4

3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


STRONG MEASURE ZERO SETS
WITHOUT COHEN REALS
January 1991
Martin Goldstern
1
Bar Ilan University
Haim Judah
1
Bar Ilan University and U.C. Chile
Saharon Shelah
1,2
Hebrew University of Jerusalem
ABSTRACT. If ZFC is consistent, then each of the following
are consistent with ZFC + 2

0
=
2
:
1. X IR is of strong measure zero i [X[
1
+ there
is a generalized Sierpinski set.
2. The union of
1
many strong measure zero sets is a
strong measure zero set + there is a strong measure
zero set of size
2
.
1
The authors thank the Israel Foundation for Basic Research, Israel Academy of Science.
2
Publication 438
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
0. Introduction
In this paper we continue the study of the structure of strong measure zero sets. Strong
measure zero sets have been studied from the beginning of this century. They were dis-
covered by E. Borel, and Luzin, Sierpinski, Rothberger and others turned their attention
to the structure of these sets and proved very interesting mathematical theorems about
them. Most of the constructions of strong measure zero sets involve Luzin sets, which
have a strong connection with Cohen reals (see [6]). In this paper we will show that this
connection is only apparent; namely, we will build models where there are strong measure
zero sets of size c without adding Cohen reals over the ground model.
Throughout this work we will investigate questions about strong measure zero sets under
the assumption that c = 2

0
=
2
. The reason is that CH makes many of the questions
we investigate trivial, and there is no good technology available to deal with most of our
problems when 2

0
>
2
.
0.1 Denition: A set X IR of reals has strong measure zero if for every sequence

i
: i < ) of positive real numbers there is a sequence x
i
: i < ) of real numbers such
that
X

i<
(x
i

i
, x
i
+
i
)
We let o P(IR) be the ideal of strong measure zero sets.
0.2 Remark: (a) if we work in

2 then X

2 has strong measure zero if
(h

)(g

n
h(n)
2)(x X)(

n)(g(n) = x[h(n))
or equivalently,
() (h

)(g

n
h(n)
2)(x X)(n)(g(n) = x[h(n))
(b) To every question about strong measure zero sets in IR there is a corresponding ques-
tion about a strong measure zero set of

2, and for all the questions we consider the
corresponding answers are the same. So we will work sometimes in IR, sometimes in

2.
0.3 Denition: Assume that H

. We say that has index H, if =
h
: h H)
and for all h H,
h
is a function with domain and n
h
(n)
h(n)
2. We let
X

:=

hH

k
[
h
(k)]
(where we let [] := f

2 : f).
We say that X

is the set dened by .
438 1 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
0.4 Fact: Assume H

is a dominating family, i.e., for all f

there is h H such
that n f(n) < h(n). Then:
(1) If has index H, then X

is a strong measure zero set.
(2) If X is a strong measure zero set, then there is a sequence with index H
such that X X

.
0.5 Denition: A set of reals X IR is a GLuzin (generalized Luzin) set if for every
meager set M IR, X M has cardinality less than c. X is a generalized Sierpinski set
if set if for every set M IR of Lebesgue measure 0, X M has cardinality less than c.
0.6 Fact: (a) If c is regular, and X is GLuzin, then X has strong measure zero.
(b) A set of mutually independent Cohen reals over a model M is a GLuzin set.
(c) If c >
1
is regular, and X is a GLuzin set, then X contains Cohen reals over L.
Proof: See [6].
0.7 Theorem: [6] Con(ZF) implies Con(ZFC + there is a GLuzin set which is not strong
measure zero).
0.8 Theorem: [6] Con(ZF) implies Con(ZFC + c >
1
+ X [IR]
c
, X a strong measure
zero set + there are no GLuzin sets).
In theorem 0.16 we will show a stronger form of 0.8.
0.9 Denition:
(1) Let Unif(o) be the following statement: Every set of reals of size less
than c is a strong measure zero set.
(2) We say that the ideal of strong measure zero sets is c-additive, or Add(o),
if for every < c the union of many strong measure zero sets is a strong
measure zero set. (So Add(o) Unif(o).)
0.10 Remark: Rothberger ([13] and [12]) proved that the following are equivalent:
(i) Unif (o)
(ii) for every h : , for every F [

n
h(n)]
<c
, there exists f

such
that for every f F there are innitely many n satisfying f(n) = f

(n).
Miller ([10]) noted that this implies the following:
Add(/) i Unif(o) and b = c
(See 0.17 for denitions)
Rothberger proved interesting results about the existence of strong measure zero sets,
namely:
If b =
1
, then there is a strong measure zero set of size
1
. (See [5].)
In this spirit, we rst prove the following result:
438 2 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
0.11 Theorem: If Unif(o) and d = c, then there exists a strong measure zero set of size
c.
We start the proof by proving the following
0.12 Fact: If d = c, then there is a set f
i
: i < c of functions in

such that for every
g

, the set
i < c : f
i

g
has cardinality less than c.
Proof of the fact: We build f
i
: i < c) by transnite induction. Let

