then in L.S ; R/G:
(5.1) ZFC holds;
(5.2) the axiom ./ holds;
(5.3) P denes a partial order on !2 and forcing with this partial order preserves
stationary subsets of !1 ;
(5.4) S denes an !1 sequence of terms for stationary subsets of !1 ;
(5.5) D denes an !1 sequence of dense subsets of the partial order given by P .
The normality condition satised by
shows
Y
L.S; /=
L.S ; R/ D
2P!1 .R/
and that os lemma applies. Thus there is a countable set R such that
D L.S; / \ R
627
and such that if G Pmax \ L.S; / is L.S; /-generic then in L.S; /G:
(6.1) ZFC holds;
(6.2) the axiom ./ holds;
(6.3) P \ L.S; / denes a partial order on !2 and forcing with this partial order
preserves stationary subsets of !1 ;
(6.4) S \ L.S; / denes an !1 sequence of terms for stationary subsets of !1 ;
(6.5) D \ L.S; / denes an !1 sequence of dense subsets of the partial order given
by P .
Fix such a countable set and x G Pmax \ L.S; / that is L.S; /-generic. Let
t 2 L.S; /G be an enumeration of the reals. Thus
L.S; /G D LS; t
and so there is a transitive inner model N and a countable ordinal such that is a
Woodin cardinal in N , LS; t N and N D LS; t \ V where is the least
strongly inaccessible cardinal of LS; t . Fix and N .
Let aG be the subset of the !1 of L.S; /G dened by G. We are using the
notation from Denition 4.44.
Let
PG !2N !2N
be the set in L.S; /G dened by P \ L.S; / and G.
Similarly let
SG !1N !1N !2N
and let
DG !1N !2N
be the sets in L.S; /G dened by G and S , and by G and D .
By the agreement between N and L.S; /G we have,
(7.1) PG is a partial order such that
.INS /N D .INS /N
(7.2) for each < !1N ,
denes a term in N
(7.3) for each < !1N ,
is dense in PG .
PG
\ N;
.; / j .; ; / 2 SG
PG
9 Extensions of L.; R/
628
Further:
(8.1) j.G/ is a lter in Pmax \ N1 ;
(8.2) j.PG / is the set of .; / 2 !2N1 !2N1 such that for some p 2 j.G/,
p .; / 2 P I
(8.3) j.SG / is the set of .; ; / 2 !1N !1N1 !2N1 such that for some p 2 j.G/,
p .; ; / 2 S I
(8.4) j.DG / is the set of .; / 2 !1N1 !2N1 such that for some p 2 j.G/,
p .; / 2 D I
(8.5) j.H / is a lter in j.PG /;
(8.6) for each < !1N1 , < !1N1 j .; ; / 2 j.SG / for some 2 H is a
stationary set in N1 ;
(8.7) for each < !1N1 ,
j.H / \ < !2N1 j .; / 2 DG ;:
629
630
9 Extensions of L.; R/
t
u
Then
L.; R/ AD R :
t
u
Combining the arguments for Theorem 9.39 and for Theorem 7.59 one obtains the
following generalization of Theorem 7.59. This also requires Theorem 9.33.
Recall that BCFACC .c/ denotes the restriction of BCFACC to posets of size c.
Theorem 9.42. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADR C is regular:
Suppose G0 Bmax is L.; R/-generic. Suppose
H0 Coll.!3 ; H.!3 //L.;R/G0
is L.; R/G0 -generic. Then
L.; R/G0 H0 ZFC C BCFACC .c/:
t
u
631
As a corollary to Theorem 9.39 and Theorem 9.42 we obtain the following consistency result.
Theorem 9.43. Assume
ZF C ADR C is regular
is consistent. Then the following are consistent.
(1) ZFC C Martins MaximumCC .c/.
(2) ZFC C Borel Conjecture C BCFACC .c/.
t
u
Another corollary of Theorem 9.39 concerns Martins Maximum and the determinacy of sets of reals which are ordinal denable.
We rst consider the closely related problem of the relationship between Martins
Maximum and quasi-homogeneous ideals.
The proof of Theorem 5.67 easily generalizes to prove the following theorem.
Theorem 9.44. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ AD R C is regular:
Suppose G Pmax is L.; R/-generic. Then
L.; R/G :
t
u
We obtain as a corollary the consistency Martins MaximumCC .c/ with the existence of a quasi-homogeneous saturated ideal.
Corollary 9.45. Assume
ZF C ADR C is regular
is consistent. Then
is consistent.
The generic extension
L.; R/G0 H0
t
u
632
9 Extensions of L.; R/
Remark 9.47. (1) By the previous theorem Martins MaximumCC .c/ does not imply there is a denable wellordering of the reals.
(2) Suppose P is a partial order such that Martins Maximum holds in V P . Suppose forcing with P adds a new subset to !1 . Then P is not homogeneous.
This is implicit in .Foreman, Magidor, and Shelah 1988/. This rules out an obvious approach to producing a model of Martins Maximum with no denable
wellordering of the reals.
(3) Martins Maximum + Conjecture implies there is a denable wellordering of the
reals. This is by the results of .Foreman, Magidor, and Shelah 1988/. In fact,
assuming Martins Maximum, the following are equivalent:
There is a denable wellordering of the reals.
There is a denable partition of
j < !2 and cof./ D !
into innitely many stationary sets.
There is a denable partition of !1 into innitely many stationary sets.
There is a denable !1 -sequence of distinct reals.
(4) Todorcevic has proved that assuming the Proper Forcing Axiom then the following are equivalent:
There is a denable wellordering of the reals.
There is a denable increasing sequence
hf W < !2 i
!
in the partial order, .! ; <F /. The order is the pointwise order modulo
nite sets.
633
(5) If the conjecture fails in V then every (uncountable) regular cardinal is measurable in HOD. Thus the consistency strength of the failure of the conjecture is
very likely beyond that of the existence of a supercompact cardinal.
(6) It looks even harder to obtain the failure of the conjecture in the presence of a
supercompact cardinal. Therefore modulo nding a new consistency proof for
Martins Maximum, the problem of nding a model of Martins Maximum in
which the conjecture fails looks quite hard.
t
u
9.3
t
u
Theorem 9.49. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC :
Suppose A R and A 2 L.; R/. Then there is a condition .hMk W
k < !i; f / 2 Qmax such that
(1) A \ M0 2 M0 ,
(2) hH.!1 /M0 ; A \ M0 i hH.!1 /; Ai,
(3) hMk W k < !i is A-iterable.
Proof. This is an immediate by Theorem 6.64 and Theorem 9.48
t
u
634
9 Extensions of L.; R/
Using Theorem 9.49, the analysis of L.R/Qmax easily generalizes to the case of
L.; R/Qmax where is a pointclass closed under continuous preimages such that
L.; R/ ADC :
Theorem 9.50. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC C is regular:
Suppose G Qmax is L.; R/-generic. Then
L.; R/G !2 -DC
and in L.; R/G:
(1) P .!1 / L.R/G;
(2) for every set B 2 P .R/ \ L.; R/, the set
X hH.!2 /; B; 2i j MX is B-iterable and X is countable
contains a club, where MX is the transitive collapse of X .
t
u
Because of the equivalence of Qmax and Qmax in L.R/ as forcing notions, Theorem 9.50 immediately gives the following version for Qmax -extensions.
Theorem 9.51. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC C is regular:
Suppose G Qmax is L.; R/-generic. Then
L.; R/G !2 -DC
and in L.; R/G:
(1) P .!1 / L.R/G;
(2) for every set B 2 P .R/ \ L.; R/ the set
X hH.!2 /; B; 2i j MX is B-iterable and X is countable
contains a club, where MX is the transitive collapse of X .
t
u
635
and that
L .A; R/ 1 L.A; R/:
Suppose
X L .A; R/G
is a countable elementary substructure with G 2 X . Let MX be the transitive collapse
of Y and let
IX D .INS /MX :
Then for each B R such that B 2 X \ L.A; R/, .MX ; IX / is B-iterable.
Proof. The proof is essentially the same as that of Theorem 6.78, using Theorem 9.51
in place of Theorem 6.77.
t
u
Another corollary of Theorem 9.51 generalizes Theorem 6.81.
Theorem 9.53. Suppose A R and that
L.A; R/ ADC :
Suppose G Qmax is L.A; R/-generic. Then in L.A; R/G the following hold.
(1) L.A; R/ ZFC.
(2) IG D INS and IG is is an !1 -dense ideal.
(3) Suppose S !1 is stationary and
f W S ! Ord:
Then there exists g 2 L.A; R/ such that 2 S j f ./ D g./ is stationary.
Proof. By Theorem 9.51,
.P .!1 //L.A;R/G L.R/G:
Therefore by Theorem 6.81,
L.A; R/G AC ;
IG D INS , and IG is a normal !1 -dense ideal in L.A; R/G.
This proves (1) and (2).
(3) follows from Theorem 9.51 and Theorem 3.42, by reducing to the case that the
t
u
range of f is bounded in L.A;R/ , cf. the proof of Lemma 6.79(3).
Suppose that P .R/ is a pointclass closed under continuous preimages such
that
636
9 Extensions of L.; R/
t
u
Remark 9.55. As formulated, FA..!1<! //c implies .!1<! /. In fact it implies that
for each stationary set S !1 there exists a function
f W !1 ! H.!1 /
which witnesses that .!1<! / holds such that
f ./ D ;
for all 2 !1 n S . This slight strengthening of .!1<! / is easily seen to follow from
t
u
+ .!1<! /.
9.4
637
Changs Conjecture
There is a curious metamathematical possibility. Perhaps there is an interesting combinatorial statement whose truth in
L.R/Pmax
cannot be proved just assuming
L.R/ AD;
but can be proved from a stronger hypothesis.
We recall the statement of Changs Conjecture.
Denition 9.56. Changs Conjecture: The set
X !2 j ordertype.X / D !1
t
u
is stationary in P .!2 /.
D. Seabold has proved the following theorem.
t
u
t
u
638
9 Extensions of L.; R/
t
u
Our goal in this section is to sketch the proof of the generalization of Theorem 9.58
to the Qmax -extension;
Theorem 9.60. Suppose
L.R/ AD
and that there exists a countable set R such that
HODL.R/ . / \ R D
and
HODL.R/ . / AD C DC:
Suppose G Qmax is L.R/-generic. Then
L.R/G Changs Conjecture:
t
u
is open.
t
u
Steel has generalized the analysis of scales in L.R/ to iterable MitchellSteel modQ i. e. MitchellSteel models relativized to R. With this machinels of the form L.R; E/;
ery the method of the core model induction used to prove Theorem 6.149 on page 425,
generalizes to prove the following theorem.
639
t
u
t
u
640
9 Extensions of L.; R/
<!
, is a formula and
containing ZFC .
By Lemma 9.52 there exists a countable transitive set N and a lter H QN
max
such that
(1.1) N T,
(1.2) p0 ; x0 ; x1 N ,
(1.3) p0 2 H ,
(1.4) H is N -generic,
(1.5) B \ N 2 N and hH.!1 /N ; B \ N; 2i hH.!1 /; B; 2i,
(1.6) .N H ; .INS /N H / is B-iterable.
Thus
N H
/; f i 2 Qmax
h.N H ; INS
and
N H
/; f i < p0
h.N H ; INS
where
641
such that j .f / D fG .
N H
Since h.N H ; INS
/; f i 2 G, H G.
Let
X L .A; R/G
be the set of b 2 L .A; R/G such that b is denable in L .A; R/G from parameters
in R \ N [ fG .
Thus X L .A; R/G. Further a 2 X since x0 2 N .
The key points are that
H G
and
hH.!1 /N ; B \ N; 2i hH.!1 /; B; 2i;
for these imply that
X \ L .A; R/ D R \ N :
Let MX be the transitive collapse of X , let FX be the image of fG under the collapsing map and let IX D .INS /MX . Thus
(2.1) .H.!2 //MX D H.!2 /N H ,
(2.2) IX D .INS /N H ,
(2.3) f D fX .
Therefore by Theorem 9.52, .MX ; IX / is iterable, and so
h.MX ; IX /; fX i 2 Qmax :
However,
N H
/; f i 2 G:
h.N H ; INS
N H
/; f i 2 G induces an iteration witTherefore the iteration witnessing h.N H ; INS
nessing
h.MX ; IX /; fX i 2 G
642
9 Extensions of L.; R/
L .A; R/ ZFC
643
(1.1) h.MX ; IX /; fX i 2 G;
(1.2) for all 2 C , fG ./ is L.MX ; fG j/-generic for Coll.!; /;
where MX is the transitive collapse of X , fX is the image of fG under the collapsing
map,
IX D .INS /MX ;
and where C is the critical sequence of the iteration
j W .MX ; IX / ! .MX ; IX /
such that j.fX / D fG .
Let
hM ; G ; j; W < !1 i
be the iteration of .MX ; IX / such that j0;!1 .fX / D fG .
Dene a sequence hX W !1 i of countable elementary substructures by induction on such that
(2.1) X0 D X ,
(2.2) XC1 D f .X \ !1 / j f 2 X ,
(2.3) if !1 is a limit ordinal then
X D [X j < :
For each !1 let M be the transitive collapse of X , let
W M ! X
be the inverse of the collapsing map, and let
j;
W M ! M
Thus
hM ; G ; j;
W < !1 i
is an iteration of .MX ; IX / such that j0;!
.fX / D fG .
1
Therefore
hM ; G ; j; W < !1 i D hM ; G ; j;
W < !1 i:
9 Extensions of L.; R/
644
NQ D X \ N
and let N be the transitive collapse of NQ . Thus N 2 M and
NQ D .N /:
Further for each !1 ,
N D j0; .N0 /:
M
MC1
!1
M
D !2
D !2 :
Let
h W !1 ! N
be such that for all limit ordinals < !1 ,
h D NQ :
t
u
HODL.R/ . / ADC :
t
u
645
!2<! ,
If Changs Conjecture fails in L.R/G then there is a function F and a corresponding term such that F is a counterexample and such that A
is 21 in L.R/.
This follows by the usual reection arguments and the fact that the pointclass
.21 /L.R/ has the scale property, Theorem 2.3.
Again by the scale property of .21 /L.R/ there must exist a condition
p0 2 Qmax \ HODL.R/
such that p0 forces that is a term for a counterexample to Changs Conjecture.
Fix a countable set R such that
HODL.R/ . / \ R D
9 Extensions of L.; R/
646
and
HODL.R/ . / ZF C AD C DC:
By Lemma 9.67, we can suppose that
HODL.R/ . / ADC :
Let S Ord be a set such that
LS D HODL.R/ :
It is easy to see that such a set S exists, essentially by Vopenkas argument. In fact one
can choose S to be a subset of L.R/ .
Let
N D L.S; /
and let
D P . / \ N:
Thus
and
D R \ L.; /:
Let g0 Qmax \ L.; / be L.; /-generic such that p0 2 g0 and let
f0 D [f j h.M; I /; f i 2 g0 :
Thus g0 is N -generic and
P . / \ N g0 D P . / \ L.; /g0 :
Therefore by Theorem 6.81 and Theorem 9.51, the following hold in N g0 .
(1.1) AC .
(1.2) The nonstationary ideal on !1 is !1 -dense.
(1.3) f0 witnesses ++ .!1<! /.
(1.4) For each p 2 Coll.!; !1 /, the set
< !1 j p 2 f0 ./
is stationary.
By modifying f0 if necessary we can suppose that for all < !1N g0 ,
f0 ./ Coll.!; /:
By (1.1) and since N0 D L.S; /,
N g0 ZFC
and further for some a0 !3N g0 ,
L.; /g0 D La0 :
647
By Theorem 5.35,
L.R/
HODa
D HODL.R/ a0 D LS; a0 D N g0 :
0
!2LS;a0 ;x
is a Woodin cardinal in
0 ;x
:
HODLS;a
S;a0
However for each < !1 , there exists x1 2 R such that for all x 2 R, if
x1 2 LS; a0 ; x then
0 ;x
HODL.R/
LS; a0 ;
P ./ \ HODLS;a
S;a0
S;a0
since HODL.R/ D LS .
Therefore there exist a transitive inner model M , containing the ordinals, and
0 < !1 such that
M ZFC;
S; a0 M ,
P .!3N g0 / \ N g0 D P .!3N g0 / \ M;
L.R/G D L.R/G:
9 Extensions of L.; R/
648
h.M0 ; I0 /; f0 i 2 G :
j W M1 ! M1 M1 h0
h.M0 ; I0 /; f0 i 2 G ;
h0 Coll.!; !1M / and h0 is M0 -generic.
Therefore there exists an iteration
h.M ; J /; G ; j; W < < !1 i
such that for all < !1 ,
M
(3.1) !2
2 G ,
(3.2) h0 is M C1 -generic,
649
where D !2
and
j W M C1 ! MC1 M C1 h0
M C1
, and h0 .
We note that by (3.1), for all < < !1 , the critical point of j; is !2M and so
for all < !1 , j0; .f0 / D f0 .
Let
j0 W .M0 ; I0 / ! .MQ 0 ; IQ0 /
be the limit embedding of the iteration.
Thus
Q
!2M0 D !1
since
j0 .!2M0 / D !1 :
M0
, and h0 .
be the generic elementary embedding corresponding to f0 ; INS
<!
+
Thus by (3.3) and since fG witnesses .!1 /,
F W !2<! ! !2
650
9 Extensions of L.; R/
MQ 0 D j.M0 / D Vj. / \ MQ :
651
9.5
t
u
652
9 Extensions of L.; R/
Denition 9.71.
t
u
t
u
t
u
Remark 9.74. (1) It is not difcult to show that WRP.!2 / is consistent with
2@1 > @2 and WRP.!2 / is consistent with CH. Thus Lemma 9.73(1) cannot
really be improved.
(2) We shall prove that SRP.!2 / implies 12 D !2 and so SRP.!2 / implies c D @2 ,
see Theorem 9.79.
(3) In fact, SRP.!2 / implies AC , as we shall note below. This gives a different
t
u
proof that SRP.!2 / implies c D @2 .
At the heart of Theorem 9.69 is the following theorem from which one can obtain
Theorem 9.69 as a corollary.
653
Theorem 9.75. Assume WRP.2/ .!2 / and that for each set A !1 , A# exists. Then
for each set A !2 , A# exists.
Proof. Fix a set A !2 . We must prove that A# exists. Clearly we may suppose that
A is conal in !2 .
For each countable set !2 let be the ordertype of and let
A
be the image of A under the transitive collapse of . Let
W !
be the collapsing map.
For each i < ! let i D !2Ci . Thus for each bounded set b !2 , i is a Silver
indiscernible of Lb.
For each formula .x0 ; y0 ; z0 / and for each pair
.s; t / 2 !2 <! i j i < !<!
of nite sets, let S.;s;t/ be the set of 2 P!1 .!2 / such that
LA A ; s; t :
Thus
P!1 .!2 / D S.;s;t/ [ S.:;s;t/ :
Fix 0 , s0 and t0 . We claim that not both S.0 ;s0 ;t0 / and S.:0 ;s0 ;t0 / are stationary in P!1 .!2 /. Assume toward a contradiction that each are stationary. Then by
WRP.2/ .!2 / there exists < !2 such that
(1.1) A \ is conal in ,
(1.2) s0 ,
(1.3) S.0 ;s0 ;t0 / \ P!1 ./ is stationary in P!1 ./,
(1.4) S.:0 ;s0 ;t0 / \ P!1 ./ is stationary in P!1 ./.
The key point is that .A \ /# exists. Therefore if
X H.!2 /
is a countable elementary substructure with A \ 2 X then .A \ /# 2 X . Let MX
be the transitive collapse of X . Thus the image of A \ under the collapsing map is
exactly A where D X \ . But the image of .A \ /# under the collapsing map is
.A /# . Thus if s0 2 X then
X \ 2 S.;s0 ;t0 /
if and only if
LA \ A \ ; s0 ; t0 :
Therefore if
LA \ A \ ; s0 ; t0
654
9 Extensions of L.; R/
t
u
One corollary of Lemma 9.76 and the proof of Theorem 5.13 is that SRP.!2 /
implies AC . This result, obtained independently by P. Larson, gives yet another proof
that SRP.!2 / implies c D @2 .
Corollary 9.77 (SRP.!2 /).
AC
holds.
t
u
655
t
u
t
u
656
9 Extensions of L.; R/
An easier version of Theorem 9.80 is the following theorem. Recall that PD is the
assertion that all projective sets are determined.
Theorem 9.81. Suppose that WRP.2/ .!2 / holds and that if g Coll.!; !1 / is
V -generic then
V g PD:
Suppose that
G Coll.!; !2 /
is V -generic. Then
V G PD:
t
u
The method of proving Theorem 9.81 amplied by some of the machinery behind
the proof of Theorem 5.104 yields the following improvements of Theorem 9.81.
Theorem 9.82. Suppose that WRP.2/ .!2 / holds and that INS is !2 -saturated. Suppose
that G Coll.!; !2 / is V -generic. Then
V G PD:
t
u
Corollary 9.83. Suppose that SRP.!2 / holds and that P is a partial order of cardinality !2 . Suppose G P is V -generic. Then
V G PD:
t
u
Remark 9.84. Theorem 9.82, and therefore Corollary 9.83, can be be strengthened
to obtain more determinacy. The main results of Steel and Zoble .2008/ improve the
results by obtaining ADL.R/ .
The proof of Theorem 9.82 can be implemented using a weakened version of
SRP.!2 /, see Theorem 9.95. This version is dened in Denition 9.88(2). Theorem 9.99 shows that this weakened version together with the assertion that INS is !2 t
u
saturated cannot imply signicantly determinacy signicantly past ADL.R/ .
WRP.!2 / implies a weak variation of Changs Conjecture.
Lemma 9.85 (WRP.!2 /). Suppose that
F W !2<! ! !2
and let CF be the set of 2 P!1 .!2 / such that there exists 2 P!1 .!2 / satisfying;
(i) and \ !1 D \ !1 ,
(ii) n ;,
(iii) F <! .
Then CF contains a closed unbounded subset of P!1 .!2 /.
657
Proof. Let
S D P!1 .!2 / n CF :
Assume toward a contradiction that S is stationary in P!1 .!2 /. Thus by WRP.!2 /
there exists !1 < < !2 such that S \ P!1 ./ is stationary in P!1 ./.
Let
Z H.!3 /
be a countable elementary substructure such that
(1.1) F 2 Z,
(1.2) 2 Z,
(1.3) Z \ 2 S .
The requirement (1.3) is easily arranged since S \ P!1 ./ is stationary in P!1 ./. By
(1.2)
Z \ Z \ !2 :
But Z \ !2 is closed under F and so Z \ !2 witnesses that Z \ 2 CF which
contradicts that Z \ 2 S .
t
u
An immediate corollary of the next lemma is that WRP.!2 / must fail in
L.A; R/Pmax where A R is such that
L.A; R/ ADC :
Lemma 9.86. Suppose that
V D LA
for some set A !2 and that for each set B !1 , B # exists. Then WRP.!2 / fails.
Proof. Consider the structure
h!2 ; A; 2i:
For each countable elementary substructure
X h!2 ; A; 2i
let AX be the image of A under the transitive collapse.
Let S be the set of X 2 P!1 .!2 / such that !1 \ X is countable in LAX . We
claim that that S is stationary in P!1 .!2 /. If not let Y0 be the set of a 2 H.!3 / such
that a is denable in the structure
hH.!3 /; A; 2i:
Thus Y0 H.!3 / and so S 2 Y0 . Since S is not stationary it follows that
Y0 \ !2 S:
Let M0 be the transitive collapse of Y0 and let A0 be the image of A under the transitive
collapse. Thus every element of M0 is denable in the structure
hM0 ; A0 ; 2i:
658
9 Extensions of L.; R/
However
M0 V D LA0
and so M0 2 LA0 . Therefore M0 is countable in LA0 and so Y0 \ !2 2 S ,
a contradiction.
Thus S is stationary in P!1 .!2 /.
We now assume toward a contradiction that WRP.!2 / holds.
Fix !1 < < !2 such that S \ P!1 ./ is stationary in P!1 ./.
Thus there exists
Z H.!3 /
such that
(1.1) Z \ 2 S ,
(1.2) A \ 2 Z.
However .A \ /# exists and so .A \ /# 2 Z.
Let Z0 D Z \ . Let MZ be the transitive collapse of Z and let AZ be the image
of A under the transitive collapse. Let AZ0 be the image of A \ under the transitive
collapse of Z0 . Trivially AZ0 is the image of A \ under the transitive collapse of Z.
However
Z \ !1 D Z0 \ !1
and so since Z0 2 S ,
V!C1 \ LAZ0 6 MZ :
which is a contradiction.
Then
L.; R/ AD R :
659
660
9 Extensions of L.; R/
661
Fix a bijection
W !2 ! M0 G
with 2 L0 .0 ; R/G. This exists since
jM0 Gj D !2
in L0 .0 ; R/G.
Therefore, by WRP.!2 /, there exists < !2 such that
(5.1) M0 G,
(5.2) !1 ,
(5.3) X 2 P!1 ./ j X 2 S is stationary in P!1 ./.
The key point is that
H.!2 /L.;R/G D H.!2 /L0 .0 ;R/G
and so in L.; R/G, the set X 2 P!1 ./ j X 2 S is stationary in P!1 ./.
Let B 2 0 be the set of x 2 R such that x codes a triple .; a; b/ such that
(6.1) < !1 ,
(6.2) a ,
(6.3) b D P ./ \ LT; a.
The set B is Suslin in L.; R/. Let TB 2 M be a tree such that
B D pTB :
Let
TB D f W !1 ! TB j f 2 M =
TB D j.TB /
where
j W L.; R/G ! N L.; R/Gg
is the associated generic elementary embedding.
Since in L.; R/G, the set X 2 P!1 ./ j X 2 S is stationary in P!1 ./,
there exists
Z hM G; ; 2i
such that
(7.1) T; TB ; 0 ; ; 2 Z,
(7.2) Z \ 2 S .
662
9 Extensions of L.; R/
Let M0 be the transitive collapse of Z and let N0 be the transitive collapse of Z \ .
By (5.2),
.!1 /M0 D .!1 /N0 D Z \ !1 :
Since TB 2 Z it follows that
P .N0 / \ LT; N0 M0 :
However Z \ 2 S and so this implies that Z \ !1 is countable in M0 , a contradiction.
t
u
There are natural weakenings of the principles, WRP.!2 / and SRP.!2 /. We discuss these briey and state some theorems. Our purpose is to illustrate how possibly
subtle variations are stratied, in the context of Pmax -extensions, by the strength of the
underlying model of ADC .
Suppose that
I P .P!1 .!2 //
is an ideal. Recall that the ideal I is normal if for all functions
F W !2 ! I;
SF 2 I where
SF D 2 P!1 .!2 / j 2 F ./ for some 2 :
The ideal is ne if for each 2 P!1 .!2 /,
2 P!1 .!2 / j 6 2 I:
(1) WRP .!2 /: There is a proper normal, ne, ideal
I P .P!1 .!2 //
such that for all T 2 P .!1 / n INS ,
X 2 P!1 .!2 / j X \ !1 2 T I
and such that if
S P!1 .!2 /
Denition 9.88.
t
u
663
Remark 9.89. WRP .!2 / simply asserts that the set of counterexamples to WRP.!2 /
t
u
generates a normal, ne, ideal which is proper on each stationary subset of !1 .
One connection between these weakened versions is given in the following lemma.
Lemma 9.90. Assume that INS is !2 -saturated and that SRP .!2 / holds. Let
I P .P!1 .!2 //
be a normal ideal witnessing that SRP .!2 / holds. Suppose that
S1 P!1 .!2 /
and
S2 P!1 .!2 /
are each I -positive. Then there exists !1 < < !2 such that S1 \ P!1 ./ and
S2 \ P!1 ./ are each stationary in P!1 ./.
Proof. Let J1 P .!1 / be the set of A !1 such that
X 2 S1 j X \ !1 2 A 2 I:
It is easily veried that J1 is a normal (uniform) ideal and so since INS is !2 -saturated,
there exists A1 2 P .!1 / n INS such that
J1 D A !1 j A \ A1 2 INS :
Similarly there exists A2 2 P .!1 / n INS such that
J2 D A !1 j A \ A2 2 INS ;
where J2 is the set of A !1 such that
X 2 S2 j X \ !1 2 A 2 I:
Choose stationary sets B1 A1 and B2 A2 such that B1 \ B2 D ;. Dene
S P!1 .!2 / to be the set of X such that;
(1.1) X 2 S1 if X \ !1 2 B1 ,
(1.2) X 2 S2 if X \ !1 2 B2 .
It follows that for each stationary set T !1 ,
X 2 S j X \ !1 2 T I:
Thus since I witnesses SRP .!2 / there exists !1 < < !2 such that
S \ P!1 ./
is closed, unbounded, in P!1 ./. This implies that both S1 \ P!1 ./ and S2 \ P!1 ./
t
u
are stationary in P!1 ./.
The following lemmas show that while WRP .!2 / is a signicant weakening of
WRP.!2 /, it is plausible that SRP .!2 / is not as signicant a weakening of SRP.!2 /.
9 Extensions of L.; R/
664
Lemma 9.91 (2@1 D @2 ). Assume WRP.!2 / and suppose that A !2 is a set such
that
H.!2 / LA:
Then
Proof. Let I be the normal ideal dened in LA, generated by sets S P!1 .!2 / such
that
(1.1) S 2 LA,
(1.2) for all !1 < < !2 , S \ P!1 ./ is not stationary in P!1 ./.
Since WRP.!2 / holds in V , I is contained in the ideal of nonstationary subsets
of P!1 .!2 /. Therefore I is a proper ideal in LA and so I witnesses WRP .!2 / in
LA.
t
u
The proof of Theorem 9.75 easily adapts, using Lemma 9.90 in place of
WRP2 .!2 /, to prove the following variation of Theorem 9.75.
Lemma 9.92. Assume SRP .!2 / and that INS is !2 -saturated. Then for each A !2 ,
t
u
A# exists.
As an immediate corollary we obtain a weak version of Theorem 9.87.
Corollary 9.93. Suppose that P .R/ is a pointclass, closed under continuous
preimages, such that
L.; R/ ADC C is regular
and such that
t
u
The situation for WRP .!2 / seems analogous to that for Changs Conjecture.
Theorem 9.94. Suppose
L.R/ AD
and that there exists a countable set R such that
HODL.R/ . / \ R D
and
HODL.R/ . / AD C DC:
Suppose G Pmax is L.R/-generic. Then
L.R/G WRP .!2 /:
t
u
665
t
u
Some information about SRP .!2 / is provided by Theorem 9.99. This theorem
places an upper bound on the consistency strength of the theory
ZFC C SRP .!2 / C INS is !2 -saturated
which is not far beyond the lower bound established by Theorem 9.95, and signicantly
below the known upper bounds for SRP.!2 /.
Theorem 9.99 involves the following determinacy hypothesis:
(ZFC) Let F be the club lter on P!1 .R/. Then
(1) F jL.R; F / is an ultralter,
(2) L.R; F / ADC .
We note the following corollary of Theorem 9.14.
Theorem 9.96. Let F be the club lter on P!1 .R/. Suppose that
L.R; F / AD:
C
t
u
Then L.R; F / AD .
The proof of Theorem 9.99 is relatively straightforward using the following theorem.
Theorem 9.97. Let F be the club lter on P!1 .R/. Suppose that
F jL.R; F /
is an ultralter and that
L.R; F / AD:
Let D
2 L.R;F /
.
.
1 /
Then:
t
u
666
9 Extensions of L.; R/
t
u
We conjecture that Theorem 9.87 holds for SRP .!2 /. This conjecture is not refuted by Theorem 9.99. The explanation lies in the subtle, but important, distinction
between models of ADC of the form L.; R/ versus models of the form L.S; ; R/
where S is a set of ordinals and is a pointclass (closed under continuous preimages).
We discuss below an example which illustrates this point.
Let F be the club lter on P!1 .R/. Suppose that, as in Theorem 9.99,
F jL.R; F /
is an ultralter and that
L.R; F / AD:
Thus
L.R; F / There is a normal ne measure on P!1 .R/:
The basic theory of ADC applied to L.R; F / shows that
L.R; F / D HODL.R;F / .R/:
Thus L.R; F / is of the form L.S; ; R/ with D ;.
However the basic theory of ADC also yields the following theorem.
Theorem 9.100. Suppose that P .R/ is a pointclass, closed under continuous
preimages, such that
L.; R/ ADC
and that
L.; R/ There is a normal ne measure on P!1 .R/:
Then
L.; R/ ADR :
t
u
Thus to obtain a model of ADC in which there is a normal ne measure on P!1 .R/,
the distinction between models of the form L.; R/ and of the form L.S; ; R/ is an
important one. We conjecture that the situation is similar for SRP .!2 /. Of course
for the other principles (SRP.!2 /, WRP.!2 /, WRP.2/ .!2 /, and WRP .!2 /) the distinction is not important. The reason is simply that these other principles are absolute
between V and L.P .!2 //.
9.6
667
t
u
Remark 9.102. In general we shall only consider Strong Changs Conjecture in the
situation that
L.P .!2 // !2 -DC:
u
t
Lemma 9.103 (ZFC). The following are equivalent:
(1) Strong Changs Conjecture.
(2) There exists a function
F W H.!3 / ! H.!3 /
such that if X H.!3 / is countable and closed under F , then there exists
Y H.!3 / such that
668
9 Extensions of L.; R/
a) F Y Y ,
b) X Y ,
c) X \ !1 D Y \ !1 ,
d) X \ !2 Y \ !2 .
(3) There exists a function
F W H.!3 / ! H.!3 /
such that if X H.!3 / is countable and closed under F , then there exists
Y H.!3 / such that
a) X Y ,
b) X \ !1 D Y \ !1 ,
c) X \ !2 Y \ !2 .
(4) There exists a transitive inner model N such that
a) P .!2 / N ,
b) N ZF C DC C Strong Changs Conjecture.
Proof. It is straightforward to show that (1) implies (2), the relevant observation is the
following. Suppose that M is a transitive set such that
M H.!3 / M:
Then there exist a countable elementary substructure
X0 M
and a function
F0 W H.!3 / ! H.!3 /
such that the following holds. Suppose that X H.!3 / is countable,
X0 \ H.!3 / X;
and X is closed under F0 . Then there exists
Y M
such that X0 Y and Y \ H.!3 / D X .
Thus it sufces to prove that (4) implies (3) and that (3) implies (1).
We rst prove that (3) implies (1), noting that for this implication one only needs
!2 -DC.
Let M be a transitive set such that
M H.!3 / M
and let X M be a countable elementary substructure. Since H.!3 / is denable in
M , there exists
F W H.!3 / ! H.!3 /
669
such that F 2 X and such that F witnesses (3). Let Y H.!3 / be a countable
elementary substructure closed under F such that
(1.1) X Y ,
(1.2) X \ !1 D Y \ !1 ,
(1.3) X \ !2 Y \ !2 ,
and let
Z D f .a/ j f W !2 ! M; f 2 X; and a 2 Y \ !2 <! :
Since M !2 M (and since !2 -DC holds),
(2.1) Z M ,
(2.2) Z \ !2 D Y \ !2 .
We nish by proving that (4) implies (3). Fix a transitive inner model N such that
(3.1) P .!2 / N ,
(3.2) N ZF C DC C Strong Changs Conjecture.
Let be a strong limit cardinal with
cof. / > jH.!3 /j;
and let M D N . Thus
M H.!3 / M
in N . Let
F W H.!3 / ! H.!3 /
be a function (in V ) such that if X H.!3 / is a countable set closed under F then
there exists
X M
such that X \ H.!3 / D X . We claim that F witnesses (3). Assume toward a
contradiction that this fails and let X H.!3 / be a countable set, closed under F ,
which witnesses that F fails to satisfy (3). However X 2 N (since H.!3 / N ) and so
by absoluteness and the choice of F , there exists a countable elementary substructure
X M
such that X 2 N and X \ H.!3 / D X . Therefore since
N Strong Changs Conjecture;
there exists a countable elementary substructure
Y M
such that
670
9 Extensions of L.; R/
(4.1) X Y ,
(4.2) X \ !1 D Y \ !1 ,
(4.3) X \ !2 Y \ !2 .
Finally Y \ H.!3 / contradicts the choice of X .
t
u
XZ0 \ !2 D X \ !2
GX <! X:
t
u
671
The primary goal of this section is to sketch the construction of a model in which
./ holds and in which Strong Changs Conjecture holds. This improvement of Theorem 9.57 will require an even stronger determinacy hypothesis. The formulation involves the sequence h W < i which is discussed at the end of Section 9.1.
The proof of Theorem 9.114 requires the following theorems concerning models
of ADC .
Theorem 9.105. Suppose that P .R/ is a pointclass closed under continuous
preimages such that
L.; R/ ADC :
Let h W < i be the -sequence of L.; R/. Suppose that either
(i) is a limit ordinal, or
(ii) if D C 1 then < where
D max < j is a Suslin cardinal in L.; R/:
Then there is a surjection
W ! \ V ! P .R/ \ L.; R/
such that is 1 -denable in L.; R/ from R.
Remark 9.106.
t
u
672
9 Extensions of L.; R/
Theorem 9.107. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC :
Let D ./L.;R/ and suppose that
A ! \ V
is ordinal denable in L.; R/. Then there exist a formula .x; y/ and a set
b 2 P ./ \ HOD such that for all
a 2 ! \ V ;
a 2 A if and only if
t
u
Theorem 9.107 easily yields the following corollary which is what we shall require.
Corollary 9.108. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC :
Let D ./L.;R/ and suppose that
a 2 P!1 .! / \ V :
Suppose that
P 2 HODL.;R/ .a/
is a partial order which is countable in V and that X is a comeager set of lters in P
such that X is ordinal denable in L.; R/ with parameters from a [ a. Suppose
that g P is a lter which is HODL.;R/ .a/-generic. Then g 2 X .
Proof. Fix 2 Ord such that a 2 V , jV j D , and such that X is denable in
L .; R/ with parameters from a [ a.
Let Y be the set of all nite sequences
ha0 ; b0 ; P0 ; 0 ; g0 i
such that:
(1.1) a0 2 P!1 .! / \ V .
(1.2) P0 is a partial order.
(1.3) P0 2 H.!1 / \ HODL.;R/ .a0 /.
(1.4) b0 2 a0<! .
(1.5) Let
X0 D g j L .; R/ 0 a0 ; b0 ; g:
Then g0 2 X0 and X0 is a comeager set of lters in P0 .
673
<!
t
u
The next theorem which we shall require generalizes Theorem 9.7. Recall that if
P .R/ is a pointclass closed under continuous images, continuous preimages, and
complements, then we have associated to a transitive set M constructed from those
sets, X , which are coded by an element of , see Denition 2.18.
674
9 Extensions of L.; R/
Theorem 9.109. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ AD R C is regular:
Let h W < i be the -sequence of L.; R/. For each < let
D sup < j is a Suslin cardinal
and let D A ! ! j w.A/ < . Then M 1 L.; R/.
t
u
Remark 9.110. By Theorem 9.19, the Suslin cardinals are closed below . Thus the
essential content of Theorem 9.109 is in the case that
< :
This is the case that is the largest Suslin cardinal below . For example if D 0
21 .
u
t
then D
We shall also need the following theorem concerning generic elementary embeddings. For this theorem it is useful to dene in the context of DC, a partial embedding,
jU , for each countably complete ultralter U .
Denition 9.111 (DC). Suppose that X ; and that U P .X / is a countably
complete ultralter. Let
jU W [LS j S Ord ! V
be dened as follows: Suppose that S Ord. Then
[jU .a/ j a 2 LS
is the transitive collapse of the ultrapower,
f W X ! LS j f 2 V =U;
and jU jLS W LS ! LjU .S / is the associated (elementary) embedding.
t
u
675
The last of the theorems which we shall need is in essence a corollary of Theorem 5.34.
Theorem 9.113. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ AD R C is regular:
Let h W < i be the -sequence of L.; R/. Then for each < ,
HODL.;R/ C1 is a Woodin cardinal:
t
u
676
9 Extensions of L.; R/
Proof. By Lemma 9.104, (2) is an immediate corollary of (1). We prove (1). Fix
D ./L.;R/ :
Let G1 Coll.!; R/ be L.; R/-generic and let
j1 W L.; R/ ! L. 1 ; R1 / L.; R/G1
be the associated embedding as given by Theorem 9.112. Thus for each set S Ord
with S 2 L.; R/,
j1 jLS D j jLS
where
2 L.; R/ is the measure on P!1 .R/ generated by the closed unbounded
subsets of P!1 .R/.
It is convenient to work in L.; R/G1 .
Fix G0 Pmax such that G0 is L.; R/-generic and such that G0 2 L.; R/G1 .
We begin by observing that a very weak version of Changs ConjectureC does hold.
Suppose that
F W !2<! ! !2
is a function in L.; R/G0 . Let CF be the set of 0 2 P!1 .!2 / such that there exists
1 2 P!1 .!2 / satisfying;
(1.1) 0 1 ,
(1.2) 0 \ !1 D 1 \ !1 ,
(1.3) F 1<! 1 .
Then in L.; R/G0 the set
P!1 .!2 / n CF
is not stationary in P!1 .!2 /; i. e. the set CF contains a closed unbounded subset.
This is an immediate corollary of Lemma 9.85 since by Theorem 9.39 and
Lemma 9.72,
L.; R/G0 WRP.!2 /:
In fact, there is a straightforward and more direct proof of this last claim which does
not require Theorem 9.39. We leave this as an exercise for the interested reader. The
alternate approach makes the generalization of this theorem to the Pmax variations we
have considered essentially routine.
Fix a surjection
W R ! H.!2 / \ L.R/
which is denable in the structure hH.!2 / \ L.R/; 2i; i. e. that is simple.
Suppose that 2 L.; R/Pmax is a term for a function
F W !2<! ! !2 :
We suppose that
1 W !2<! ! !2 :
Then since ADR holds in L.; R/ it follows that there exists a function
h W R<! ! R
677
such that:
(2.1) h 2 L.; R/;
(2.2) Suppose 2 P!1 .R/ and h <! . Then for a comeager set of lters
g Pmax \ , if p0 2 Pmax is a condition such that p0 < q for each q 2 g
then there exist a condition p 2 Pmax and a countable set Z !2 such that
a) \ !2 Z,
b) \ !1 D Z \ !1 ,
c) p Z <! Z,
d) p < p0 .
The function h is easily dened from a winning strategy for Player II in the game
dened as follows. The players alternate choosing pairs of reals dening a sequence
h.xi ; yi / W i < !i after ! many moves. As usual Player I begins by choosing .x0 ; y0 /.
The rules are that for each i < !, .xi / 2 Pmax and that
.xiC1 / < .xi /:
The rst player to violate the rules loses, otherwise Player II wins if for each p0 2 Pmax
such that
p0 < .xi /
for all i < !, there exist p 2 Pmax and a countable set Z !2 such that (2.2a)(2.2d)
hold where
D xi ; yi j i < !:
The game is determined in L.; R/ and by the weak version of Changs Conjecture
described above, Player I cannot have a winning strategy in this game.
We shall need to code terms in L.; R/Pmax for functions
F W !2<! ! !2
by sets of reals. Suppose that 2 L.; R/Pmax is a term such that
1 W !2<! ! !2 :
Let A
be the set of x 2 R such that .x/ D .p; s; / where
(3.1) p 2 Pmax ,
(3.2) s 2 !2<! ,
(3.3) < !2 ,
(3.4) p .s/ D .
678
9 Extensions of L.; R/
We let A
be the code of . Let be the set of A 2 such that A D A
for some
term .
Fix a surjection
W ! !
such that is 1 denable in L.; R/, such a function exists by Theorem 9.105.
We now come to the rst key point. Suppose that A 2 and that 2 L.; R/Pmax
is a term such that A D A
. Suppose in addition that s 2 ! is such that both A and
R n A have scales which are 11 .B/ where B D .s/. Let
D R \ HODL.;R/ s:
The key claim is that for every lter
g HODL.;R/ s \ Pmax ;
if g is HODL.;R/ s-generic and if p0 2 Pmax is a condition such that p0 < q for each
q 2 g, then there exist a condition p 2 Pmax and a countable set Z !2 such that
(4.1) \ !2 Z,
(4.2) \ !1 D Z \ !1 ,
(4.3) p Z <! Z,
(4.4) p < p0 .
This follows from Corollary 9.108 by Theorem 9.109 and Theorem 9.105.
More generally suppose that X ! is a countable set such that every set
A 2 X has a scale in X . Let
D R \ HODL.;R/ .X /:
Let Y L.; R/Pmax be a countable set of terms such that
\ X D A
j 2 Y :
Finally suppose that
g HODL.;R/ .X / \ Pmax
is a lter which is HODL.;R/ .X /-generic and p0 2 Pmax is a condition such that
p0 < q for each q 2 g. Then there exist a condition p 2 Pmax and a countable set
Z !2 such that
(5.1) \ !2 Z,
(5.2) \ !1 D Z \ !1 ,
(5.3) for each term 2 Y ,
(5.4) p < p0 .
p Z <! Z;
679
This too follows from Corollary 9.108, using Theorem 9.109 and Theorem 9.105.