= g
j
: j < c.
We will ensure that for j < i, f
i
,<

g
j
. This will be sucient.
But this is easy to achieve, as for any i, the family g
j
: j < i is not dominating, so there
exists a function f
i
such that for all j < i, for innitely many n, f
i
(n) > g
j
(n).
This completes the proof of 0.12.
0.13 Proof of 0.11: Using d = c, let f
i
: i < c) be a sequence as in 0.12. Let F :


[0, 1] Q be a homeomorphism. (Q is the set of rational numbers.) We will show that
X := F(f
i
) : i < c is a strong measure zero set.
Let
n
: n < ) be a sequence of positive numbers. Let r
n
: n = Q. Then
U
1
:=

n
(r
n

2n
, r
n
+
2n
) is an open set. So K := [0, 1] U
1
is closed, hence compact.
As K rng(F), also F
1
(K)

is a compact set. So for all n the projection of F
1
(K)
to the nth coordinate is a compact (hence bounded) subset of , say g(n). So
F
1
K f

: f

g
Let Y := X U
1
K. Then Y F(F
1
(K)) F(f
i
) : f
i

g is (by assumption on
f
i
: i < c)) a set of size < c, hence has strong measure zero. So there exists a sequence of
real numbers x
n
: n < ) such that Y U
2
, where
U
2
:=

n
(x
n

2n+1
, x
n
+
2n+1
)
and X U
1
U
2
. So X is indeed a strong measure zero set.
In section 2 we will build models where Add(o) holds and the continuum is bigger than

1
without adding Cohen reals. First we will show in 3.4:
0.14 Theorem: If ZFC is consistent, then
ZFC + c =
2
+ o = [IR]

1
+ there are no Cohen reals over L
is consistent.
Note that c =
2
and o = [IR]

1
implies
(1) Add(o). (Trivially)
(2) b = d =
1
. (By 0.11)
438 3 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
The same result was previously obtained by Corazza[3]. In his model the nonexistence of
strong measure zero sets of size c is shown by proving that every set of size c can be mapped
uniformly continuously onto the unit interval (which is impossible for a strong measure
zero set). Thus, the question arises whether is possible to get a model of o = [IR]
<c
+
c =
2
+ not all set of size c can be continuously mapped onto [0, 1].
By adding random reals to our construction, we answer this question in the following
stronger theorem:
0.15 Theorem: If ZFC is consistent, then
ZFC + c =
2
+ o = [IR]

1
+ there are no Cohen reals over L
+ there is a generalized Sierpinski set
is consistent. (See 0.5.)
By a remark of Miller [8, 2] a generalized Sierpinski set cannot be mapped continuously
onto [0, 1] (not even with a Borel function).
Pawlikowski [11] showed that the additivity of the ideal o of strong measure zero sets does
not imply the additivity of the ideal / of meager sets. For this he built a model satisfying
Add(o) + c =
2
+ b =
1
. He used a nite support iteration of length
2
. So he adds
many Cohen reals, and in the nal model Cov(/) holds (i.e., IR can not be written as the
union of less than c many meager sets). We will improve his result in the next theorem:
0.16 Theorem: If ZFC is consistent, then
ZFC + c = d =
2
> b + Add(o) + no real is Cohen over L
is consistent.
(Note that by 0.11, d = c + Add(o) implies that there is a strong measure zero set of
size c.)
0.17 Notation: We use standard set-theoretical notation. We identify natural numbers
n with their set of predecessors, n = 0, . . . , n 1.
A
B is the set of functions from A into
B, A
<
:=

n<
n
A. [A[ denotes the cardinality of a set A. P(A) is the power set of a
set A, A B means A B&A ,= B. A B is the set-theoretic dierence of A and B.
[A]

:= X A : [X[ = . [A]
<
and [A]

are dened similarly. (We write A := B or

B =: A to mean: the expression B denes the term (or constant) A.)
Ord is the set of ordinals. cf() is the conality of an ordinal . S