We now x 2 W . We rst apply this last claim in L. 1 ; R1 / where
j1 W L.; R/ ! L. 1 ; R1 / L.; R/G1
is the generic elementary embedding associated to G1 .
Let
HOD1 D j1 .HODL.;R/ /
and let
N G0 !2 -DC:
Let
M0 D f W !2<! ! !2 j f 2 N G0 [ H.!2 /N G0 :
Dene, in N G0 , a set
as follows. T is the set of
T P!1 .M0 /
2 P!1 .M0 /
9 Extensions of L.; R/
680
(7.1) \ !2 Z,
(7.2) \ !1 D Z \ !1 ,
(7.3) f Z <! Z for each function
f W !2<! ! !2
such that f 2 .
We caution, and emphasize, that in dening P!1 .M0 / we are working in N G0 . If
for example has countable conality in L.; R/ then
.P!1 .M0 //N G0 .P!1 .M0 //L.;R/G0 :
Let
S D P!1 .M0 / n T :
Suppose that
H0 Coll.!3 ; P .!2 //N G0
is N G0 -generic with H0 2 L.; R/G1 . Thus
N G0 H0 ZFC:
We prove that in N G0 H0 , the set S is not stationary in P!1 .M0 /. (1) follows from
this claim.
We shall need to use the stationary tower Q< as dened in N G0 H0 where
D C1 :
By Theorem 9.113, is a Woodin cardinal in
HODL.;R/
and so since N G0 H0 is a generic extension of HODL.;R/ for a partial order P
with
L.;R/
< ;
jP jHOD
it follows that is a Woodin cardinal in N G0 H0 .
Assume toward a contradiction that S is stationary in P!1 .M0 /. We shall prove
that this contradicts (6.1).
Let
GS .Q< /N G0 H0
be N G0 H0 -generic with S 2 GS and let
g0 Coll.!; P ./ \ N G0 H0 /
be an N G0 H0 -generic lter with
g0 2 N G0 H0 GS :
We choose GS 2 L.; R/G1 . Note that since
N D HODL.;R/ . /;
681
it follows that
N g0 ZFC:
Let
jS W N G0 H0 ! N .S/ N G0 H0 GS
be the associated generic elementary embedding. Thus since M0 D [S and since
S 2 GS ,
jS M0 2 jS .S/:
From the denition of S it follows that the following must hold in jS N G0 :
(8.1) There exists p0 2 jS .G0 / such that p0 < p for all p 2 G0 and such that for
all sets
Z 2 jS .P!1 .!2 //
if
a) jS !2L.;R/ Z,
b) !1L.;R/ D Z \ jS .!1L.;R/ /,
then there exist p 2 jS .G0 / and 2 Y such that
a) p jS . /Z <! 6 Z,
b) p < p0 .
This we will show contradicts (6.1).
Suppose A 2 . Let SA and TA be trees on ! such that in N ;
A D pTA D R n pSA :
1
Let TA D j1 .TA / D j .TA /. Note that TA1 2 N . We suppose that the set
.A; TA ; SA / j A 2 2 N G0 H0 ;
this is possible since N G0 H0 ZFC.
For each A 2 let A1 be the set pTA1 as dened in L.; R/G1 . The rst
key points are:
(9.1) A1 D j1 .A/;
(9.2) jS .A/ D j1 .A/ \ N .S/ .
Let hAi W i < !i be the N G0 H0 -generic enumeration of dened by g0
in the natural fashion. Dene in L.; R/G1 , B to be the set of x 2 R1 such that
x codes hxi W i < !i such that for each i < !, xi 2 pTA1i . Merging the trees
hTA1i W i < !i denes in a natural fashion a tree
T 1 2 N G0 H0 g0
such that in L.; R/G1 ;
B D pT 1 :
A key point is that the tree T 1 is < weakly homogeneous in N G0 H0 g0 and
that is a limit of Woodin cardinals in N G0 H0 g0 where D L.;R/ . Thus it
follows that:
682
9 Extensions of L.; R/
.S/
t
u
As an immediate corollary of Theorem 9.114, we obtain the following improvement of Theorem 9.39.
Theorem 9.115. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ AD R C is regular:
Let h W < i be the -sequence of L.; R/. For each < let
D A ! ! j w.A/ <
and let
N D HODL.;R/ . /:
Let W be the set of < such that
(i) D ,
(ii) N is regular,
(iii) cof./ > !:
Suppose that 2 W , G0 Pmax is N -generic and that
H0 .Coll.!3 ; P .!2 ///N G0
is N G0 -generic. Then
(1) N G0 H0 ZFC C Martins MaximumCC .c/,
(2) N G0 H0 Strong Changs Conjecture.
Proof. By Theorem 9.114,
N G0 ZF C Strong Changs Conjecture:
Further
P .!2 /N G0 D P .!2 /N G0 H0 :
There by Lemma 9.103,
N G0 H0 Strong Changs Conjecture:
t
u
9.7 Ideals on !2
9.7
683
Ideals on !2
t
u
For the denition of the second saturation property for JNS that we shall dene it is
convenient to dene the notion of a canonical function.
Denition 9.117. Suppose h W !2 ! !2 . Then h is a canonical function if there exists
an ordinal < !3 and a surjection
W !2 !
such that
!2 n < !2 j f ./ D ordertype./
t
u
is not stationary in !2 .
Denition 9.118. The ideal JNS is weakly presaturated if for every function
f W !2 ! !2
and for every set S 2 P .!2 / n JNS , there exists a canonical function h W !2 ! !2 such
that
2 S j f ./ h./ JNS :
t
u
Remark 9.119. Suppose that JNS is weakly presaturated and that
G .P .!2 / n JNS ; /
is V -generic. Let
j W V ! .M; E/ V G
be the associated generic elementary embedding. Then j.!2V / D !3V .
t
u
684
9 Extensions of L.; R/
Lemma 9.120 (Shelah). Suppose that is a regular cardinal and that P is a partial
order such that
V P cof. / < cof.j j/:
Then C is not a cardinal in V P .
t
u
t
u
In contrast, by the results of .Foreman, Magidor, and Shelah 1988/ if is supercompact and if G Coll.!2 ; < / is V -generic then in V G, JNS is !-presaturated.
The same theorem is true with only the assumption that is a Woodin cardinal.
If GCH holds then JNS is not weakly presaturated. In fact if GCH holds then there
is a single function
f W !2 ! !2
such that if G .P .!2 / n JNS ; / is V -generic then !3V < j.f /.!2V /.
An even easier argument proves the following lemma which shows that Martins
Maximum does not imply that JNS is weakly presaturated.
Lemma 9.122. Assume that JNS is weakly presaturated. Suppose that N is a transitive
inner model containing the ordinals such that
(i) N ZFC,
(ii) !2V is inaccessible in N .
Let D !2V . Then . C /N < !3 .
9.7 Ideals on !2
685
Proof. Dene
f W !2 ! !2
by f ./ D .jjC /N .
Let
D ..!2V /C /N . For each <
let
W !
be a surjection such that 2 N . Dene
f W !2 ! !2
by f ./ D ordertype. /.
Thus for each <
there exists a closed unbounded set C !2 such that
f ./ < f ./
for all 2 C .
Suppose G .P .!2 / n JNS ; / is V -generic and let
j W V ! M V G
be the associated generic elementary embedding. Thus for each <
,
D j.f /.!2V / < j.f /.!2V /:
Therefore
t
u
< !2V j f ./ 2 U
t
u
t
u
686
9 Extensions of L.; R/
Remark 9.125. In light of Shelahs theorem that no normal ideal extending JNS can be
presaturated, semi-saturation (together with !-presaturation) is perhaps the strongest
saturation property that such an ideal can have. It implies, for example, that every
t
u
normal ideal which extends JNS is precipitous and much more.
The next theorem, which is essentially an immediate consequence of the denitions, shows, in essence, that semi-saturated ideals on !2V in V correspond to semisaturated ideals on !1V g in V g where g Coll.!; !1V / is V -generic. We state the
theorem only for the nonstationary ideal, the general version is similar.
Theorem 9.126. Suppose that g Coll.!; !1 / is V -generic. The following are
equivalent.
(1) V The nonstationary ideal on !2 is semi-saturated.
(2) V g The nonstationary ideal on !1 is semi-saturated.
t
u
Theorem 9.127 and Lemma 9.128 correspond to Lemma 4.27 and Corollary 4.28
respectively. The proofs are similar, we leave the details to the reader.
Theorem 9.127. Suppose that I P .!2 / is a normal uniform ideal such that the
ideal I is semi-saturated. Suppose that U is a V -normal ultralter which is set generic
over V and such that
U P .!2 / n I
and let
j W V ! M V U
be the associated embedding. Then j.!2V / D !3V .
t
u
The next lemma is an immediate corollary of Theorem 9.127. This lemma shows,
for example, that if JNS is semi-saturated then every function
f W !2 ! !2
is bounded by a canonical function modulo JNS . Thus ! .!2 / implies that JNS is not
semi-saturated and so by Shelahs theorem on ! .!2 / .Shelah 2008/, if 2!1 D !2
then JNS is not semi-saturated.
Lemma 9.128. Suppose that I P .!2 / is a normal uniform ideal such that the
ideal I is semi-saturated. Suppose that
f W !2 ! !2 :
Then there exists a canonical function
h W !2 ! !2
such that
< !2 j h./ < f ./ 2 I:
t
u
9.7 Ideals on !2
687
t
u
If ./ holds then every (normal) semi-saturated ideal on !2 must properly extend
JNS . Therefore we shall only be considering ideals which properly extend JNS but we
note that there are several obvious questions concerning the general case of arbitrary
normal ideals on !2 , with no restriction on 2!1 .
(1) Can JNS be semi-saturated?
(2) Is is possible for every function
f W !2 ! !2
to be bounded by a canonical function pointwise on a closed unbounded set?
(3) Can the nonstationary ideal on !2 be semi-saturated?
(4) Let I be the nonstationary ideal on !2 restricted to the ordinals of conality !1 .
Can the ideal I be semi-saturated?
(5) Suppose that there exists a normal uniform ideal
I P .!2 /
such that I is semi-saturated and contains JNS . Suppose that J P .!2 / is a
normal uniform semi-saturated ideal. Must
JNS J
Remark 9.130. (1) It is plausible that if there is a huge cardinal, then in a generic
extension of V one can arrange that every function
f W !2 ! !2
is bounded pointwise on an !1 -club by a canonical function. Granting this, a
negative answer to the rst question would in effect be an interesting dichotomy
theorem.
(2) The likely answer to the second question is no. Theorem 9.131 shows that if
P .!2 /# exists and if the nonstationary ideal on !2 is semi-saturated, then a
generalization of Theorem 3.19(4) to !2 must hold. This seems impossible.
(3) Let I be the nonstationary ideal on !2 restricted to the ordinals of conality !1 .
It is not known whether the ideal I can be !3 -saturated. This is a well known
problem. Question (3) is a weaker question, possibly signicantly weaker as the
t
u
results concerning JNS show.
Theorem 9.131. Assume P .!2 /# exists. Let I be the nonstationary ideal on !2 . Suppose that I is semi-saturated and that C !2 is closed and unbounded. Then there
exists a set A !1 such that
j !1 < < !2 and L A is admissible C:
688
9 Extensions of L.; R/
t
u
The next theorem, which is a corollary of Theorem 9.126, shows that if the nonstationary ideal on !2 is semi-saturated then one formulation of the Effective Continuum
Hypothesis must hold.
Theorem 9.132. Let I be the nonstationary ideal on !2 . Suppose that I is semisaturated. Suppose that M is a transitive inner model containing the reals such that
M ZF C DC C AD
and such that every set X 2 P .R/ \ M is weakly homogeneously Suslin in V . Then
M !2 :
Proof. Assume toward a contradiction that
M > !2 :
Let A 2 P .R/ \ M be such that
!2
11 .A/
and let 2 Ord be least such that
L .A; R/ ZF C DC:
We note that the existence of is immediate since (trivially) there must exist a
measurable cardinal in V . By the choice of A,
!2 < :
Fix a partial map
W R ! !2
such that:
(1.1) 2 L .A; R/;
(1.2) .t / j t 2 dom./ D !2 ;
1
(1.3) .x; y/ j .x/ .y/ 2
1 .A/;
1
(1.4) Suppose Z dom./ is
1 , then
.t / j t 2 Z
is bounded in !2 .
9.7 Ideals on !2
689
9 Extensions of L.; R/
690
(3.1) X Y jg .!2V /,
(3.2) jY j D !1 in Lg .Ag ; Rg /.
Let
D sup.X /:
By (3.1) and (3.2), is singular in Lg .Ag ; Rg / and so since by the elementarity of
jg , jg .!2V / is a regular cardinal in Lg .Ag ; Rg /, it follows that
< jg .!2V /:
Fix t 2 jg .dom.// such that
Let 2 V
in V ,
Coll.!;!1 /
g .t / D :
1
Y D pT :
and
Y D pT pT D dom./:
9.7 Ideals on !2
691
1 2 pT!1 :
collapsing map and let 0 be the image of . Finally suppose that g Coll.!; !1 0 /
is M0 -generic and let tg be the interpretation of 0 by g . Thus tg 2 dom./ and
by absoluteness,
.t0 / < .tg /:
But tg 2 pT0 and T0 D T where
M0
D !1
D Z0 \ !1 :
t
u
There are three closely related results which improve slightly on the results of
.Foreman and Magidor 1995/; these are stated as Theorem 9.134, Theorem 9.135
and Theorem 9.136 below. These theorems are straightforward corollaries of Theorem 10.62, Theorem 10.63, and Lemma 10.65. We leave the details to the interested
reader.
Remark 9.133. (1) The condition (iii) of Theorem 9.134 is trivially implied by, for
example, the hypothesis that 2@2 < @! .
(2) The condition (ii), that the ideal I be !-presaturated, is certainly easier to
achieve than the condition that I be presaturated. If is a Woodin cardinal
and
G Coll.!2 ; </
692
9 Extensions of L.; R/
Theorem 9.135. Assume that there is a measurable cardinal and that there is a normal, uniform, !3 -saturated ideal on !2 . Then 12 < !2 .
t
u
t
u
9.7 Ideals on !2
693
We end this chapter with two consistency results which can be obtained. As we
have previously noted, .Shelah 2008/ proves that ! .!2 / follows from 2!1 D !2 and
so in particular if 2!1 D !2 then by Theorem 9.127, JNS cannot be semi-saturated. Our
original motivation for obtaining the kind of results below concerned the consistency
strength of
ZF C ADR C is regular:
The point is that the existence of a !3 -saturated ideal on !2 seems likely to be quite
strong beyond the level of superstrong cardinals and so obtaining approximations
starting from a model of ADR in which is regular might provide evidence that this
theory is also quite strong with consistency strength beyond that of superstrong cardinals. However recent results of .Sargsyan 2009/ have shown that the consistency
strength of
ZF C ADR C is regular
is below that of a Woodin cardinal which is a limit of Woodin cardinals.
Theorem 9.137. Suppose P .R/ is a pointclass closed under continuous preimages, S Ord, and that
L.S; ; R/ AD R C is regular:
Suppose G0 Pmax is L.S; ; R/-generic. Suppose
H0 Coll.!3 ; H.!3 //L.S;;R/G0
is L.S; ; R/G0 -generic. Then in L.S; ; R/G0 H0 the following hold.
(1) Martins MaximumCC .c/.
(2) JNS is precipitous.
(3) JNS is weakly presaturated.
t
u
Theorem 9.138. Suppose P .R/ is a pointclass closed under continuous preimages, S Ord, and that
L.S; ; R/ AD R C is regular:
Suppose that G0 Pmax is L.S; ; R/-generic and that
H0 Coll.!3 ; H.!3 //L.S;;R/G0
is L.S; ; R/G0 -generic. Then in L.S; ; R/G0 H0 :
(1) Martins MaximumCC .c/.
(2) For each stationary set S < !2 j cof./ D ! there is a normal uniform
ideal I P .!2 / such that
a) JNS I and S I ,
b) I is semi-saturated.
t
u
Chapter 10
Further results
One fundamental open question is the consistency of Martins Maximum with the axiom ./. The results of Section 10.2 strongly suggest that Martins Maximum does not
imply ./ even in the context of large cardinal assumptions. The situation for Martins
MaximumCC seems more subtle, these issues are discussed briey at the beginning of
Section 10.2.4.
However we shall also prove in Section 10.2 that the axiom ./ is independent of
Martins MaximumCC .c/. This will follow from Theorem 9.39 and Theorem 10.70.
Nevertheless the axiom ./ can be characterized in terms of a bounded form of
Martins Maximum, this is main result of Section 10.3. Such variations of Martins
Maximum were introduced by Goldstern and Shelah.
Theorem 2.53 shows that forcing axioms can be reformulated as reection principles. A natural question therefore is whether Martins Maximum is implied by the
axiom ./ together with some natural reection principle such as the principle SRP of
Todorcevic.
Theorem 10.1 (Larson (Martins Maximum)). Suppose that axiom ./ holds. Then
there is an .!1 ; 1/-distributive partial order P such that
V P ZFC C SRP C :Martins Maximum C Axiom./:
t
u
\ V:
This special case of Martins Maximum has been studied by Q. Feng, he denes this as
the Conal Branch Principle (CBP). It is easily veried that CBP implies SRP.
Theorem 10.2 (Larson (Martins Maximum)). Suppose that axiom ./ holds. Then
there is an .!1 ; 1/-distributive partial order P such that
V P ZFC C CBP C :Martins Maximum C Axiom./:
t
u
695
the forcing notions preserve !1 then once there are Woodin cardinals then essentially
anything can be accomplished (at least if V is any of the current inner models). We
give some denitions and state without proof some results relevant to this somewhat
general question.
Suppose M is a countable transitive model of ZFC. Suppose that is a Woodin
cardinal in M and that EQ D hE W < i is a weakly coherent Doddage in M such
that the sequence is in M and the sequence witnesses that is a Woodin cardinal in M .
More precisely:
(1) For each < , E is a set of . ; /-extenders which are -strong and of
hypermeasure type; i. e. for each E 2 E , cp.jE / D and jE . / .
(2) is strongly inaccessible and < if < < .
(3) E 2 M and EQ 2 M .
Q \ V D EQ \ V for each E 2 E .
(4) (Weak coherence) jE .E/
(5) For each A 2 M; A M there exists < such that for all
<
jE .A/ \
D A \
696
10 Further results
(3) We restrict to iteration trees which are nonoverlapping and normal in the sense
of .Martin and Steel 1994/. The notion of an iteration scheme is dened there in
terms of winning strategies in iteration games.
t
u
We say that an iteration scheme I is < -weakly homogeneously Suslin if the associated set of reals I is < -weakly homogeneously Suslin. The set of reals I is
dened by xing a surjection W R ! H.!1 / which is denable in hH.!1 /; 2i. I is
the preimage of I under . The canonical choice for is 1 denable in hH.!1 /; 2i.
Denition 10.4. Weakly Homogeneous Iteration Hypothesis (WHIH):
(1) There is a proper class of Woodin cardinals.
(2) There exist a Woodin cardinal and a weakly coherent Doddage
hE W < i
which witnesses is a Woodin cardinal such that if > and is inaccessible
then there exists a countable elementary substructure
X V
such that
Q 2 X,
a) ; E
b) hM; EQM i has a iteration scheme which is 1 -homogeneously Suslin,
where M is the transitive collapse of X and EQM is the image of EQ under the
collapsing map.
u
t
WHIH holds in all of the current inner models in which there is a proper class of
Woodin cardinals.
The existence of 1 -weakly homogeneously Suslin iteration schemes for a countQ trivializes the question of what can happen in set generic extenable structure hM; Ei
sions of M . If M elementarily embeds into a rank initial segment of V then similarly
essentially anything can happen in some generic extension of V .
Q is a countable structure where EQ is a weakly coherent
Theorem 10.5. Suppose hM; Ei
Doddage in M witnessing that is a Woodin cardinal for some 2 M . Suppose there
Q which is 1 -weakly homogeneously Suslin. Then
is a an iteration scheme for hM; Ei
any sentence true in a rank initial segment of V is true in a rank initial segment of a
set generic extension of M .
t
u
Theorem 10.6. Suppose there are ! 2 many Woodin cardinals less than . Suppose
Q is a countable structure where EQ is a weakly coherent Doddage in M witnesshM; Ei
ing that is a Woodin cardinal for some 2 M . Suppose there is a an iteration scheme
Q which is < -weakly homogeneously Suslin. Then there is a set R such
for hM; Ei
that M. / is a symmetric extension of M for set forcing and
t
u
M. / ADC :
697
Since the symmetric extension M. / is a model of ADC , the analysis of both Pmax
and Qmax can be carried out in M. /. This yields the following corollary.
Theorem 10.7. Suppose there are ! 2 many Woodin cardinals less than . Suppose
Q is a countable structure where EQ is a weakly coherent Doddage in M witnesshM; Ei
ing that is a Woodin cardinal for some 2 M . Suppose there is a an iteration scheme
Q which is < -weakly homogeneously Suslin. Then:
for hM; Ei
(1) There is a set generic extension of M in which the axiom ./ holds.
(2) There is a set generic extension of M in which the nonstationary ideal on !1 is
t
u
!1 -dense.
As a corollary we obtain, for example:
Theorem 10.8 (WHIH). There exists a partial order P such that
V P ./:
t
u
We now generalize the notion of an iteration scheme to allow the use of generic
elementary embeddings in the construction of the iterations.
Suppose M is a countable transitive model of ZFC. Suppose that 0 ; 1 are Woodin
cardinals in M with 0 < 1 . Suppose hE W < 1 i is a weakly coherent Doddage of
sets of extenders in M1 such that the sequence is in M and the sequence witnesses that
Q is a function
1 is a Woodin cardinal in M . A mixed iteration scheme for hM; 0 ; Ei
which assigns to each countable generic iteration
hM ; G ; j; W < i
an iteration scheme for .M ; EQ / where:
(1) M0 D M , G0 is M -generic for the stationary tower forcing P<0 dened in M .
(2) j W M ! M is a commuting system of elementary embeddings.
(3) G is M -generic for the stationary tower forcing j0; .P<0 /.
(4) MC1 is the generic ultrapower of M given by G and j;C1 is the corresponding elementary embedding.
(5) For each limit ordinal 0 < , M is the direct limit of hM W < i.
Q
(6) EQ D j0; .E/.
We can view a mixed iteration scheme as a partial function
I W H.!1 / ! H.!1 /:
698
10 Further results
(2) There exist a Woodin cardinals 0 < 1 and a weakly coherent Doddage
EQ D hE W < 1 i
which witnesses 1 is a Woodin cardinal such that if 1 < and if is inaccessible then there exists a countable elementary substructure,
X V
containing 0 ; 1 and EQ such that
hM; 0M ; EQM i
has a mixed iteration scheme which is 1 -homogeneously Suslin. Here M is
Q under the
the transitive collapse of X and .0M ; EQM / is the image of .0 ; E/
collapsing map.
t
u
A variant of MIH which we denote by MIH is obtained by modifying part (2)
replacing 0 by !1 , adding the assumption that the nonstationary ideal on !1 is precipitous and considering mixed iterations where rst one iterates by generic ultrapowers
using the nonstationary ideal instead of using the stationary tower.
All of the large cardinals within reach of the current inner model theory are (relatively) consistent with the existence of a wellordering of the reals which is
21 .1 -WH/,
over the base theory
ZFC C There is a proper class of Woodin cardinals:
A natural conjecture is that any large cardinal with an inner model theory is consistent
2 1
with a
1 . -WH/ wellordering of the reals. This can be proved with certain general
assumptions on the inner models.
2 1
t
Theorem 10.10 (MIH or MIH ). There is no
1 . -WH/ wellordering of the reals. u
699
t
u
700
10 Further results
where the product is computed with countable support. If the normal lter generated
by X is proper then Q.X / is .!; 1/-distributive.
Theorem 10.16. Suppose that there is a proper class of Woodin cardinals. Suppose
that is a measurable cardinal,
is a normal ultralter on and that
G Coll.!; </ Q.
/
is V -generic. Suppose that is a 2 sentence and that there exists a partial order
P 2 V such that
V P :
Then there exists a partial partial order Q 2 V G such that
Q
t
u
t
u
Note that the set a indicated in Theorem 10.17 is set generic over M (by Vopenkas
Theorem).
Theorem 10.18. Suppose that P .R/ is a pointclass closed under continuous
preimages such that
L.; R/ ZF C DC C ADR :
There exists a 2 -sentence such that the following hold where
M D HODL.;R/
and where D .0 /L.;R/ .
701
is an ultralter.
702
10 Further results
One obvious approach to coding information into L.P .!1 // is indicated in the
next lemma.
Lemma 10.19. Suppose that for each set A R there exists a sequence
hB; W < < !1 i
of borel sets such that
AD
[ \
B; :
>
Then:
(1) P .R/ L.P .!1 //.
(2) For each function f W R ! R there exists a set a !1 such that for all x 2 R,
f \ L.a; x/ 2 L.a; x/:
t
u
The following theorem is well known. The proof is an easy exercise using Solovays
method for generically coding information using almost disjoint families in P .!/.
Theorem 10.20. There is a -centered boolean algebra B such that if
GB
is V -generic then in V G the following holds. Suppose A R. Then there exists a
sequence hB; W < < !1 i of borel sets such that
[ \
AD
t
u
B; :
>
There is a Pmax version of Theorem 10.20, involving quite different methods. The
effect is similar in that one can arrange for example in the resulting extension that there
exists a sequence, hB; W < < !1 i, of borel sets such that
[ \
R# D
B; :
>
hH.!2 /; 2iL.R/
for any partial order P 2 L.R/. Of course this is a trivial claim if P is .!; 1/distributive in L.R/. The general case, for arbitrary partial orders P 2 L.R/, is more
subtle. It is a plausible conjecture that if
L.R/ AD
703
L.R/ AD
and that for each 2 -sentence , if
Pmax
704
10 Further results
c D i j 2iC1 2 b :
Suppose Y P .!/. An ordinal is a Y -uniform indiscernible if is an indiscernible of La for each a 2 Y . We caution that the Y -uniform indiscernibles are
not necessarily Y Y -uniform indiscernibles. However the following lemma is easily
proved.
Lemma 10.21. Suppose that R is a nonempty set such that for all a 2 <! ,
a# 2 Lx for some x 2 . Suppose that and that
min 2 S
j > !1 D min 2 S j > !1
where
(i) !1 D sup.!1 /Lx j x 2 ,
(ii) S is the set of -uniform indiscernibles,
(iii) S
is the set of -uniform indiscernibles.
Then S
n !1 D S .
t
u
Denition 10.22. Suppose that hSi W i < !i is a sequence of pairwise disjoint stationary subsets of !1 and suppose that z !. Let S D hSi W i < !i. We associate to the
pair .S; z/ two subsets of P .!/;
X(Code) .S; z/ D [X j <
and
Y(Code) .S; z/ D [Y j <
where
S(Code) .S; z/ D h. ; X ; Y / W < i
is the maximal sequence generated from .S; z/ as follows.
705
t
u
Remark 10.23. (1) It might seem more natural to dene the set b ! in (iv) using
the rst ! many uniform Y -uniform indiscernibles. This would allow the use of
single stationary set S in the decoding process instead of a sequence hSi W i < !i
of stationary sets. In fact such an approach is possible, the details are quite
similar. One advantage is that with further modications the coded set, X, is 1
denable, from parameters, in the structure
hH.!2 /; 2i;
instead of the expanded structure, hH.!2 /; INS ; 2i. For our applications this
feature is at best more difcult to achieve; cf. Theorem 10.55, and there are
more elegant ways to achieve this (by simply making the coded set !1 -borel).
706
10 Further results
However one can, by further renements, arrange in the resulting extension that
there exists A !1 such that if N is any transitive set such that
A 2 N,
.INS /N D INS \ N ,
N ZFC ,
then N 2@0 D @2 .
This yields a 2 sentence in the language of the structure,
hH.!2 /; INS ; 2i;
which if true implies c D !2 .
(2) The sequence h W < i can be generated in a variety of ways rather just using
the Y -uniform indiscernibles. Similarly the condition (iv)(b) can be modied
to further thin the sequence. This in effect we shall do in Section 10.2.4, see
Denition 10.72 and and the subsequent Remark 10.73.
(3) If for every x 2 R, x # exists, then S(Code) .S; z/ has length !2 .
(4) The requirement that X0 D ;, rather than X0 D ;, is just for convenience. u
t
We extend these notions to sequences of models.
Denition 10.24. Suppose
hMk W k < !i
is a sequence such that for all k < !,
(1) for each t 2 R \ Mk , t # 2 MkC1 ,
(2) Mk is countable, transitive and
Mk ZFC;
(3) Mk MkC1 and
S D hSi W i < !i
and that S is a sequence of pairwise disjoint sets such that for all i < !,
Si 2 .P .!1 / n INS /M1 :
Suppose z ! and z 2 M0 .
Let
S(Code) .hMk W k < !i; S; z/ D h. ; X ; Y / W < i;
707
let
X(Code) .hMk W k < !i; S; z/ D [X j < ;
and let
Y(Code) .hMk W k < !i; S; z/ D [Y j < ;
where h. ; X ; Y / W < i is the maximal sequence such that for all < , there
exists n 2 ! with
h. ; X ; Y / W < i D S(Code) .Mk ; S; z/ j
t
u
Suppose that hMk W k < !i is a sequence of countable transitive sets satisfying the
conditions in Denition 10.24. We note that since for all k < !,
.INS /MkC1 \ Mk D .INS /MkC2 \ Mk
M
the following holds. Suppose k < !, S 2 .P .!1 //M0 and that < !2 k . Then for
all i > k,
Q MkC1 :
.SQ /Mi \ D .S/
This observation yields the following important corollary which concerns the behavior
of X(Code) .hMk W k < !i; S; z/ under iterations. Suppose that
j W hMk W k < !i ! hMk W k < !i
is an iteration. Then
j X(Code) .hMk W k < !i; S; z/ X(Code) .hMk W k < !i; j.S/; z/
and
j Y(Code) .hMk W k < !i; S; z/ Y(Code) .hMk W k < !i; j.S/; z/:
j.Y(Code) .hMk W k < !i; S; z// D Y(Code) .hMk W k < !i; j.S/; z/:
These claims are easily veried using the properties (1)(3) of Denition 10.24.
It certainly can happen that
X(Code) .hMk W k < !i; S; z/ 2 M0 :
Thus if
is an iteration it may be that j.X(Code) .hMk W k < !i; S; z// does not coincide with the
denition given above. In the cases we shall be interested in,
X(Code) .hMk W k < !i; S; z/ 2 M0
and the two possible denitions of j.X(Code) .hMk W k < !i; S; z// coincide, cf. Remark 10.27(5).
708
10 Further results
10.2.2 Q.X/
max
Suppose the nonstationary ideal on !1 is !1 -dense and there are innitely many
Woodin cardinals with a measurable cardinal above. The covering theorems show that
L.P .!1 // is close to L.R/ below L.R/ . A natural question is whether it must necessarily be the case that L.P .!1 // is a generic extension of L.R/ or whether covering
must hold between L.P .!1 // and L.R/. We note that assuming AD
L.R/ D L.P .!1 //
and so covering trivially holds between these inner models in this case.
We focus on the case when the nonstationary ideal is !1 -dense since this case is
the most restrictive. It eliminates the possibility that sets appear in L.P .!1 // because
they are coded into the structure of the boolean algebra
P .!1 /=INS :
Suppose that X P .!/ is a set such that
L.X; R/ ADC :
.X/
We dene a variation, Q.X/
max , of Qmax such that if G Qmax is L.X; R/-generic
then in L.X; R/G, the nonstationary ideal is !1 -dense and X 2 L.P .!1 //. In fact in
L.X; R/G, X is a denable subset of H.!2 /. Thus, for example, if X codes R# then
in L.X; R/G,
R# 2 L.P .!1 //:
Before dening Q.X/
max it is convenient to dene a renement of Qmax .
709
Qmax
and y D z.
t
u
710
10 Further results
t
u
t
u
Q.X/
max .
711
0 D supk j k < !:
Construct by induction a sequence hgk ; hk W k < !i of generics such that
(1.1) gk Coll .!; Ck \ k /, gk is Lzk -generic, and gk 2 LzkC1 ,
(1.2) hk Coll .!; k / hk is Lzk gk -generic, and hk 2 LzkC1 .
Construct by induction a sequence hGk W k < !i of generics such that the following
conditions are satised. As in the proof of Theorem 6.64 these conditions uniquely
specify the generics. For each k < ! let
bk D 2iC1 j i 2 x0 [ 3iC1 j i 2 zk :
712
10 Further results
Gk \ Coll .!; .; //
is the LzkC1 gh-least lter, F , such that
a) F is Lzk gh-generic,
b) for all < and for all i 2 bkC1 ,
.0; i / 2 F jColl .!; / $ .0; i / 2 G0 jColl .!; /;
c) for all < and for all i bkC1 ,
.0; !/ 2 F jColl .!; / $ .0; i / 2 G0 jColl .!; /;
where g D Gk \ Coll .!; Ck \ /, h D GkC1 \ Coll .!; /, is the least
element of CkC1 above , and for each < , is the th indiscernible of
Lzk above .
Dene
713
Therefore
.hMk W k < !i; f; z0 / 2 Q.X/
max
t
u
and is as required.
Lemma 10.30. Suppose
.hNk W k < !i; g; z/ < .hMk W k < !i; f; z/
in
Q.X/
max
and let
Qmax .
Therefore, since
.j; hMk W k < !i; hMk W k < !i/ 2 N0 ;
M
N0
N1
\ Mk D INS
\ Mk :
INS kC1 \ Mk D INS
From this it follows from the denitions that for all i < !,
S(Code) .hMk W k < !i; Sg ; z/
is an initial segment of S(Code) .Ni ; Sg ; z/. Therefore
S(Code) .hMk W k < !i; Sg ; z/
is an initial segment of
S(Code) .hNk W k < !i; Sg ; z/:
The rst claim of the lemma, (1), follows by the elementarity of j .
We prove (2). Note that .hNk W k < !i; j.h/; x/ 2 Q
max and
.hNk W k < !i; j.h/; x/ < .hMk W k < !i; h; x/
in
Q
max .
714
10 Further results
Let
h. ; X ; Y / W < i D S(Code) .Ni ; Sg ; z/:
and let 0 be such that for all i < !,
S(Code) .hMk W k < !i; j.Sf /; z/ D S(Code) .Ni ; j.Sf /; z/j0 :
Since
.hMk W k < !i; h; x/ 2 Q.X/
max
it follows that
Since .hMk W k < !i; j.f // is an iterate of .hMk W k < !i; f / and since
.hMk W k < !i; f / 2 Q
max ;
it follows that the Z -uniform indiscernibles above !1N0 coincide with the Z-uniform
indiscernibles above !1N0 . Since
.hMk W k < !i; j.h/; x/ 2 Q.X/
max
it follows that the Y0 0 -uniform indiscernibles above !1N0 coincide with the Z -uniform
indiscernibles above !1N0 . Further these coincide with the Y0 -uniform indiscernibles
above !1N0 .
715
Finally j.h/./ D g./ for all < !1N0 such that is a Z-uniform indiscernible
and such that > !1M0 .
Therefore by induction on it follows that if 0 C < then
0
Y.0 C / n Y0 D Y.
n Y0 0
0
C /
0
and
0
X.0 C / n X0 D X.
n X0 0 :
0
C /
0
t
u
(2) follows.
Remark 10.31. There is an important difference between Q.X/
max and Qmax . Suppose
716
10 Further results
Proof. (3) is an immediate consequence of (2) and the denition of the order on Q.X/
max .
(2) follows from (4) since
.hMk W k < !i; f; z/ 2 N0 :
To see this suppose
k W hNk W k < !i ! hk.Nk / W k < !i
is a countable iteration. Then by elementarity it follows that
k.X / D k.X(Code) .hNk W k < !i; Sg ; x// [ k.j.X(Code) .hMk W k < !i; Sf ; z///
where
X D X(Code) .hNk W k < !i; j.Sf /; z/:
Therefore k.X / X.
We construct the iteration j to satisfy (1) and (4).
Fix c0 2 X(Code) .hMk W k < !i; Sf ; z/ and let
b0 D 2iC1 j i 2 c0 [ 3iC1 j i 2 x:
Dene a function
g0 W !1N0 ! N0
by perturbing g as follows. Let C be the set of uniform [P .!/ \ Mk j k < !indiscernibles below !1N0 and above !1M0 . Let D C be the set of 2 C such that
C \ has ordertype .
For each < !1N0 , g0 ./ D g./ unless D C where 2 D, < and
is the th element of C past . In this case
g0 ./ D g./
if .0; i / 2 f ./ and i 2 b0 , otherwise
g0 ./ D .0; !/ _ p j p 2 g./:
Let
j W hMk W k < !i ! hMk W k < !i
be the iteration of length .!1 /N0 determined by g0 . Clearly j 2 LxG.
We come to the key claims. Let
S(Code) .hMk W k < !i; Sf ; z/ D h ; X ; Y W < i;
let
j.S(Code) .hMk W k < !i; Sf ; z// D h ; X ; Y W < i;
and let
S(Code) .hNk W k < !i; j.Sf /; z/ D h ; X ; Y W < i:
717
is the least [P .!/ \ Mk j k < !-uniform indiscernible above !1N0 . The latter
implies that
sup j < D [Mk \ Ord j k < !
by clause (4(h)) in the denition of Q.X/
max .
We prove the second claim. From the rst claim
Y
D [Y j <
and
D [Mk \ Ord j k < !:
Let
Y D [Y j < :
Q.X/
max ,
Si D j .0; i / 2 g./
and let
Ti D j .0; i / 2 j.f /./:
718
10 Further results
For each i !, let SQi be the set computed from Si and let TQi be the set computed
from Ti , each computed relative to [Nk j k < !. Thus for each < !1N0 and for
each i < !, if i 2 b0 and if 2 Ti then 2 TQi otherwise 2 TQ! .
For each i < !, Ti is stationary in [Nk j k < ! and so putting everything
together, if D C !1N0 then
(2.1) YC1
D Y
[ x,
D X
(2.2) XC1
[ c0 ,
(2.3) C1
is the least indiscernible of Lx above !1N0 ,
SQi \ Z D TQi \ Z
where Z is the set of indiscernibles of such that is an indiscernible of Lx
and such that
2 [Nk j k < !:
Thus j has the desired properties and this proves the lemma.
t
u
719
Nevertheless the properties (1.1)(1.3) sufce to implement the proof of Lemma 10.32.
This yields an iteration
j W hMO k W k < !i ! hj.MO k / W k < !i
k
such that
(2.1) j 2 N0 ,
(2.2) for all k < !,
Ok
Lemma 10.34. Suppose that X P .!/ and that for all t 2 R, t # exists. Then Q.X/
max
is homogeneous.
Proof. This follows by Lemma 10.30(2).
Suppose that
.hMk W k < !i; f; z/ 2 Q.X/
max
and that
Let t 2 R code the pair .hMk W k < !i; hMk0 W k < !i/.
By Theorem 10.29, there exists
.hNk W k < !i; g; x/ 2 Q.X/
max
such that t 2 Lx and such that the lter
G Coll.!; <!1 /N0
given by g is Lx-generic.
720
10 Further results
such that
.hNk W k < !i; j.f /; z/ 2 Q.X/
max ;
such that
and
Q.X/
max j.hNk W k < !i; g; x/
are isomorphic.
Finally
.hNk W k < !i; j.f /; z/ < .hMk W k < !i; f; z/
and
t
u
721
t
u
t
u
722
10 Further results
Coll.!; !1M /
and where for each < !1M , is the th Z-uniform indiscernible above !1M0 ;
(5) h.!1M / D g;
(6) For all < !1M ,
(7) hp W < !1M i 2 M0 .
723
Nk D MQ 0 k :
be a countable iteration, of limit length, such that j.!1N0 / > and such that
q 2 j.f /./:
The iteration exists since the critical sequence of any iteration of
hNk W k < !i
is an initial segment of the Z-uniform indiscernibles above !1N0 . Let hi W i < !i be
O
be the (unique) iteration such that ki .fi / D j.f /ji , and let
Q.X/
max
M0.p/
where we let
724
10 Further results
Let IZ be the class of Z-uniform indiscernibles and let IZ be the class of Z -uniform
indiscernibles. Thus
N
IZ D IZ n !1 0 :
Further Z 2 M0.p/ and Z is countable in M0.p/ .
The key point is the following. Let
M
.p/
h W < !1 0 i
be the increasing enumeration of IZ and let
I D C j D and 0 < < !1M :
Suppose that fO 2 M.p/ is a function such that
0
.p/
M
(1.1) dom.fO/ D !1 0 ,
.p/
M0
fO./ Coll.!; /
M0
n I,
f.p/ ./ D fO./:
Since hAk W k < !i 2 M0.p/ , we can choose fO so that requirement (4) of the lemma
is satised by the condition .hMk.p/ W k < !i; fO; z/ by modifying fOjI if necessary.
But this implies that requirement (4) is satised by any condition q 2 Q.X/
max such that
.p/
O
q < .hM W k < !i; f ; z/:
k
is dense in
Q.X/
max
\ M g. Let
.hMk W k < !i; h; z/ 2 D! M \ M g
1
t
u
725
726
10 Further results
hp j < i 2 M Gj ;
[P .!/.p / j < D P .!/ \ M Gj ;
f . / D GjColl.!; /;
Suppose that
.0; / 2 GjColl.!; . C /M /;
< < . C /M , and that for each < , is a P .!/.p / -uniform
indiscernible.
Then < .!1 /M0 and
.0; !/ 2 f ./:
727
F W !1N0 ! .!1N0 ; /
728
10 Further results
(2.1) L t ZFC,
(2.2) is an inaccessible cardinal in .!1 /L t .
Since t codes .M;
/, E D. Further E 2 Lt and in Lt , E contains a subset
which is a club in !1N0 .
For each 2 E let be the least ordinal, , satisfying (2.1)(2.2), and let F be
the function F as computed in L t .
The point of all of this is reection. Let > be an ordinal such that
L t ZFC
and suppose
X L t
is a countable elementary substructure containing t and F . Let D X \ !1N0 . Then
2 E and F is the image of F under the collapsing map.
For each < ! N0 let
T D 2 E j .0; / 2 f . /:
Thus
M
Let
729
hp W < !1N0 i
730
10 Further results
!1N0 ,
k W .M; / ! .j.M /; j. //
be the iteration map corresponding to (5.3) and let T D k.T / where T 2 M is the
tree on ! used to dene . The key point is that since .M; I/ is strongly Y -iterable
it follows that
hV!C1 \ j.M /j.G/; pT \ j.M /j.G/; 2i hV!C1 ; Y; 2i:
The proof of this claim involves noting that if
kO W .M;
/ ! .MO ;
/
O
731
t
u
732
10 Further results
Theorem 10.40. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC C is regular:
Suppose that X P .!/ is a set such that
X 2 L.; R/:
Q.X/
max
AC
holds.
.X/
Proof. By Lemma 10.34, Q.X/
max is homogeneous, by Lemma 10.33, Qmax is !-closed.
The claim that
L.; R/G !2 -DC
follows from (5) since (5) implies that R can be wellordered in L.; R/G.
We work in L.; R/G and prove (1)(5).
(1) is immediate, G is the set of
.hMk W k < !i; f; z/ 2 Q.X/
max
such that z D zG , there is an iteration
j W hMk W k < !i ! hj.Mk / W k < !i
such that j.f / D fG . This iteration is uniquely specied by fG and the requirement
that j.f / D fG .
(2) follows from Lemma 10.38, using Theorem 10.39 and Theorem 10.29 to supply
the necessary conditions.
(3) and (4) follow from (2) and the denitions.
Finally by (2)(4),
X(Code) .SfG ; zG / D [j .X(Code) .hMk W k < !i; Sf ; z// j .hMk W k < !i; f; z/ 2 G
where as above, for each .hMk W k < !i; f; z/ 2 G,
j W hMk W k < !i ! hMk W k < !i
is the iteration such that j.f / D fG .
733
Therefore
X(Code) .SfG ; zG / X [ ;:
By the genericity of G and by Lemma 10.35,
X [ ; [j .X(Code) .hMk W k < !i; Sf ; z// j .hMk W k < !i; f; z/ 2 G
and so X(Code) .SfG ; zG / D X [ ;. A similar argument, using Lemma 10.36, proves
that Y(Code) .SfG ; zG / D P .!/.
Finally (6) follows by an argument essentially identical to the proof that AC holds
t
u
in the Qmax -extension.
Lemma 10.41. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC C is regular:
Suppose that X P .!/ is a set such that
X 2 L.; R/:
Suppose that
G Q.X/
max
is L.; R/-generic.
Then in L.; R/G, for every set A 2 P .R/ \ L.; R/ the set
X hH.!2 /; A; 2i j MX is A-iterable and X is countable
contains a club, where MX is the transitive collapse of X .
Proof. The proof is a modication of the proof of Lemma 5.107.
We cannot really use the proof of Lemma 6.77, which is the version of this lemma
for Qmax . The minor difculty is that Q.X/
max Qmax , and so there are no conditions
.hMk W k < !i; f; z/ 2 Q.X/
max
such that
hV!C1 \ M0 ; A \ M0 ; 2i hV!C1 ; A; 2i:
This fact accounts for the various differences between the presentation of the anal
ysis of the Q.X/
max -extension and that of the Qmax -extension.
We work in L.; R/G.
Let
H.!2 /G D [H.!2 /M0 j .hMk W k < !i; f; z/ 2 G
where for each .hMk W k < !i; f; z/ 2 G,
j W hMk W k < !i ! hMk W k < !i
is the (unique) iteration such that j.f / D fG .
By Theorem 10.40(1),
P .!1 / D P .!1 /G ;
734
10 Further results
and so
H.!2 / D H.!2 /G :
This is the key to the proof, just as it was the for the proof of Lemma 5.107.
Let
F W H.!2 /<! ! H.!2 /
be a function such that for all Z H.!2 / if F Z Z then
hZ; A \ Z; G \ Z; 2i hH.!2 /; A; G; 2i:
Suppose X H.!2 / is a countable subset such that
hX; F \ X; G \ X; 2i hH.!2 /; F; G; 2i:
Let MX be the transitive collapse of X . We prove that MX is A-iterable.
Let hsi W i < !i enumerate X .
Let hNi W i < !i be a sequence of elements of X such that the following hold for
all i < !.
(1.1) !1 N0 .
(1.2) Ni 2 NiC1 .
(1.3) si 2 Ni .
(1.4) hNi ; A \ Ni ; G \ Ni ; 2i hH.!2 /; A; G; 2i.
Since !1 N0 , for each i < !, Ni is transitive.
Since
H.!2 /G D H.!2 /;
there exist sequences
h.hMki W k < !i; fi ; z/ W i < !i;
hai W i < !i;
and
hbi W i < !i
such that for all i < !,
(2.1) .hMki W k < !i; fi ; z/ 2 G \ NiC1 \ X ,
(2.2) .hMkiC1 W k < !i; fiC1 ; z/ < .hMki W k < !i; fi ; z/,
(2.3) ai 2 M0i ,
(2.4) for all p 2 Zi \ G,
.hMkiC1 W k < !i; fiC1 ; z/ < p;
(2.5) ji .ai / D Ni ,
(2.6) ji .bi / D si ,
735
j .A \ ai / D A \ j .ai /
736
10 Further results
737
X
However for each i < !, NiX is transitive and NiX 2 NiC1
. Further
MX D [NiX j i < !:
Therefore
and so
MO X D [jO.NiX / j i < !
jO.A \ MX / D A \ MO X :
t
u
Therefore MX is A-iterable.
738
10 Further results
Proof. (1)(4) follow from Theorem 10.40. (5) follows from (1), Lemma 10.41 and
from Theorem 3.42 using the chain condition of Q.X/
max to reduce to the case that
f WS !
where < L.A;X;R/ , cf. the proof of Lemma 6.79.
t
u
Perhaps our covering theorems do not capture all the covering consequences of the
assumption that the nonstationary ideal is !1 -dense, particularly if in addition large
cardinals are assumed to exist. Theorem 10.44 is the version of Theorem 10.42 which
addresses this question.
Theorem 10.43. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADR :
Then for each set A 2 there exists an inner model L.S; R/ such that
(1) S Ord and A 2 L.S; R/,
(2) HODSL.S;R/ There is a proper class of Woodin cardinals.
t
u
t
u
739
.;/
10.2.3 Pmax
.;/
We dene and briey analyze Pmax
which is the version of Q.X/
max which corresponds to
.;/
Pmax but with X D ;. Our interest in Pmax lies in Theorem 10.70. This theorem shows
that
Martins Maximum CC .c/ C Strong Changs Conjecture
hM; a; zi
such that the following hold.
(1) M is a countable transitive set such that
M ZFC C ZC:
(2) M is iterable.
(3) M
AC .
740
10 Further results
t
u
.;/
.;/
The nontriviality of Pmax
is an immediate corollary of the analysis of L.; R/Qmax
where
P .R/
such that
(1) A \ M 2 M and hM \ V!C1 ; A \ M; 2i hV!C1 ; A; 2i,
(2) M is A-iterable,
(3) X(Code) .M; a; z/ D ;.
741
AC
holds.
t
u
742
10 Further results
.;/
It is convenient to organize the analysis of Pmax
following closely that of Q.X/
max . The
reason is simply that most of the proofs adapt easily to the new context. The next four
lemmas summarize the basic iteration facts that one needs. These lemmas are direct
analogs of the lemmas we proved as part of the analysis of Q.X/
max . We leave the details
to the reader.
.;/
, such as the !-closure
The rst two easily yield elementary consequences for Pmax
.;/
and homogeneity of Pmax , the latter two allow one to complete the basic analysis.
.;/
Pmax
and let
j W M0 ! M0
t
u
.;/
As we have indicated, the iteration lemmas required for the analysis of Pmax
are
.;/
.;/
routine generalizations of those for Qmax . The situation for Pmax is actually quite a bit
.;/
conditions are simpler and there is more freedom in
less complicated since the Pmax
constructing iterations.
t
u
743
t
u
744
10 Further results
.M; I/ 2 H.!1 /
is such that .M; I/ is strongly Y -iterable. Let 2 M be the Woodin cardinal associated
to I. Suppose t !, t codes M and
.;/
hN ; a; zi 2 Pmax
;
is a condition such that t 2 Lz. Let
2 M be a normal .uniform/ measure and let
.M ;
/ be the !1N -th iterate of .M;
/. Then there exists a sequence
hp W < !1N i 2 N
and there exists .b; x/ 2 N such that
.;/
(1) hN ; b; xi 2 Pmax
,
t
u
.;/
is homogeneous.
Lemma 10.53. Suppose that L.R/ ADC . Then Pmax
t
u
!1M .
.;/
Suppose that G Pmax
is a semi-generic lter. Then
745
AC
holds.
.;/
.;/
Proof. By Lemma 10.52, Pmax
is !-closed and by Lemma 10.53, Pmax
is homogeneous.
By the usual arguments, (2) and the assertion that
L.R/G ZF C !1 -DC
each follow from Lemma 10.51 using Theorem 10.47 to supply the necessary conditions.
.;/
. (5) implies
(4), (5) and (6) follow from (2) and the denition of the order on Pmax
that R can be wellordered in L.R/G and so
L.R/G ZFC:
By (2), if C !1 is closed, unbounded, then C contains a closed, unbounded, subset
which is constructible from a real. Thus
.INS /L.R/AG D .INS /L.R/G \ L.R/AG :
This implies that
.S(Code) .SAG ; zG //L.R/AG D .S(Code) .SAG ; zG //L.R/G ;
and so L.R/Ag ZFC.
The generic lter G can be dened in L.R/Ag as the set of all
.;/
hM; a; zG i 2 Pmax
746
10 Further results
(1.1) j.a/ D AG ,
(1.2) j 2 L.M; AG /,
747
Thus
.;/
hM0 ; a0 ; zG i 2 Pmax
.;/
G0 Pmax
5j .A0 /
which is a contradiction.
.;/
There is an interesting absoluteness theorem for Pmax
.
Pmax
.;/
Pmax
:
:
Proof. Fix the 2 sentence . We give the proof in the case that none of the predicates
for the sets Y occur in . The general case is similar.
As usual we may suppose that
D .8x0 .9x1 .x0 ; x1 ///
where
748
10 Further results
O
j W M ! M
H.!2 /M
j.b0 /; b1 :
L .R/ ZFC :
.;/
We claim there exists a semi-generic lter F Pmax
such that the following hold
where
hSi W i < !i D SAF :
(2.1) hM; a; zi 2 F .
(2.2) There exists
Y L .R/
.;/
such that !1 Y and such that F \ Y is Y -generic for Pmax
.
749
.;/ L .0 /
Let g0 .Pmax
/ 0
be an L 0 .0 /-generic lter with
L 0 .0 /g0 ZFC ;
(3.4) the P .!/F uniform indiscernibles above !1 are exactly the .P .!/ \ X0 /uniform indiscernibles above !1 ,
where as above, P .!/F D P .!1 /F \ P .!/.
.;/
Fix a semi-generic lter F Pmax
which satises (2.1)(2.4). Let
B0 D j1 .b0 /
where
j1 W M ! M
750
10 Further results
Similarly let hTk0 W k < !i be the image of hTk W k < !i under the collapsing map and
let B00 be the image of B0 .
Fix t0 ! which codes N . Let
.;/
O d; zi
hM;
O 2 Pmax
be such that t0 is recursive in z.
O
Let
j2 W N ! N
O
Si \ C D Si \ C;
O
where C is the set of < !1 such that is an indiscernible of Lz.
(4.2) IF0 D P .!1 /F0 \ INS .
(4.3) Let be the !1th .X \ P .!//-uniform indiscernible. Then
3iC1 j i 2 zO D i < ! j .Si / \ is stationary:
O Note that we do not require that
Such an iteration is easily constructed in M.
.INS /N D .INS /M \ N :
We now come to the key points. First
.;/
O j2 .a0 /; zi
O 2 Pmax
hM;
and
O j2 .a0 /; zi
O < hM; a; zi:
hM;
where b 2 H.!2 /N
O j2 .a0 /; zi
O
hM;
t
u
751
.;/
As we have indicated, our interest in Pmax
lies in Theorem 10.70 which shows that
752
10 Further results
and let
d D i < ! j 3iC1 2 b:
By the genericity of G again, we can suppose that c D t and that d codes M.
Let
S(Code) .S; z/ D h. ; X ; Y / W < !2 i:
t
u
753
Theorem 10.58. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC C is regular:
Suppose is a 2 sentence in the language for the structure
hH.!2 /; 2; Y W Y 2 L.; R/; Y Ri;
and that
Pmax
:
Then
.;/
Pmax
:
t
u
t
u
We shall obtain this as corollary of a slightly more general theorem, Theorem 10.62, which requires the following generalization of one of the main denitions
of .Foreman and Magidor 1995/.
Denition 10.60. Suppose that P is a partial order. P is weakly proper if for every
ordinal ,
.P!1 .//V
is conal in
P
.P!1 .//V :
t
u
754
10 Further results
Remark 10.61. Foreman and Magidor dene a partial order P to be reasonable if for
every ordinal ,
.P!1 .//V
is stationary in
P
.P!1 .//V :
It is not difcult to show that this notion is strictly stronger than that of being weakly
proper. Recall that P is proper if for every and for every set S P!1 ./, if S is
stationary then S is stationary in V P . Foreman and Magidor, .Foreman and Magidor
1995/, prove Theorem 10.62 and (implicitly) a strong version of Theorem 10.63 for
reasonable partial orders; this version does not require the hypothesis,
L.A; R/ AD:
The special case of L.R/; i. e. A D ;, has also been examined by Neeman and Zapletal,
but, as here, in the context of the relevant determinacy hypothesis.
t
u
Theorem 10.62. Suppose that P is a partial order such that P is weakly proper. Suppose that jP j D and that for each set A , A# exists. Then
P
.12 /V D .12 /V :
Proof. There exists a tree T on ! 2 such that if g P is V -generic then in V g;
pT D x # j x 2 R:
It is convenient to work in V g. Since jP jV !2V , !3V is a cardinal in V g.
Assume toward a contradiction that
.12 /V < .12 /V g :
Fix x0 2 RV g such that
.12 /V < 0
where 0 is the least ordinal above !1V such that L 0 x0 is admissible.
Thus x0# 2 pT . Since P is weakly proper, there exists a subtree S0 T such that
(1.1) S0 2 V ,
(1.2) jS0 jV D !,
(1.3) x0# 2 pS0 .
Let S be the transitive collapse of S0 so that S is a tree on ! for some countable
ordinal, which is isomorphic to the tree S0 . Let x 2 RV code S and for each !1
let be the least ordinal above such that L x is admissible.
Thus in V the following hold,
(2.1) for all t 2 pS , t D z # for some z 2 R,
(2.2) for all < !1 and for all t 2 pS ,
rank.M.z # ; ! C // <
where z # D t and where M.z # ; ! C / is the ! C model of z # .
We note that (2.2) holds by boundedness.
755
By absoluteness, (2.1) and (2.2) hold in V g. Therefore in V g, for all < !1 ,
rank.M.x0# ; ! C // < :
By reection
rank.M.x0# ; !1 // < !1 ;
which contradicts the choice of x0 since necessarily
!1 < .12 /V :
t
u
There is a closely related theorem. Recall the following which is formally stated
as Theorem 2.30. Suppose that A R is such that every set in P .R/ \ L.A; R/
is -weakly homogeneously Suslin and that P is a partial order such that P 2 V .
Suppose that T is a -weakly homogeneous tree such that
A D pT :
Finally, suppose that G P is V -generic. Then there is a generic elementary embedding
jG W L.A; R/ ! L.AG ; RG /
such that
(1) jG .A/ D AG D pT V G ,
(2) RG D RV G ,
(3) L.AG ; RG / D jG .f /.a/ j a 2 RG ; f W R ! L.A; R/ and f 2 L.A; R/.
Further the properties (1)(3) uniquely specify jG .
Theorem 10.63. Suppose that A R,
L.A; R/ AD;
and that every set in P .R/ \ L.A; R/ is -weakly homogeneously Suslin. Suppose
that P 2 V is a partial order such that P is weakly proper. Suppose that G P is
V -generic and let
jG W L.A; R/ ! L.AG ; RG /
be the associated generic elementary embedding. Then
jG ./ D
for all 2 Ord.
Proof. Assume toward a contradiction that for some 2 Ord,
jG ./ :
Let be the least ordinal such that jG ./ ; i. e. the critical point of jG . Necessarily,
< L.A;R/ :
Let
W dom./ !
756
10 Further results
V G
t
u
The technical lemma, Lemma 10.59, which we require for the proof of Theorem 10.69 is an immediate corollary of the next theorem.
757
Theorem 10.64. Suppose P is a partial order such that P is !3 -cc and such that
P
.!1 /V D .!1 /V . Then P is weakly proper.
Proof. By the chain condition of P , !3V is a cardinal in V P and so both !1V and !3V
are cardinals in V P . Therefore by Lemma 9.120,
.cof.!2V //V
> !:
.P!1 .//V ;
where D !2V . But this is immediate.
t
u
Theorem 9.134, Theorem 9.135, and Theorem 9.136 (these are the theorems of
Section 9.7 concerning ideals on !2 ) are immediate corollaries of the boundedness
theorems, Theorem 10.62 and Theorem 10.63, together with the next lemma.
Lemma 10.65. Suppose that I P .!2 / is a normal uniform ideal such that
j cof./ D ! 2 I:
Let P D hP .!2 / n I; i. Suppose that either
(1) I is !3 -saturated, or
(2) I is !-presaturated and that P is @! -cc, or
(3) 2@2 D @3 and
.!1 /V D .!1 /V :
Then P is weakly proper.
Proof. (1) is an immediate corollary of Theorem 10.64. The proof of (2) is straightforward. The relevant observation is that since the ideal I is !-presaturated and since
< !2 j cof./ D ! 2 I;
it follows that for each k < !,
V
//V
.cof.!kC1
> !:
> !;
and of course one is only concerned with those values of k < ! such that P is not
!kC1 -cc; i. e. with cardinals below the chain condition satised by P .
758
10 Further results
Since 2@2 D @3 , P is !4 -cc in V and so all cardinals above !3V are preserved.
Therefore we need consider only the cases k 2. For k D 0 this is immediate and the
case k D 1 follows by appealing to the generic ultrapower associated to the V -generic
lter G P . This leaves the case k D 2; i. e. !3V . But this case now follows by
Lemma 9.120.
t
u
Lemma 10.68, below, isolates the application of Lemma 10.59 within the proof of
Theorem 10.69. This lemma in turn requires the following two lemmas.
Lemma 10.66. Suppose that hS W < !1 i is a sequence of stationary subsets of
!1 and that h W !1 i is a closed increasing sequence of cardinals such that for
each < !1 , C1 is measurable and such that !1 <
. Suppose that S !1 is
stationary and let Z be the set of X 2 P!1 .
/ such that
(1) X \ !1 2 S ,
(2) For each X \ !1 ,
ordertype.X \ / 2 S :
Then Z is stationary in P!1 .
/.
Proof. Suppose T !1 is stationary. Let GT be the game played on !1 for T :
The players alternate choosing countable ordinals, i , for i < ! with Player I
choosing i for i even. Player I wins if
supi j i < ! 2 T:
Since T is stationary, Player II cannot have a winning strategy.
For each < !1 let G be the game of length !
.1 C / dened as follows. A play
of the game is an increasing sequence
h W < !
.1 C /i
of countable ordinals. Player I chooses for even and Player II chooses for
odd.
Player II wins if for some ,
sup j < !
.1 C / S :
We claim that for each , Player II cannot have a winning strategy in G . This
is easily proved by induction on . Let 0 be least such that Player II has a winning
strategy in G0 and let
W !1<!.1C0 / ! !1
be a winning strategy for Player II. Clearly we may suppose that 0 is least for all
possible choices of hS W 0 i.
If 0 D 0 or if 0 is a successor ordinal then one obtains a contradiction by producing a winning strategy for Player II in GS0 .
759
If 0 is a limit ordinal then again one can construct a winning strategy for Player II
in the game GS0 by using an increasing ! sequence conal in 0 . One constructs a
strategy
W !1<! ! !1
for Player II in GS0 such that if hi W i < !i is a play against then there exists a
play,
h W < !
.1 C 0 /i;
against such that
(1.1) for all < 0 ,
sup j < ! .1 C / 2 S ;
Y H.
C /
such that
(2.1) X Y ,
(2.2) X \ !1 D Y \ !1 ,
(2.3) for each X \ !1 ,
ordertype.Y \ / 2 S :
If not then Player II has a winning strategy in G where D X \ !1 . This follows
from the following observation. Suppose
Z H.
C /
is a countable elementary substructure and 2 Z is a measurable cardinal. Let
2 Z
be a normal measure on and let
2 \A 2 Z j A 2
:
Let
Z D f ./ j f 2 Z:
760
10 Further results
Then
(3.1) Z H.
C /,
(3.2) Z \ V D Z \ V .
Using this it is straightforward to prove the claim above; if Y H.!2 / does not exist
then Player II has a winning strategy in G where D X \ !1 .
Thus Y H.
C / exists as required and the lemma follows.
t
u
Lemma 10.67. Suppose that hSi W i < !i is a sequence of pairwise disjoint stationary
subsets of !1 and that there exist !1 many measurable cardinals. Then there is a
partial order P such that P is .!; 1/-distributive and such that if G P is V -generic
then
.INS /V D .INS /V G \ V
and in V G there exists a sequence
hTi W i < !i
of pairwise disjoint subsets of !1 and an ordinal such that:
(1) For each i < !, Ti !1 and for each S 2 P .!1 / \ V n INS , both S \ Ti INS
and S n Ti INS .
(2) !1 < < !2 and cof. / D !1 .
(3) There exists a closed conal set C such that for each i < !,
hC; SQi \ C; 2i h!1 ; Ti ; 2i:
(4) Suppose that
W !1 !
is a surjection and that < .
a) Suppose that i < !,
S D < !1 j ordertype./ 2 Si ;
and that S is stationary. Then for each k < !, both S \ Tk and S n Tk are
stationary in !1 .
b) Suppose that cof./ D !1 , C is closed and conal, S !1 is
stationary and that for some i < !,
hC; SQi \ C; 2i h!1 ; S; 2i:
Then for each k < !, both S \ Tk and S n Tk are stationary in !1 .
761
is V -generic where the product partial order is computed with countable support. Then
in V g:
(3.1) V ! V .
(3.2) For each < !1 , 2 AQ .
(3.3) For each < !1 let
g D g \
P . ; A /:
<
Then
.INS /V gC1 D .INS /V g \ V g C1 :
762
10 Further results
2!1 n
where D C 1, and so
V g D V g C1 g C1;!1
where
g C1;!1 D g \
Y
V
P . ; A /
2!1 n
where D C 1.
Thus (3.3) follows by applying Lemma 10.66 in V g C1 and arguing as above.
Let G0 Coll.!1 ; !1 / be V -generic and in V G0 let for each k < !,
Tk D fG1
.k/
0
where fG0 W !1 ! !1 is the function given by G0 .
For each i < ! let i be the i th measurable cardinal. For each (nonzero) limit
ordinal < !1 let CiC1 be the . C i /th -measurable cardinal where i < ! and let
D sup j < :
Let Q0 be the product partial order, dened in V G:
Y
P . ; Si /;
Q0 D
2Z
where
(5.1) Z D [Tk j k < !,
(5.2) for each 2 Z, i D k where
2 Tk :
763
H0
Y
V G0
P . ; Si /
2Z\
and where
H0;!1
D H0 \
Y
P . ; Si /
V G0 H0
:
2Zn
V
P . ; Si /
2Z\
Y
V G0
P . ; Si /
2Z\
since V ! V in V G0 . Therefore it follows that for each < !1 , G0 is V H0 generic for Coll.!1 ; !1 /.
Thus for each < !, if
C1
S 2 .P .!1 / n INS /V H0
H0C1 G0 ,
then in V
for each k < !, both S \ Tk and S n Tk are stationary. Both
(4(a)) and (4(b)) follow from this.
t
u
Lemma 10.68. Suppose that V D L.S; A; R/G and
(i) L.S; A; R/ ADC ,
(ii) S Ord and A R,
.;/
(iii) G Pmax
is an L.S; A; R/-generic lter.
Suppose 2 Ord,
L.S; A; R/G is a Woodin cardinal;
and that
L.S; A; R/G There is a measurable cardinal above :
Suppose that 2 Ord is such that < , S 2 V , and such that
V 2 V:
764
10 Further results
Suppose that
Y L .S; A; R/G
is a countable elementary substructure with G; S; A; Y , let M0 be the transitive collapse of Y , and let .0 ; g0 / be the image of .; G/ under the collapsing map.
Suppose that P0 2 M0 g0 is a partial order such that
M0 g0 P0 is !2 -cc and jP0 j D !2
and such that
M0 g0 INS D .INS /V
P0
\ V :
and an iteration
j W .M0 g0 h0 h1 ; I0 / ! .M0 j.g0 /j.h0 /j.h1 /; j.I0 //
of length !1N such that
a) j 2 N ,
b) j..INS /M0 g0 h0 / D .INS /N \ M0 j.g0 /j.h0 /,
c) for each p 2 j.g0 /, hN ; a; zi < p.
Proof. The key point is that, by Theorem 10.64,
M g0 P0 is weakly proper
and so by Lemma 10.67,
.12 /M0 g0 D .12 /M0 g0 h0 :
Let a0 D .Ag0 /M0 g0 and let z0 D .zg0 /M g0 . Thus
X(Code) .M g0 ; Sa0 ; z0 / D X(Code) .M g0 h0 ; Sa0 ; z0 /
since by Theorem 10.54(5),
P .!/ \ M0 D Y(Code) .M0 g0 ; Sa0 ; z0 /:
Here is the second place we make full use of the thinning requirement, Denition 10.22(iv(b)), the rst place was in the proof of Theorem 10.55.
Let
hsi0 W i < !i D .Sa0 /M0 g0
765
766
10 Further results
767
Then,
(3.1) j.0 / < ,
(3.2) S(Code) .N ; j.Sa0 /; z0 /jj.0 / 2 M0 j.g0 /j.h0 /j.h1 /,
D X D ;,
(3.3) Xj.
0/
(3.4) Yj.
D Yj.
[ z.
0 /C1
0/
Let
(4.1) D j.0 / C .
.
(4.2) For each < , D j.
0 /C1C
t
u
Theorem 10.69. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ AD R C is regular:
.;/
Suppose G0 Pmax
is L.; R/-generic. Suppose
H Coll.!3 ; H.!3 //L.;R/G0
is L.; R/G0 -generic. Then
L.; R/G0 H0 ZFC C Martins MaximumCC .c/:
Proof. Using Lemma 10.68, the proof is quite similar to the proof of Theorem 9.39. u
t
Combining Theorem 10.55, Theorem 10.57, Theorem 10.69, and Lemma 10.56 we
obtain the following theorem. The proof follows closely that of Theorem 9.114 which
is outlined in Section 9.6 where the hypothesis is discussed. This theorem shows that
for Theorem 4.76 it is essential that the predicate INS be part of the structure.
Theorem 10.70. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ AD R C is regular:
Let h W < i be the -sequence of L.; R/. For each < let
D A ! ! j w.A/ <
and let
N D HODL.;R/ . /:
Let W be the set of < such that D and such that N is regular.
.;/
Suppose that 2 W , G0 Pmax
is L.; R/-generic and that
H0 .Coll.!3 ; P .!2 ///N G0
is N G0 -generic. Let M D N G0 H0 . Then:
768
10 Further results
Pmax
hH.!2 /; Y; 2 W Y R; Y 2 L.R/iL.R/
:
Then
hH.!2 /; Y; 2 W Y R; Y 2 L.R/iM I
(3) M :./.
Proof. The claim that
follows by Theorem 10.69 and (2) follows by Theorem 10.55. By Theorem 10.57,
X(Code) .SAG ; zG / D ;
and so (3) follows from Lemma 10.56.
The proof that
M Strong Changs Conjecture
requires adapting the proof of Theorem 9.114. This is straightforward, we leave the
details to the reader.
u
t
.;;B/
10.2.4 Pmax
It is not difcult to prove the following theorem. One uses the proof that Martins
Maximum implies ! .!2 / (Theorem 5.11 from the rst edition) together with the fact
that, assuming Martins Maximum, if G is V -generic for Namba forcing then
.12 /V < .12 /V G :
Theorem 10.71. Assume Martins Maximum. Suppose that
S D hSi W i < !i
is a sequence of pairwise disjoint stationary subsets of !1 . Then for each z 2 P .!/,
X(Code) .S; z/ D P .!/:
t
u
.S; z/,
It is straightforward to dene minor variations of X(Code) .S; z/, say X(Code)
for which it seems very unlikely that Theorem 10.71 holds; i. e. for which it is unlikely
that Martins Maximum implies
X(Code)
.S; z/ D P .!/:
For example one could simply add the requirement, in the calculation of
X(Code) .S; z/, that at every stage the stationary subsets of !1 given by SQi \ be independent from all the previous stationary sets (cf. Denition 10.22(iv)). This gives a
769
plausible approach to showing that Martins Maximum does not imply ./, even if one
assumes in addition that large cardinals are present.
The situation for Martins MaximumCC is more subtle. Indeed the question of
whether Martins MaximumCC implies ./ assuming some additional large cardinal
hypothesis, is in essence the question of whether some large cardinal hypothesis implies that there exists a semiproper partial order P such that
V P ./:
However there is a natural modication in Denition 10.22 which plausibly yields an
approach to showing that Martins MaximumCC does not imply ./ outright. This in
.;;B/
.;/
turn yields another variation of Q.X/
max which we denote Pmax . As is the case for Pmax ,
.;;B/
.;/
Pmax is in essence a variation of Qmax , but with a new parameter B R.
.;;B/
-extension yields the following result.
The analysis of the Pmax
Fix B R with B 2 L.R/. Then
Martins Maximum CC .c/ C Strong Changs Conjecture
together with all the 2 consequences of ./ for the structure
hH.!2 /; INS ; B; 2i
does not imply ./. Thus for the characterization of ./ (Theorem 4.76), using the
converse of the absoluteness theorem, it is essential that predicates be added for
conally many sets Y R with Y 2 L.R/. This result complements the results of the
previous section which show that the predicate for INS must be added.
Denition 10.72. Suppose that hSi W i < !i is a sequence of pairwise disjoint stationary subsets of !1 Suppose that z ! and that B R. Let S D hSi W i < !i and
let
ES D C i C 1 j < !1 ; is a limit ordinal and 2 Si :
Let
A(Code) .S; z; B/ D [A j <
where S(Code) .S; z/ D h. ; X ; Y / W < i and hA W < i is the sequence:
(i) A0 D ;.
(ii) Suppose is not the successor of an ordinal of conality !1 . Then
A D [A j < :
(iii) Suppose has conality !1 and let
b D i < ! j SQi is stationary in :
Let d D i j 3
iC1
770
10 Further results
t
u
Remark 10.73. (1) Thus, with notation from Denition 10.72, new elements are
added to A(Code) .S; z; B/ only at stages when new elements could be added to
X(Code) .S; z/ and various additional side conditions are satised. Again this can
be modied. For example one could replace (iii(e)) with the condition that
N ADC
and ES is N -generic for P ; where P D Smax , or P D Bmax etc.
More subtle effects can be achieved by modifying (iii(f)). For example if one replaces Pmax by Smax in (iii(e)) then in light of the absoluteness theorem for Smax ,
it would be natural to make the analogous change in (iii(f)) requiring in addition
that Suslin trees be preserved. This is the correct analog of A(Code) .S; z; B/ for
Smax .
(2) Note that A(Code) .S; z; B/ X(Code) .S; z/. This is because we have dened
a D 2iC1 j i 2 b:
We could easily decouple A(Code) .S; z; B/ and X(Code) .S; z/ by setting
a D 5iC1 j i 2 b:
However by adopting the former approach, certain aspects of the analysis of the
.;;B/
.;/
-extension can be reduced to the analysis of the Pmax
-extension.
Pmax
(3) By (iii(b)), (iii(e)), (iii(f)), and (essentially) Lemma 10.56,
a) !1 D .!1 /N ,
b) D D .!2 /N ,
c) S(Code) .S; z/j D .S(Code) .S; z//N ES .
t
u
Lemma 10.56, which analyzes X(Code) .S; z/ in the context of ./, generalizes to
characterize ./ in terms of the behavior of A(Code) .S; z; B/.
771
772
10 Further results
t
u
773
AC .
t
u
.;/
.;/
Remark 10.77. Let P Pmax
be the suborder of Pmax
dened by
.;/
j M is B-iterable:
P D hM; a; zi 2 Pmax
.;;B/
. This is because we have dened A(Code) .S; z; B/ so
Then P is a suborder of Pmax
that
A(Code) .S; z; B/ X(Code) .S; z/:
This observation which we have discussed in Remark 10.73, allows one to infer the
.;;B/
.;/
from that of Pmax
.
u
t
nontriviality of Pmax
774
10 Further results
t
u
.;/
.;;B/
As for the analysis of Pmax
it is convenient to organize the analysis of Pmax
fol.X/
.;/
lowing closely that of Qmax (and Pmax ). Again the reason is simply that the proofs
.;;B/
adapt easily to prove the corresponding lemmas for the Pmax
analysis. The next four
lemmas give the basic iteration facts one needs.
j W M0 ! M0
t
u
.;;B/
are routine
The remaining iteration lemmas required for the analysis of Pmax
.;/
generalizations of those for Pmax
.
775
t
u
776
10 Further results
Z D P .!/ \ M
t
u
.M; I/ 2 H.!1 /
is such that .M; I/ is strongly Y -iterable. Let 2 M be the Woodin cardinal associated
to I. Suppose t !, t codes M and
.;;B/
;
hN ; a; zi 2 Pmax
is a condition such that t 2 Lz. Let
2 M be a normal .uniform/ measure and let
.M ;
/ be the !1N -th iterate of .M;
/. Then there exists a sequence
hp W < !1N i 2 N
and there exists .b; x/ 2 N such that
.;;B/
(1) hN ; b; xi 2 Pmax
,
777
From Lemma 10.79 and Lemma 10.80 one easily obtains the homogeneity and the
.;;B/
!-closure of Pmax
.
.;;B/
is
Lemma 10.83. Suppose that B R and that L.B; R/ ADC . Then Pmax
!-closed.
t
u
.;;B/
Lemma 10.84. Suppose that B R and that L.B; R/ ADC . Then Pmax
is
homogeneous.
t
u
!1M .
.;;B/
Suppose that G Pmax
is a semi-generic lter. Then
is L.R/-generic. Then
L.R/G ZFC
and in L.R/G:
(1) L.R/G D L.R/fG ;
(2) P .!1 /G D P .!1 /;
778
10 Further results
AC :
Finally (3) can be proved by adapting the proof of the analogous claim for Pmax . As in
.;/
the case of Pmax
, (3) can also be proved by rst proving that for each set
Z 2 L.R/ \ P .R/
there exists a countable elementary
X hH.!2 /; Z; 2i
such that MX is Z-iterable, where MX is the transitive collapse of X . See the proof of
Theorem 10.54(3).
t
u
779
j W .M; I / ! .M ; I /
such that j.a/ D AG and such that
I D .INS /L.R/G \ M :
Then
(1) FG is a lter in Pmax ,
(2) FG \ D ; for each dense set D Pmax which is denable in hH.!1 /; B; 2i
from parameters in H.!1 /,
(3) P .!1 /G D P .!1 /FG .
Proof. By Theorem 10.85, the hypothesis of Lemma 4.74 holds in L.R/G and so by
Lemma 4.74, FG is a lter in Pmax .
We assume toward a contradiction that (2) fails and we x t 2 R such that there is
a dense set D t Pmax such that
(1.1) FG \ D t D ;,
(1.2) D t is denable in the structure
hH.!1 /; B; 2i
from t .
Fix a condition hM; a; zi 2 G which forces (1.1).
Let M0 be a countable transitive set such that
(2.1) M0 ZFC C 1 -Replacement,
(2.2) R \ M0 2 M0 ,
(2.3) B \ M0 2 .L.R//M0 ,
(2.4) hV!C1 \ M0 ; B \ M0 ; 2i hV!C1 ; B; 2i,
(2.5) M0 is B-iterable,
(2.6) M 2 .H.!1 //M0 ;
and such that M0 \ Ord is a small as possible.
780
10 Further results
if and only if 2 Ti .
Let a D j.a/ and let
h. ; X ; Y / W < i D .S(Code) .Sa ; z//M0 :
It follows from (3.1) and (3.2) that
and so Y0 D P .!/ \ M , where 0 D .!2 /M . Thus for each < !1M0 ,
0 C D ;
and so if 1 D 0 C .!1 /M0 then
Y1 D Y0 [ x:
Therefore by the choice of M0 , minimizing M0 \ Ord,
.A(Code) .Sa ; z; B//M0 D ;
.;;B/
and
and so hM0 ; a ; zi 2 Pmax
jG W M0 ! M0
781
jO W M ! MO
is the (unique) iteration such that jO.a/ D AG . Clearly we may assume that the condition hM; a; zi forces this. Repeating the construction given above yields a contradiction.
t
u
.;;B/
As an immediate corollary we obtain the desired absoluteness theorem for Pmax
.
:
Then
.;;B/
Pmax
:
.;;B/
Proof. Suppose that G Pmax
is L.R/-generic.
Let FG be the set of h.M; I /; ai 2 Pmax such that in L.R/G there exists an
iteration
j W .M; I / ! .M ; I /
782
10 Further results
z
if and only if
.;;B/
Pmax
z
t
u
.;;B/
in the usual
Theorem 10.88. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ADC C is regular
.;;B/
and that B 2 \ L.R/. Then Pmax
is !-closed and homogeneous. Further, suppose
.;;B/
G Pmax is L.; R/-generic. Then
t
u
The proof of Theorem 10.69 easily adapts to prove the corresponding version for
.;;B/
.;;B/
Pmax
. One uses Lemma 10.67 to produce the analog of Lemma 10.68 for Pmax
in
essentially the same manner.
783
Theorem 10.89. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ AD R C is regular
and that B 2 \ L.R/.
.;;B/
Suppose G0 Pmax
is L.; R/-generic. Suppose
H Coll.!3 ; H.!3 //L.;R/G0
is L.; R/G0 -generic. Then
t
u
L.; R/G0 H0 ZFC C Martins MaximumCC .c/:
Putting everything together we obtain Theorem 10.90. The proof of Theorem 10.90(2) follows closely that of Theorem 9.114 which is outlined in Section 9.6
where the hypothesis is discussed. This theorem shows that for Theorem 4.76 predicates for conally many sets in P .R/ \ L.R/ must be added to the structure
hH.!2 /; INS ; 2i.
Theorem 10.90. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ AD R C is regular:
Let h W < i be the -sequence of L.; R/. For each < let
D A ! ! j w.A/ <
and let
N D HODL.;R/ . /:
Let W be the set of < such that D and N is regular. Suppose that
.;;B/
B 2 P .R/ \ L.R/, 2 W , G0 Pmax
is L.; R/-generic, and that
H0 .Coll.!3 ; P .!2 ///N G0
is N G0 -generic. Let M D N G0 H0 . Then:
(1) M Martins MaximumCC .c/;
(2) M Strong Changs Conjecture;
(3) Suppose is a sentence in the language for the structure
hH.!2 /; INS ; B; 2i;
L.R/Pmax
. Then hH.!2 /; INS ; B; 2iM ;
and that hH.!2 /; INS ; B; 2i
(4) M :./.
Proof. (1) follows by Theorem 10.89 and as we have indicated above, (2) is proved by
the methods of Section 9.6.
(3) is an immediate corollary of Theorem 10.87 since
P .!1 /M D P .!1 /G :
Similarly, by Theorem 10.88,
.A(Code) .hSi W i < !i; zg ; B//M D ;
where hSi W i < !i D SAG . Therefore (4) follows by the equivalences to ./ given in
Lemma 10.74.
t
u
784
10 Further results
\ V:
\ V:
\ V:
t
u
Remark 10.92. It follows from the results of this section and the preceding section
that Bounded Martins Maximum does not imply Bounded Martins MaximumCC . This
can easily be strengthened to show both:
(1) Bounded Martins Maximum does not imply Bounded Martins MaximumC ;
(2) Bounded Martins MaximumC does not imply Bounded Martins MaximumCC .
t
u
The following lemmas of Bagaria give useful reformulations of Bounded Martins
Maximum and of Bounded Martins MaximumCC .
785
t
u
t
u
AC
holds.
786
10 Further results
AC
787
t
u
t
u
The axiom ./ implies a very strong form of Bounded Martins MaximumCC . This
is the content of the next theorem, Theorem 10.99, which in essence is simply a reformulation of the fundamental absoluteness theorem, Theorem 4.64, for Pmax .
Remark 10.98. The requirement on N , in Theorem 10.99, that for each partial order
P 2 N,
1
N P
2 -Determinacy;
can be reformulated in terms of large cardinals. In fact, since R N , it is equivalent
to the assertion that for each set a 2 N , with a Ord, .M1 .a//# 2 N where M1 .a/ is
computed in V . M1 .a/ denotes the minimum (iterable) ne structure model of
ZFC C There is a Woodin cardinal
containing the ordinals and constructed relative to the set a. The formal denition
involves the ne structure theory of Mitchell and Steel .1994/.
t
u
Theorem 10.99 (Axiom ./). Suppose that for each partial order P ,
1
V P
2 -Determinacy:
788
10 Further results
\VI
j W .M; I / ! .M ; I /
a) .P .!1 //M N ;
b) !1N D !1M ;
c) Q3 .S / N , for each S 2 N such that S !1N ;
d) If S !1N , S 2 M and if S I then S is a stationary set in N .
Then
789
N 1 N P I
4
1
i. e. that 4 statements with parameters from V are absolute between N and N P . Fix
a set E Ord such that E 2 N and such that
H. /N 2 LE
where D .jP [ H.!2 /jCC /N .
Suppose that G P is N -generic and assume toward a contradiction
hH.!2 /; INS ; 2i .:/A:
Let g Coll.!; sup.E// be N G-generic. Then E # together with the iteration
j W .M; I / ! .M ; I /
which sends a to A, witness that (2.1) fails in N Gg which contradicts
N 1 N Gg:
4
t
u
Remark 10.100. One corollary of Theorem 10.99 is that the consistency of Bounded
Martins MaximumCC is relatively weak even in conjunction with large cardinal axioms. For example if
ZFC C There is a proper class of Woodin cardinals
is consistent then so is
ZFC C There is a proper class of Woodin cardinals
C Bounded Martins MaximumCC :
In contrast,
ZFC C There is a measurable cardinal
C Martins Maximum
implies the consistency of
ZFC C There is a proper class of Woodin cardinals
and much more. The latter is by results of Steel combined with results of Schimmerling.
t
u
Suppose that P .R/ is a pointclass closed under continuous preimages such
that
790
10 Further results
2P
max
2P
max
generalizes to L.; R/
. In particular,
t
u
t
u
791
By altering the choice of the inner model, N , in the proof of Lemma 10.102, one
can also prove the following variation.
Lemma 10.103. Suppose that P .R/ is a pointclass closed under continuous
preimages such that
L.; R/ AD R :
Suppose that G 2 Pmax is L.; R/-generic.
N L.; R/G containing R [ G such that
t
u
t
u
792
10 Further results
1
Denition 10.105. ZF
2 -Determinacy:
t
u
Remark 10.106. Of course, (2) of Denition 10.105 implies (1), and so an equivalent
notion is obtained by eliminating (1) from the denition.
t
u
Theorem 10.107. Suppose that for each partial order P ,
1
V P
2 -Determinacy:
Then there exists a transitive inner model N containing the ordinals such that
(1) P .!1 / N ,
(2) N ZFC,
1
(3) for each partial order P 2 N ,N P ZF
2 -Determinacy,
793
Let
G0 Coll.!1 ; H.!2 //
be N -generic. Thus there exists A !1 such that A 2 N G0 and such that
N G0 D L# .A/:
Let
A D < !1 j jjL
# .A\/
D !:
# .A\/
D !:
# .A\/
O
D !:
t
u
The second theorem, Theorem 10.108, is closely related Theorem 9.75 and Theorem 9.81, which show that closure properties of P .!1 / transfer upwards to closure
properties of P .!2 /, assuming WRP.2/ .!2 /.
Theorem 10.108. Suppose that
1
V Coll.!;!1 /
2 -Determinacy:
794
10 Further results
Proof. We prove that for each set a Ord, if a 2 N then a# 2 N ; i. e. that for each
partial order P 2 N ,
1
NP
1 -Determinacy:
The proof of the theorem is similar.
Fix a Ord with a 2 N . Assume toward a contradiction that
a# N:
Let D sup.a/ and suppose that
G0 Coll.!1 ; /
be N -generic. Thus there exists A !1 such that
A 2 N G0
and such that A# N G0 . Therefore there exists b0 2 H.!1 / such that in N G0 ,
< !1 j b0 2 .A \ /#
is stationary and co-stationary in !1 . However
.INS /N D .INS /N G0 \ N
and so by Lemma 10.94,
hH.!2 /; INS ; 2iN 1 hH.!2 /; INS ; 2iN G0
since N Bounded Martins MaximumCC .
Therefore there exists AO !1 such that AO 2 N and such that
< !1 j b0 2 .A \ /#
is stationary and co-stationary in !1 . But by the hypothesis of the lemma, AO# exists
and so the set
< !1 j b0 2 .A \ /#
cannot be both stationary and co-stationary. This is a contradiction.
t
u
795
Theorem 10.109. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ AD R C is regular:
.;;B/
Let B R be a set in L.R/. Suppose that G0 Pmax
is L.; R/-generic and let
L.;R/
M D a 2 .HOD/
RG0 j j rank.a/ <
where D ./L.;R/ . Then:
M P ADL.R/ :
(2) M :./.
(3) Suppose that N M is a transitive inner model such that
a) P .!1 /M N ,
b) N ZFC,
1
c) For each partial order P 2 N , N P
2 -Determinacy.
Then N Bounded Martins MaximumCC .
t
u
Denition 10.110 (FengMagidorWoodin). Suppose that A R. Then A is universally Baire if for any compact Hausdorff space X and any continuous function,
W X ! R;
the set a 2 X j .a/ 2 A has the property of Baire in X .
t
u
The next theorem gives a useful characterization of the sets A R which are
universally Baire.
Theorem 10.111 (FengMagidorWoodin). Suppose A R.
equivalent.
796
10 Further results
Theorem 10.112. Suppose there is a proper class of Woodin cardinals and that
A R. Then following are equivalent.
(1) A is universally Baire.
(2) A is 1 -weakly homogeneously Suslin.
(3) A is 1 -homogeneously Suslin.
t
u
t
u
797
t
u
t
u
t
u
t
u
Remark 10.120. (1) Theorem 10.119 is proved using basic method for proving
Theorem 5.104; i. e. the proof uses core model methods. We note that the theorem is false at the projective level.
(2) It is open whether the actual converse to Theorem 10.117 holds.
We shall need the generalization of Lemma 10.117 to L.R# /.
Lemma 10.121. Suppose that for each partial order P ,
V P ADL.R / :
#
t
u
798
10 Further results
\ V:
t
u
\ V:
t
u
799
t
u
t
u
800
10 Further results
Theorem 10.127. Assume there is a proper class of measurable cardinals which are
limits of Woodin cardinals. Then the following are equivalent.
(1) ./.
(2) Suppose that N is a transitive set such that
a) P .!1 / N ,
b) N ZFC,
c) N Every set which is
1! .R# / is universally Baire.
Then for each set A 2 L.R/ \ P .R/,
N A-Bounded Martins MaximumCC :
Proof. We rst prove that (2) follows from ./. The proof is a generalization of the
proof of Theorem 10.99.
Fix A R with A 2 L.R/ and suppose that N is a transitive set such that
(1.1) P .!1 / N ,
(1.2) N ZFC,
1
#
(1.3) N Every set which is
! .R / is universally Baire.
We note that
801
(2.3) Suppose that h.M1 ; I1 /; a1 i < h.M0 ; I0 /; a0 i, .M1 ; I1 / is A-iterable, and that
A \ M1 2 M1 . Let
j W .M0 ; I0 / ! .M0 ; I0 /
be the iteration such that a1 D j.a0 /. Then
hH.!2 /; A \ M1 ; I1 ; 2iM1 a1 :
Now suppose P 2 N is a partial order and that G0 P is N -generic with
.INS /N D .INS /N G0 \ N:
Let be the least strongly inaccessible cardinal of N such that
P 2 N :
Let G1 Coll.!; / be N G0 -generic. We work in N G0 G1 D N G. The key
point is that since R# is universally Baire in N and since
L.R/ AD;
it follows, by Theorem 10.118 and by Theorem 10.119, that .R# /N G D .ZG /N G
where Z D .R# /N , and so
hH.!1 /; A; 2i hH.!1 /; AG ; 2iN G
where AG is the interpretation of A in N G;
AG D [pT j T 2 N and A D pT \ N :
Thus if N H is a set generic extension of N with H 2 N G, then in N G,
.R# /N H D .R# /N G \ N H :
We can now apply Theorem 10.125, to obtain in N G a structure
.M1 ; I1 ; / 2 H.!1 /N G
such that
(3.1) M1 is transitive,
(3.2) M1 ZFC C is a Woodin cardinal,
(3.3) I1 D .I< /M1 ,
(3.4) .M1 ; I1 / is AG -iterable,
(3.5) H.!2 /M1 D H.!2 /N .
N G
By Lemma 4.40, there exists a condition h.M; I /; ai 2 Pmax
such that
802
10 Further results
803
(10.1) j.a/ D X ,
(10.2) I D INS \ M .
By (9.1), X Lt for each t 2 RN G and so since
N G ./;
it follows by Theorem 4.60, that the set X is L.R/N G -generic for .Pmax /N G . Let
GX .Pmax /N G be the L.RN G /-generic lter given by X . By (9.1),
FX D GX \ N:
Suppose that D Pmax is dense with D 2 L.RN /. Let A be the set of t 2 RN such
N G
that t codes an element of D. Let DG be the set of p 2 Pmax
such that there exists
#
t 2 AG which codes p. Since R is universally Baire in N it follows that
.R# /N G D ..R# /G /N G :
N G
and further, by (9.1), that
Thus it follows that DG is dense in Pmax
t
u
804
10 Further results
t
u
t
u
805
t
u
t
u
Q1! .a/
t
u
can be dened in
806
10 Further results
Theorem 10.135. Assume there exists a proper class of Woodin cardinals. Suppose
that a Ord and that
R La:
Suppose that b [a. Then the following are equivalent.
(1) b 2 Q1! .a/.
(2) b 2 N for each transitive inner model of ZFC such that for some 2 Ord;
a) a 2 N ,
b) for each z Ord, if z 2 N then M! .z/# 2 N ,
c) is a Woodin cardinal in N .
t
u
Corollary 10.136. Assume there exists a proper class of Woodin cardinals and that N
is a transitive inner model of ZFC with
P .!1 / [ Ord N:
Suppose that is a Woodin cardinal in N and that
N R# is universally Baire:
Then for each a 2 N \ P ./,
Q1! .a/ N:
t
u
Theorem 10.137 gives our nal characterization of ./. It is quite likely that the
closure requirements on N cannot be signicantly weakened, more precisely that the
requirement:
for each set a Ord, if a 2 N then .M! .a//# 2 N ,
does not sufce.
Theorem 10.137. Suppose that there exist a proper class of Woodin cardinals. Then
the following are equivalent.
(1) ./.
(2) Suppose that N is a transitive set such that
a) P .!1 / N ,
b) N ZFC,
c) for each set a Ord, if a 2 N then Q1! .a/ N .
Then for each set A 2 L.R/ \ P .R/,
N A-Bounded Martins MaximumCC :
t
u
.;;B/
The following corollary of the analysis of Pmax
shows that the equivalences given
in Theorem 10.128 and Theorem 10.133 are essentially the best possible.
10.4 -logic
807
Theorem 10.138. Suppose P .R/ is a pointclass closed under continuous preimages such that
L.; R/ AD R C is regular:
.;;B/
Suppose that B R is a set in L.R/. Suppose that G0 Pmax
is L.; R/-generic
and let
M D a 2 .HOD/L.;R/ RG0 j rank.a/ <
t
u
10.4 -logic
The absoluteness theorems associated to Pmax and its variations can more naturally
be formulated using -logic this strengthening of !-logic was introduced in the
rst edition of this book. The presentation given in this edition reects a number of
expository changes, specically what was -logic in that edition, is now the logical
relation for -logic, T and it is with this denition that we begin. An equivalent
denition based on the generic-multiverse generated by V is given in .Woodin 2009/.
Denition 10.139. Suppose that:
(i) There exists a proper class of Woodin cardinals.
(ii) T is a theory containing ZFC.
(iii) is a sentence.
Then T if for all partial orders P and for all 2 Ord, if VP T then
t
u
VP .
808
10 Further results
Using the generic elementary embeddings associated to the stationary tower one
can prove the following absoluteness theorem. This requires using the full stationary
tower, P< , rather than the restricted tower, Q< , which we have used almost exclusively up to this point.
Theorem 10.140. Suppose that there exist a proper class of Woodin cardinals. Suppose that T is a theory containing ZFC and is a sentence. Then
T
if and only if for each partial order P , V P T .
The previous theorem strongly suggests that the validities of ZFC in -logic which
are 2 -sentences somehow capture the extent of the inuence of large cardinals. This
intuition leads to the denition of the proof relation for -logic and to the Conjecture.
A central aspect of this denition of the proof relation for -logic involves the notion of an A-closed model where A R is universally Baire. Recall that if A R is
universally Baire then A has a canonical interpretation, AG , in any set generic extension, V G, of V ;
AG D [pT j T 2 V and A D pT V :
The denition we shall give of when a transitive set M is A-closed involves AG . However this can be dened without reference to AG . For example if
M ZFC
then there is a very natural reformulation using the Stone spaces, XP , dened in V
from partial orders P 2 M .
Denition 10.141. Suppose that A R and that A is universally Baire. A transitive
set M is A-closed if for each partial order
P 2 M;
P
t
u
t
u
10.4 -logic
809
We have dened in Denition 4.66, Q3 .a/ for each set a 2 H.!1 /. A similar
denition applies to dene Q3 .a/ for an arbitrary set a provided for example that there
exists a proper class of Woodin cardinals (much less is required): Let b be the transitive
closure of a. Then Q3 .a/ is the set of all Y b such that the following hold.
(1) There exists a transitive inner model M of ZFC such that:
a) Ord M;
b) a 2 M;
c) for some ; a 2 V , 2 M and is a Woodin cardinal in M;
d) Y 2 M.
(2) Suppose that M is a transitive inner model of ZFC such that:
a) Ord M;
b) a 2 M;
c) for some ; a 2 V , 2 M and is a Woodin cardinal in M.
Then Y 2 M.
Remark 10.143. Suppose there exists a proper class of Woodin cardinals and that M
is a transitive set such that, M ZFC. Then the following are equivalent:
(1) M is A-closed for each 13 set A.
(2) For each set a Ord if a 2 M then Q3 .a/ M .
t
u
Remark 10.144. Suppose there exists a proper class of Woodin cardinals and that M
is a transitive set such that, M ZFC. Then the following are equivalent:
(1) M is R# -closed.
(2) For each set a Ord if a 2 M then M! .a/# 2 M .
t
u
t
u
We shall dene T ` only for the language of set theory and for theories T
containing the axioms of ZFC. More general denitions are naturally possible.
810
10 Further results
t
u
We note that the proof relation for -logic can be dened without reference to
the universally Baire sets, referring instead to iterable structures. In this approach,
A closure, for the relevant universally Baire sets, is reformulated in terms of closure
under the (unique) iteration strategies of canonical countable structures.
Remark 10.147. One very natural generalization of -logic would allow additional
unary predicates to be interpreted by designated universally Baire sets. This in fact
will be implicit in some of what follows; cf. Theorem 10.169.
t
u
One application of the basic descriptive set theory of ADC is the following important absoluteness theorem, the relevant theorem of ADC is Theorem 9.7.
Theorem 10.148. Suppose that there exist a proper class of Woodin cardinals.
Suppose that T is a theory containing ZFC and is a sentence. Then
T `
if and only if for each partial order P ,
V P T ` .
t
u
The following soundness theorem is easily proved from the denitions using either
the generic elementary embeddings associated with the stationary tower, I< , where
is a Woodin cardinal, or by using the closure properties of the pointclass of all universally Baire sets.
Theorem 10.149. Suppose that there exist a proper class of Woodin cardinals, is a
sentence, and ZFC ` : Then ZFC .
The Conjecture is the conjecture that the converse of Theorem 10.149 holds, in
effect this is simply the conjecture that the Completeness Theorem holds.
10.4 -logic
811
Denition 10.150 (The Conjecture). Suppose that there is a proper class of Woodin
t
u
cardinals. Then for all sentences , ZFC ` if and only if ZFC .
The following theorem shows that the Conjecture is a consequence of iteration
hypotheses.
Theorem 10.151 (WHIH). Suppose that T is a theory containing ZFC and that is a
sentence. Then the following are equivalent:
(1) T ` .
t
u
(2) T .
As we have indicated, many of the theorems relating to Pmax and its variations are
more naturally presented in the context of -logic. For example the general existence
can be strengthened as follows.
theorem, Theorem 5.49, for conditions in Pmax
Theorem 10.152. Suppose that there exists a proper class of Woodin cardinals and
that A R is universally Baire. Suppose that is a 2 sentence such that
ZFC C
is -consistent. Then there is a condition .hMk W k < !i; a/ 2 Pmax
such that
M0 ZFC C ;
and such that
(1) A \ M0 2 M0 ,
(2) hH.!1 /M0 ; A \ M0 i hH.!1 /; Ai,
(3) hMk W k < !i is A-iterable,
and further the set of such conditions is dense in Pmax
.
t
u
Pmax
:
t
u
In fact by Theorem 4.76, one has the following equivalence for ./. This reformulation does not involve forcing at all. Of course the denition of -logic that we have
given does implicitly involve forcing. However, as we have noted, there is another
denition of -logic in terms of iterable structures which does not involve forcing either. Further this equivalence for ./ still holds if one weakens -logic by restricting
the collection of test models to just R# -closed models. As noted in Remark 10.144,
R# -closure has a reformulation purely in ne structural terms.
812
10 Further results
Theorem 10.154. Suppose that there exists a proper class of Woodin cardinals. Then
the following are equivalent.
(1) ./.
(2) For each 2 sentence in the language for the structure
hH.!2 /; INS ; 2; A W A 2 P .R/ \ L.R/i;
if
ZFC C hH.!2 /; INS ; 2; A W A 2 P .R/ \ L.R/i
is -consistent, then
hH.!2 /; INS ; 2; A W A 2 P .R/ \ L.R/i :
t
u
The discussion of Section 10.1 can also naturally be recast in terms of -logic, for
example we note the following reformulation of Theorem 10.17.
Theorem 10.155. Suppose that there exists a proper class of Woodin cardinals. Suppose that P .R/ is a pointclass closed under continuous preimages such that
L.; R/ ZF C DC C ADR
and such that every set in is universally Baire. There exists a countable set a0 !1
such that
L.;R/ a
0
a0 !1HOD
and such that the following holds where
M D HODL.;R/ a0 :
Suppose that is a 2 -sentence such that
ZFC C
is -consistent. Then there exists a countable set a !1 such that
(1) .INS /M D M \ .INS /M a ,
(2) M a .
t
u
t
u
With the analogous results for the universally Baire sets this theorem can easily be
generalized to yield the following.
813
Theorem 10.157. Assume there exists a proper class of Woodin cardinals and that
A R is universally Baire. Then for each sentence , either
(1) ZFC ` L.A; R/ , or
(2) ZFC ` L.A; R/ :.
t
u
21
814
10 Further results
Theorem 10.158 (WHIH). Suppose there exists a proper class of measurable Woodin
cardinals. Then for each 1 sentence , either
(1) ZFC C CH ` hH.!2 /; R; 2i , or
(2) ZFC C CH ` hH.!2 /; R; 2i :.
t
u
We state two theorems though they are not really optimal. The rst involves the
stationary tower and it is a corollary of a strengthened version of the 21 absoluteness
theorem of .Woodin 1985/, which deals with integer games of length !1 . The second theorem involves weakly homogeneous iteration schemes and the conclusion is
stronger.
Theorem 10.159. Suppose that there is a proper class of measurable Woodin cardinals. Then there is a set A !1 such that
hH.!2 /; 2iLA
where is any 2 sentence such that
ZFC C hH.!2 /; 2i
t
u
is -satisable.
Theorem 10.160. Suppose that there exits a proper class of Woodin cardinals.
Suppose M is a countable transitive model of ZFC and
hE W < i 2 M
is a weakly coherent Doddage in M such that;
(i) is a Woodin cardinal in M ,
(ii) hE W < i M ,
(iii) hE W < i witnesses that is a Woodin cardinal in M ,
(iv) there exists 2 M such that < and such that is a measurable Woodin
cardinal in M .
Q has an iteration scheme in V which is
Suppose .M; E/
Suslin. Then there is a set
A 2 .P .!1 //M
-weakly homogeneously
such that
hH.!2 /; 2iLA
where is any 2 sentence such that
ZFC C hH.!2 /; 2i
is -consistent.
t
u
815
These theorems shows that under fairly general circumstances there cannot be
(nontrivial) consistent 2 sentences for hH.!2 /; 2i which are mutually inconsistent.
The correct version of these theorems is given in the following conjecture:
Conjecture: Suppose there exists a measurable Woodin cardinal and CH holds. Then
there exists A !1 such that for all B !1 , if
A 2 LB
then
Th.LA/ D Th.LB/:
t
u
t
u
Theorem 10.166 supports our claim that the partial order Coll.!1 ; R/ is an analog
of Pmax for CH. The theorem requires the notion of weakly A-good iteration schemes.
Denition 10.162. Suppose that
Q / 2 H.!1 /
.M; E;
and that
(1) M ZFC,
(2) EQ 2 M is a weakly coherent Doddage (in M) witnessing is a Woodin cardinal
in M.
Suppose that A is universally Baire and that M is A-closed.
Q / is weakly A-good if every iterate of M conAn iteration scheme, I , for .M; E;
structed according to I , is A-closed.
t
u
816
10 Further results
817
Theorem 10.166. Suppose that there exists a proper class of Woodin cardinals. Let
be the set of A R such that A is universally Baire. Suppose that 0 is a
pointclass such that:
(i) L.0 ; R/ \ P .R/ D 0 .
(ii) For each A 2 0 there exists
Q / 2 H.!1 /
.M; E;
such that
a) M ZFC,
b) EQ 2 M is a weakly coherent Doddage (in M) witnessing is a Woodin
cardinal in M,
c) in M there is a measurable Woodin cardinal above ,
d) M is A-closed,
Q has an iteration scheme in M which is weakly A-good.
e) .M; E/
0
Suppose that is a 2 sentence in the language for the structure
hH.!2 /; INS ; 2; A W A 2 0 i
and that
Coll.!1 ;R/
ZFC C hH.!2 /; INS ; 2; A W A 2 0 iV
is -consistent. Suppose that G Coll.!1 ; R/ is V -generic. Then
hH.!2 /; INS ; 2; A W A 2 0 iL.0 ;R/G :
t
u
Remark 10.167.
(2) It seems quite likely that if there exists a proper class of measurable Woodin
cardinals then the pointclass of the universally Baire sets necessarily satises
the requirement (ii) of Theorem 10.166.
t
u
We note the following theorem.
Theorem 10.168. Suppose that there exists a proper class of Woodin cardinals and
Q / 2 H.!1 / such that
that A R is universally Baire. Then there exists .M; E;
(1) M ZFC and M is A-closed,
(2) EQ 2 M is a weakly coherent Doddage (in M) witnessing is a Woodin cardinal
in M,
Q has a weakly A-good iteration scheme in M where P .R/ is the
(3) .M; E/
pointclass of all universally Baire sets.
u
t
818
10 Further results
Coll.!1 ;R/
t
u
t
u
819
Theorem 10.171. Suppose that there exists a model hM; Ei such that
hM; Ei ZFC C CH;
and such that for each 2 sentence if there exists a partial order P such that
hH.!2 /; 2iV
VP
P Q
;
, then
hH.!2 /; 2ihM;E i :
Assume there exists a proper class of inaccessible cardinals. Then for all partial orders P ,
V P ADL.R/ :
u
t
Remark 10.172. The proof of Theorem 10.171 uses the core model induction, this
is the machinery used to prove Theorem 5.104 and Theorem 6.149; and the conclusion can be strengthened. A plausible upper bound in the consistency strength of the
hypothesis is an inaccessible limit of Woodin cardinals which are limits of Woodin
cardinals. Of course without the assumption that
hM; Ei CH;
the hypothesis is relatively weak, the upper bound, eliminating the inaccessibles, being
the consistency strength of
1
ZFC C ./ C For all P , V P
2 -Determinacy;
by Theorem 4.69.
t
u
820
10 Further results
(2) There exist a Woodin cardinals 0 < 1 and a weakly coherent Doddage
EQ D hE W < 1 i
which witnesses 1 is a Woodin cardinal such that if 1 < and if is inaccessible then there exists a countable elementary substructure,
X V
containing 0 ; 1 and EQ such that
hM; 0M ; EQM ; 1M i
has a mixed iteration scheme which is 1 -homogeneously Suslin and weakly
A-good for each universally Baire set A 2 X . Here M is the transitive collapse
Q 1 / under the collapsing map. u
t
of X and .0M ; EQM ; 1M / is the image of .0 ; E;
Theorem 10.175. Assume there is a proper class of measurable cardinals which are
limits of Woodin cardinals. Then for each ordinal there exists a transitive inner
model containing the ordinals such that
(1) V N ,
(2) N ZFC C MIHC ,
WH N
WH
(3) .1
/ 1
.
t
u
There are two natural candidates for canonical models of the form, L.P .!1 //, in
the context of CH.
(I) Suppose that 0 is 1 -huge. Suppose that
G0 Coll.!; < 0 /
is V -generic and that
G1 .Coll.!1 ; < 1 //V G0
is V G0 -generic. The rst candidate is
L.P .!1 //V G1 :
(II) Suppose that P .R/ is a pointclass, closed under continuous preimages,
such that
L.; R/ ZF C ADR C is regular:
Suppose that G Coll.!1 ; R/ is L.; R/-generic. The second candidate is
L.P .!1 //L.;R/G :
The rst class of models, or at least the background models V G1 , have two interesting
features. These models were the subject of Theorem 6.28, which shows that a much
stronger version of (1) below actually holds.
821
t
u
822
10 Further results
Denition 10.177 (Strong ADC Conjecture). Suppose that L.A; R/ and L.B; R/
each satisfy ADC . Suppose that every set
X 2 .L.A; R/ [ L.B; R// \ P .R/
is !1 -universally Baire. Then either A 2 L.B; R/ or B 2 L.A; R/.
t
u
Theorem 10.178. Assume there exists a proper class of Woodin cardinals. Let 1 be
the pointclass of all A R such that A is universally Baire and let T D Th.H.!2 //
Then following are equivalent.
(1) There exists a sentence such that
V ZFC C
for some , and such that for each sentence , either
a) ZFC C ` H.!2 / , or
b) ZFC C ` H.!2 / :.
(2) T is is 1 denable (equivalently 1 -denable) in the structure
hM 1 ; 2; Ri:
t
u
Remark 10.179. (1) First order logic is denable in V! and as a result the theory of
V! cannot be nitely axiomatized over ZFC in rst order logic. This of course
is the essence of the incompleteness theorems of Godel.
The key question raised by Theorem 10.178 concerns the intrinsic complexity of
-logic; i. e. of the set:
0 D j ZFC `
for this places a limit on how large a fragment of V one can consistently assert has a theory which is nitely axiomatized over ZFC in -logic. The only
immediate upper bound is V where is the second Woodin cardinal, noting
that Wadge determinacy holds in this case for the sets A R which are homogeneously Suslin for each < . Neeman has proved that if there is a
Woodin cardinal then all universally Baire sets are determined and using this
result, the set 0 is denable in V0 C1 where 0 is the least Woodin cardinal.
The set 0 cannot be dened in H.!1 / and assuming ./ it cannot be dened in
H.!2 /.
(2) Assume there exists a proper class of Woodin cardinals. Let 1 be the pointclass of all A R such that A is universally Baire. Then 0 has the same Turing
degree as the 1 -theory of the structure
hM 1 ; 2; RiI
in fact each is recursively reducible to the other. Thus the complexity of -logic
t
u
is the same as that of the complete 21 . 1 / subset of !.
823
Theorem 10.180 (CH C ADC Conjecture). Assume there exists a proper class of
Woodin cardinals. Let 1 be the pointclass of all A R such that A is universally
Baire. Let T be the 1 theory of hM 1 ; 2; Ri. Then either
(1) T is 2 denable in the structure; hH.!2 /; INS ; 2i; or
(2) T is 2 denable in the structure; hH.!2 /; INS ; 2i:
t
u
There is a version of Theorem 10.180 which is not dependent on the ADC Conjecture. This theorem is proved using the core model induction.
Theorem 10.181. Assume there exists a proper class of Woodin cardinals and that
either Martins Maximum.c/ holds or there is an !1 -dense ideal on !1 .
Let 1 be the pointclass of all A R such that A is universally Baire and let T
t
u
be the 1 theory of hM 1 ; 2; Ri. Then T is denable H.c C /.
As a corollary of Theorem 10.180, using Tarskis theorem on the undenability of
truth, one obtains the rst theorem regarding CH. This theorem shows that the most
optimistic possibility of a version of Pmax for CH must fail.
Theorem 10.182 (ADC Conjecture). Suppose that there exist a proper class of Woodin
cardinals and that is a sentence such that
V
for some strongly inaccessible cardinal, . Suppose that for each sentence , either
(i) ZFC C ` H.!2 / , or
(ii) ZFC C ` H.!2 / :.
t
u
Then CH is false.
The axiom ./ is a natural example of an axiom which axiomatizes the theory of
H.!2 / in -logic. An immediate consequence of ./ is that there exists a surjection
W R ! !2
such that the induced prewellordering,
.x; y/ j .x/ .y/
is 13 .
Let express: There exists a surjection W R ! !2 such that
is 2 -denable in the structure
hH.!2 /; INS ; 2i;
(without parameters),
824
10 Further results
is .21 /L.A;R/ .
The second theorem regarding CH generalizes the fact that ./ implies
12 D !2 .
It is a corollary of Theorem 10.183.
Theorem 10.184 (ADC Conjecture). Suppose that there exist a proper class of Woodin
cardinals and that is a sentence such that
V
for some strongly inaccessible cardinal, . Suppose that for each sentence , either
(i) ZFC C ` H.!2 / , or
(ii) ZFC C ` H.!2 / :.
Then holds.
t
u
825
We note the following theorem which shows that Theorem 10.184 is essentially the
strongest possible. This theorem was independently proved by Neeman.
Theorem 10.185. Assume there exists a proper class of Woodin cardinals. Suppose
that A R is universally Baire, 0 is sentence and
(i) L.A; R/ 0 ,
(ii) for all B 2 L.A; R/ \ P .R/, either L.B; R/ D L.A; R/, or L.B; R/ :0 .
Let A D ./L.A;R/ . Then there exists a sentence such that:
(1) For each sentence , either
a) ZFC C ` H.!2 / , or
b) ZFC C ` H.!2 / :.
(2) ZFC C is -consistent.
t
u
In the fall of 2009, Aspero, Larson, and Moore showed that there are 2 -sentences
1 and 2 such that both 1 and 2 can be forced to hold with CH but .1 ^2 / implies
the :CH. The main question which remains is:
Question. Can there exist a sentence such that for all 2 sentences, , either
(1) ZFC C CH C ` H.!2 / , or
(2) ZFC C CH C ` H.!2 / :;
and such that ZFC C CH C is -consistent?
t
u
Remark 10.186. (1) A natural conjecture is that if the answer to the question above
is yes, then under suitable large cardinal hypotheses, or suitable determinacy
hypotheses, the witness for , is simply the sentence:
H.!2 / 2 H.!2 /V
Coll.!1 ;R/
826
10 Further results
(3) We note the following corollary of Theorem 10.171. Suppose that CH holds in
V and that if is a 2 sentence for which there exists a partial order P such that
hH.!2 /; 2iV
CH C ;
t
u
Finally if the Conjecture is false (and there is a proper class of Woodin cardinals)
then a very interesting question is the following.
Question. Can there exist a sentence such that for all either
(1) ZFC C CH C H.!2 / , or
(2) ZFC C CH C H.!2 / :;
and such that ZFC C CH C is -satisable?
t
u
The next theorem, in conjunction with Theorem 10.184, shows that this question must have a negative answer in any sufciently iterable model provided that
ZFC C CH C is -consistent in V .
Theorem 10.187. Suppose that there exists a proper class of Woodin cardinals and
that
Q / 2 H.!1 /
.M; E;
is such that:
(i) M is transitive and
M ZFC C There exists a proper class of Woodin cardinals:
(ii) M is a Woodin cardinal.
(iii) E 2 M and in M is a weakly coherent Doddage witnessing that is a Woodin
cardinal.
Q has an iteration scheme which is universally Baire.
(iv) .M; E/
Suppose that T 2 M is a theory containing ZFC, is a sentence and that
M T :
Then T ` .
t
u
827
t
u
Denition 10.189. Axiom ./CC : There exists a pointclass P .R/ and a lter
g Pmax such that
(1) L.; R/ ADC ,
(2) g is L.; R/-generic,
(3) P .R/ L.; R/g.
t
u
Then cof./ D !1 .
Assuming Conality Conjecture an argument using the core model induction and
which is quite similar to the proof of Theorem 10.181 yields the following theorem and
it is a corollary of this theorem that the two axioms, ./C and ./CC , are equivalent.
The statement of the theorem involves the following notation for each set A R,
./A abbreviates:
(1) L.A; R/ ADC ,
(2) L.P .!1 /; A/ D L.A; R/G, for some L.A; R/-generic lter G Pmax .
Theorem 10.191 (Conality Conjecture). Let be the pointclass of sets A R such
that ./A holds and suppose that A and B are in . Then
L.A; B; R/ ADC :
t
u
828
10 Further results
Related to the problem of Martins Maximum vs. ./ is the following question: Is
ZFC C Martins Maximum C ./CC
consistent?
A simpler question concerns the value of
12 in V g where g is V -generic for
Namba forcing. Note that if
1 D !2
2
then necessarily,
.12 /V < .12 /V g :
A bound for .12 /V g is provided by the following theorem.
Theorem 10.192. Assume that for all A !2 , A exists. Suppose that g is V -generic
for Namba forcing. Then in V g:
(1) For all x 2 R, x # exists;
(2) 12 !3V .
Proof. We sketch the proof. Let P be the Namba partial order.
The elements of P are pairs .s; t / such that
(1.1) t !2<! and t is closed under initial segments,
(1.2) s 2 t ,
(1.3) for all a 2 t if s a then
j < !2 j a _ 2 t j D !2 :
The order on P is dened in the natural fashion:
.s ; t / .s; t /
if s s and t t .
As usual we identify the generic g with the corresponding function
[s j .s; t / 2 g;
which is a conal function from ! to !2V .
Fix x 2 RV g and let P be a term for x. By the usual fusion arguments there
exist a condition .s; t / 2 g and a function
W t ! ! <!
such that for all a 2 t and for all b 2 t ,
(2.1) .a/ 2 ! dom.a/ ,
(2.2) if a b then .a/ .b/
829
pT D .y; y # / j y 2 RLAg :
As usual we regard the innite branches of T as triples .y; z; f / where y 2 ! ! ,
z 2 ! ! and f 2 ! .
Thus working in LA there exists a condition
.s ; t / .s; t /
in PA and a function
W t ! T
830
10 Further results
In summary we have proved that in V g, for all x 2 RV g , x # exists and that !3V
is an indiscernible of Lx. However
!1V g D !1V
and so it follows that .12 /V g !3V .
t
u
Theorem 10.193 (Axiom ./CC ). Suppose that for each A !2 , A exists. Suppose
that g is V -generic for Namba forcing. Then
.12 /V g D .!3 /V :
Proof. Clearly
!3 D ./L.;R/ :
Recall that AG denotes the set [a j h.M; I /; ai 2 G. By modifying G if necessary
we can suppose that
!1 D .!1 /LAG
and so, since AC holds, there exists a surjection
W !2 ! R
which is 1 denable in the structure
hH.!2 /; INS ; 2i
from AG .
Fix an ordinal 2 !3 n !2 and x a set B 2 such that B codes a a surjection
W R ! :
Let F G be the set of
h.M; I /; ai 2 G
such that
(1.1) M ZFC C ./,
(1.2) B \ M 2 M,
(1.3) .M; I / is B-iterable,
(1.4) hV!C1 \ M; B \ M; 2i hV!C1 ; B; 2i.
By Lemma 4.52 and Lemma 4.56, F is dense in G.
Suppose that h.M; I /; ai 2 F and let
jG W .M; I / ! .M ; I /
be the (unique) iteration such that
jG .a/ D AG :
It follows from (1.1)(1.4) that
hV!C1 \ M ; 2i hV!C1 ; 2i
831
Then s 2 M where
jG W .M; I / ! .M ; I /
t
u
832
10 Further results
t
u
Version II: Suppose that X !1 . Then there exists a pointclass P .R/ such that
L.; R/ ZF C DC C ADR ,
there is a lter g Coll.!1 ; R/ such that
(1) g is L.; R/-generic,
(2) X 2 L.; R/g.
t
u
Version III: Suppose that X !1 . Then there exists a pointclass P .R/ such
that
L.; R/ ZF C DC C ADR ,
there is a lter g Coll.!1 ; R/ such that
(1) g is L.; R/-generic,
(2) X 2 L.; R/g,
(3) .INS /L.;R/g D INS \ L.; R/g.
t
u
Remark 10.194. Clearly, assuming CH, (Version I) and (Version II) are each expressible by a 2 sentence in the structure
hH.!2 /; 2i:
Further (Version III) is expressible by a 2 sentence in the structure
hH.!2 /; INS ; 2i:
It is a corollary of the proof of Theorem 10.180 that (Version III) cannot be implied by
CH in -logic. In fact one can show that the sentence
There exists a partial order P such that RV D .R/V
V
cannot be a validity of -logic.
(Version III)
t
u
833
t
u
834
10 Further results
Theorem 10.197 (Long Game Conjecture, CH). Assume there exists a proper class of
measurable Woodin cardinals. Suppose that X !1 . Then there exists a set A R
such that
(1) L.A; R/ ADC ,
(2) there is a lter g Coll.!1 ; R/ such that
a) g is L.A; R/-generic,
b) X 2 L.A; R/g.
t
u
We now consider the problem of obtaining (Version II) from CH. This is closely
related to the question concerning CH, listed at the end of the previous section:
Does there exist a sentence such that
ZFC C CH C
is -consistent and such that for all 2 sentences, , either
ZFC C CH C ` H.!2 / , or
ZFC C CH C ` H.!2 / :?
It is convenient to dene a slight strengthening of (Version II).
(Version II)C : Suppose that X !1 and that A R is universally Baire. Then there
exists a pointclass P .R/ such that
A 2 ,
L.; R/ ZF C DC C ADR ,
there is a lter g Coll.!1 ; R/ such that
(1) g is L.; R/-generic,
(2) X 2 L.; R/g.
t
u
We remark that the assumptions (i)(iii) of Theorem 10.198 should hold in any
ne structural inner model in which there exists a proper class of measurable Woodin
cardinals. Further it seems quite plausible that (i), (iii) and a sufcient fragment of (ii)
are provable consequences of the existence of a proper class of measurable Woodin
cardinals; i. e. that the stronger theorem, obtained by eliminating the assumptions (i)
(iii), is actually true.
Theorem 10.198. Assume there exists a proper class of measurable Woodin cardinals
and that:
(i) Long Game Conjecture holds;
(ii) WHIHC holds;
835
Coll.!1 ;R/
V Coll.!1 ;R/
` H.!2 / , or
` H.!2 / :.
t
u
Suppose that LE is a MitchellSteel inner model with a superstrong cardinal, and
a proper class of Woodin cardinals, in which the countable initial segments of LE are
-iterable for every . Then one can show that in LE, the 2 theory of H.!2 / is not
nitely axiomatized over ZFC in -logic. With additional assumptions one can also
show that in LE, (Version II) must fail.
Thus any attempt to prove (Version II) from CH would seem to require large cardinals beyond superstrong.
836
10 Further results
t
u
Remark 10.200. There are two natural variations of the Effective Singular Cardinals
Hypothesis:
(1) One could require that GCH holds below
, or
(2) that the Effective Generalized Continuum Hypothesis holds below
.
The Effective Generalized Continuum Hypothesis is the obvious variation of Effective
Singular Cardinals Hypothesis:
Suppose that
is an innite cardinal and that g Coll.!;
/ is V -generic.
Suppose that
M V g
is a transitive inner model such that in V g:
R M;
M ZF C AD;
Every set A 2 P .R/ \ M is 1 -homogeneously Suslin.
Then M < .
CC /V .
t
u
We give a brief summary of a few relevant results and which are proved in Chapter 7
of .Woodin 2010b/. These results are primarily concerned with the following related
problem. Suppose
is a singular strong limit cardinal and that
2 V Coll.!;/
1
837
(1) ! Coll.!;
/,
(2) p 2 Coll.!;
/,
(3) p x 2 ,
where x 2 V Coll.!;/ is the term for a subset of ! given by ;
n 2 x D _q 2 Coll.!;
/ j .n; q/ 2 :
t
u
838
10 Further results
t
u
t
u
839
yields a structure theory for L.VC1 / which in many aspects is analogous to the structure theory for L.R/ in the context of ADL.R/ . Note that by Kunens theorem on the
nonexistence of an elementary embedding of V to V ,
must be the ! th element of the
critical sequence of j .
The next theorem shows that from this hypothesis one obtains a weak failure of
the Effective Singular Cardinals Hypothesis. The proof of this theorem and of related
theorems can be found in Chapter 7 of .Woodin 2010b/.
t
u
Theorem 10.205. Suppose that there exists an elementary embedding
j W L.VC1 / ! L.VC1 /
with critical point below
and that g Coll.!;
/ is V -generic. Then in V g there
exists a transitive inner model
M L.VC1 /g
such that
(1) RV g M ,
(2) M ZF C ADC ,
(3) .
CC /L.VC1 / < M .
t
u
Remark 10.206. It is a natural conjecture that the inner model M of Theorem 10.205
can be chosen such that
./L.VC1 /g D M :
It is immediate that ./L.VC1 /g is simply the least ordinal such that in L.VC1 /,
is not the surjective image of VC1 . We denote this ordinal by L.VC1 / , this is the
natural generalization of L.R/ to L.VC1 /.
This in turn suggests the following problem. Suppose there exists an elementary
embedding
j W L.VC1 / ! L.VC1 /
with critical point below
. Must
L.VC1 / <
CC
t
u
Chapter 11
Questions
The following is a list of questions, including many which have appeared in earlier
chapters. The order simply reects roughly the place within the book where the question is discussed, either explicitly or implicitly, and there is signicant overlap among
various of these questions. Comments have been asserted in italics for those questions
which either have been solved or otherwise affected by developments of which I am
aware since the rst edition.
(1) Assume L.R/ AD. Must L.R/ !3 ?
(2) Can there exist countable transitive models M and M such that
M ZFC C The nonstationary ideal on !1 is saturated;
2
N/
(5) Suppose that INS is !2 -saturated and that P .!1 /# exists. Suppose that A !1
and let
A D sup.!2 /LZ j Z !1 ; A 2 LZ; and RLA D RLZ :
Must A < !2 ?
11 Questions
841
(6) Assume there exists a proper class of Woodin cardinals. Do either of the following imply :CH?
a) Every function f W !1 ! !1 is bounded on a closed conal subset of !1
by a canonical function.
b) Suppose that A R is universally Baire and that
f W !1 ! A:
Then there exists a tree T on ! !1 such that that A D pT and such that
< !1 j f ./ 2 pT j
contains a closed conal subset of !1 .
Solved by Larson and Shelah: The answer is no.
(7) Suppose the nonstationary ideal on !1 is !1 -dense.
a) Must c D !2 ?
b) Must
12 D !2 ?
c) Must L.R/ !3 ?
(8) Assume Martins Maximum.c/. Suppose that P .R/ is a pointclass, closed
under continuous preimages, such that
a) L.; R/ ADC ,
b) !3 D ./L.;R/ .
Suppose that G Pmax is an L.; R/-generic lter such that G 2 V and such
that
P .!1 / D P .!1 /G :
Must
L.; R/G ADR
(9) Assume ./. Suppose
W !1 2 ! 0;1
is a partition with no homogeneous rectangle for 0 of (proper) cardinality @1 .
Must there exist a set X !1 such that E
.3/ X is nonstationary in !1 ?
Justin Moore has proved that the associated partition relation is false .Moore
2006/, but the status of this question remains unclear.
(10) Assume .
a) Must INS be semi-saturated?
b) Must HODR AD?
842
(11)
11 Questions
?
c) Can L.P .!! // ?
11 Questions
843
(16) (Conjectures)
a) There exists a regular (uncountable) cardinal and a denable partition of
< j cof./ D !
into innitely many stationary sets.
b) Suppose that there is a proper class of supercompact cardinals. Then (a)
holds.
c) Assume Martins Maximum. There is a denable wellordering of the reals.
These conjectures are all implied by the HOD-Conjecture of .Woodin 2010b/
where a number of relevant results are proved.
(17) Suppose that P .R/ is a pointclass closed under continuous preimages such
that
L.; R/ ADC
and let M D .HOD/L.;R/ .
Suppose that a !1 is a countable set such that
M a ./:
Must .!1 /
< .!1 /
M a
844
11 Questions
(21) Suppose that 1 and 2 are 2 sentences (in the language for the structure
hH.!2 /; INS ; 2i) such that both
ZFC C CH C hH.!2 /; INS ; 2i 1
and
ZFC C CH C hH.!2 /; INS ; 2i 2
are -consistent. Let D .1 ^ 2 /. Is
ZFC C CH C hH.!2 /; INS ; 2i
-consistent?
Solved by Aspero, Larson, and Moore in fall, 2009: The answer is no.
(22) Can there exist a sentence such that for all 2 sentences, , either
ZFC C CH C ` H.!2 / , or
ZFC C CH C ` H.!2 / :;
and such that ZFC C CH C is -consistent? The reformulation with CH C
replaced by either generic- or , is also open and discussed in Woodin .2003/.
(23) (Conjecture) Assume there exists a proper class of Woodin cardinals. Let be a
2 sentence. Then the following are equivalent.
a) ZFC C is -consistent.
b) There exists a partial order P such that V P .
This is the Conjecture.
(24) Are the following mutually consistent?
a) .ZF C DC/ There exists a cardinal such that for every cardinal
, there
exists an elementary embedding
j WV !V
with cp.j / D and j. / >
.
b) .ZF C DC/ For all x 2 R, x is OD if and only if for some A 2 1 , x is
OD in L.A; R/.
(25) Assume there exists an elementary embedding
j W L.VC1 / ! L.VC1 /
with critical point below
. Dene L.VC1 / to be;
sup 2 Ord j there exists a surjection, f W P .
/ ! , with f 2 L.VC1 /:
Must
L.VC1 / <
CC
Bibliography
846
Bibliography
Jech, T. and W. Mitchell (1983). Some examples of precipitous ideals. Ann. Pure
Appl. Logic 24(2), 99212.
Kanamori, A. (2008). The higher innite. Perspectives in Mathematical Logic. Berlin:
Springer-Verlag. Large cardinals in set theory from their beginnings; second edition.
Kechris, A., D. A. Martin, and R. Solovay (1983). An introduction to Q theory.
In Cabal Seminar 7981, Volume 1019 of Lecture Notes in Mathematics. SpringerVerlag.
Ketchersid, R., P. Larson, and J. Zapletal (2007). Increasing
12 by Nambia-style
forcing. JSL 72, 13721378.
Koellner, P. and W. H. Woodin (2010). Large cardinals from determinacy. In M. Foreman and A. Kanamori (Eds.), Handbook of Set Theory-volume 3, Volume XIV,
pp. 19512120. New York: Springer-Verlag.
Larson, P. (2004). The Stationary Tower: Notes on a Course by W. Hugh Woodin.
University Lecture Series (American Mathematical Society). Oxford, U.K.: Oxford
University Press.
Laver, R. (1976). On the consistency of Borels conjecture. Acta Math. 137, 151169.
Law, D. (1994). An abstract condensation property. Ph. D. thesis, Caltech.
Levy, A. and R. Solovay (1967). Measurable cardinals and the continuum hypothesis.
Israel J. Math. 5, 234248.
Martin, D. A. and J. Steel (1983). The extent of scales in L.R/. In Cabal Seminar
7981, Volume 1019 of Lecture Notes in Mathematics. Springer-Verlag.
Martin, D. A. and J. Steel (1989). A proof of projective determinacy. J. Amer. Math.
Soc. 2, 71125.
Martin, D. A. and J. R. Steel (1994). Iteration trees. J. Amer. Math. Soc. 7(1), 173.
Mitchell, W. J. and J. R. Steel (1994). Fine structure and iteration trees. Berlin:
Springer-Verlag.
Moore, J. (2006). A solution to the L space problem. J. Amer. Math. Soc. 19(3),
717736.
Moschovakis, Y. N. (1980). Descriptive set theory. Amsterdam: North-Holland Publishing Co.
Neeman, I. (2004). The determinacy of long games, Volume 7 of de Gruyter Series
in Logic and its Applications. Berlin: Walter de Gruyter & Co.
Ostaszewski, A. J. (1975). On countably compact perfectly normal spaces. Journal
of the London Mathematical Society 14(2), 505516.
Bibliography
847
Index
|NS , 578
CC
|NS , 579
Changs ConjectureC , 667
closed set (general), 34
closed, unbounded (general), 34
M|NS , 511
|
M0 NS , 569
coding elements of H.!1 /, 21
coding elements of H.c C /, 21
850
Index
I< , 199
I _ S , 288
indecomposable ultralter, 422
1
-borel set A, 610
INS , 2
iterable structure, 53
iteration of a structure,
h.Nk ; Jk / W k < !i, 119
iteration of a structure,
hNk W k < !i, 124
iteration of a structure,
h.Mk ; Ik / W k < !i, 201
iteration (full), 205
iteration scheme, 695
iteration scheme (mixed), 697
iterations by stationary tower, 200
iteration of a structure,
.M; I; a/, 511
iteration of a structure, .M; I/, 116
IU;F , 505
JNS , 683
Long Game Conjecture, 833
MIH, 698
MIHC , 820
mixed iteration scheme, 697
Martins Maximum, 38
Martins MaximumZF .c/, 40
Martins MaximumC , 38
Martins MaximumC .c/, 39
Martins MaximumCC , 39
Martins MaximumCC .c/, 39
M -normal ultralter, 124
M , 29
M3 .a/, 262
M! .a/, 804
nonregular ultralter, 421
N , 29
!1 -dense ideals
and Suslin trees, 331
-Conjecture, 810
-logic, 807
!1 -dense ideal, 306
-proof, 809
!-presaturated ideal, 683
PF , 578
PFA; Proper Forcing Axiom, 37
AC , 185
BC , 487
, 398
C
, 400
S , 426
C
S , 427
Pmax , 136
0
Pmax
, 233
|
NS
Pmax
, 508
2
Pmax , 290
, 221
Pmax
(T)
Pmax
, 207
PNS , 288
pointclass, 22
projection for measures, 24
proper partial order, 36
Proper Forcing Axiom; PFA, 37
weakly proper partial order, 753
AC , 193
AC .I /, 441
AC , 221
P< , 35
PU , for ultralters on !, 476
PU , for ultralters on !1 , 501
.;/
, 739
Pmax
.;;B/
Pmax
, 773
QF , 581
Qmax , 307
KT
Qmax , 384
KT
Qmax , 391
M
Qmax , 408
2
Qmax , 371
Q
max , 708
Qmax , 334
Q< , 35
Q3 .X /, 172
Index
1 ./, 31
2
1 .< -WH/, 31
2 1
1 . -WH/, 31
Smax , 428
Suslin cardinal, 613
Suslin sets of reals, 22
SPFA; Semiproper Forcing Axiom, 38
SRP.!2 /, 652
SRP .!2 /, 662
stationary set (general), 34
stationary subset (general), 34
stationary tower, 35
SQ , 184
Strong ADC Conjecture, 822
NS
Umax
, 557
universal function, F , 598
universally Baire set A, 795
U -restricted 2 formula, 363
851
852
Index
YA .F; I /, 404
YA .F; /, 418
YBC .I /, 451
YBC .I /, 451
Y(Code) .S; z/, 704
YColl .I /, 307
ZBC .I /, 450
ZFC , 52
ZFC , 404
Zh;F , 505
Zp;F , 505