:=

: cf() =

. In particular, S
1
2
is the set of all ordinals <
2
of uncountable conality.
IR is the set of real numbers. c = [IR[ is the size of the continuum. For f, g

we let
f < g i for all n f(n) < g(n), and f <

g if for some n
0
, n n
0
f(n) < g(n). The
bounding number b and the dominating number d are dened as
b :=min[H[ : H

, g

h H (h <

g)
d :=min[H[ : H

, g

h H g < h
=min[H[ : H

, g

h H g <

h
(It is easy to see
1
b d c.)
438 4 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
We call a set H

dominating, if g

h H g < h.
/ is the ideal of meager subsets of IR (or of

2). o is the ideal of strong measure zero sets.
For any ideal P(IR), Add() abbreviates the statement: The union of less than c
many sets in is in . Cov() means that the reals cannot be covered by less than c
many sets in .
If f is a function, dom(f) is the domain of f, and rng(f) is the range of f. For A dom(f),
f[A is the restriction of f to A. For 2
<
, [] := f

2 : f.
0.18 More Notation: If Q is a forcing notion, G
Q
is the canonical name for the generic
lter on Q. We interpret p q as q is stronger (or has more information) than p. (So
p q q[ p G
Q
.)
When we deal with a (countable support) iteration P

, Q

: < ), we write G

for the
canonical name of the generic lter on , and G() for the generic lter on Q

. If there
is a natural way to associate a generic real to the generic lter on Q

, we write g

for
the real given by G(). We write [

for the forcing relation of P

. If < , G

always
stands for G

. V = V
0
is the ground model, and V

= V [G

]. P

is the countable
support limit of P

: < ). P

/G

is the P

-name for p P

: p[ G

(with the
same relation as P

). The forcing relation with respect to P

/G

(in V

) is denoted by
[

.
There is a natural dense embedding from P

into P

/G

. Thus we always identify

P

-names for P

/G

-names with the corresponding P

-names.

is the weakest condition of P

, and

is usually abbreviated to [

. (So
[

([

) i [

).
0.19 Even more Notation: The following notation is used when we deal with trees of
nite sequences:
For V
<
, i V ,

i is the function [[, i) V

<
.
p
<
is a tree if p ,= , and for all p, all k < [[, [k p. Elements of a tree are
often called nodes. We call [[ the length of . We reserve the word height for
the notion dened in 2.2.
For p
<
, p, we let succ
p
() := i :

i p.
If p is a tree, p, let p
[]
:= p : or .
If p
<
is a tree, b p is called a branch, if b is a maximal subset of p that is linearly
ordered by .
Clearly, if p succ
p
() ,= , then a subset b p is a branch i b is of the form
b = f[n : n for some f

.
We let stem(p) be the intersection of all branches of p.
438 5 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
1. A few well known facts
We collect a few more or less well known facts about forcing, for later reference.
1.1 Denition: An ultralter | on is called a P-Point, if for any sequence A
n
: n )
of sets in | there is a set A in | that is almost contained in every A
n
(i.e., n AA
n
is
nite).
1.2 Denition: For any ultralter | on , we dene the P-point game G(|) as follows:
There are two players, IN and NOTIN. The game consists
of many moves.
In the n-th move, player NOTIN picks a set A
n
|, and player
IN picks a nite set a
n
A
n
.
Player IN wins if after many moves,

n
a
n
|.
We write a play (or run) of G(|) as
A
0
; a
0
A
1
; a
1
A
2
; . . . ).
It is well known that an ultralter | is a P-point i player NOTIN does not have a winning
strategy in the P-point game.
For the sake of completeness, we give a proof of the nontrivial implication (which is
all we will need later):
Let | be a P-point, and let be a strategy for player NOTIN. We will construct a run of
the game in which player NOTIN followed , but IN won.
Let A
0
be the rst move according to . For each n, let /
n
be the set of all responses of
player notin according to in an initial segment of a play of length n in which player
IN has played only subsets of n:
/
n
:= A
k
: k n, A
0
; a
0
A
1
; . . . ; a
k1
A
k
) is an
initial segment of a play in which NOTIN
obeyed , and a
0
, . . . , a
k1
n
Note that /
0
= A
0
, and for all n, /
n
is a nite subset of |.
As | is a P-point, there is a set X | such that for all A

n
/
n
, X A is nite.
Let X A
0
n
0
, and for k > 0 let n
k
satisfy
n
k
> n
k1
and A /
n
k1
X A n
k
Either

k
[n
2k
, n
2k+1
) |, or

k
[n
2k+1
, n
2k+2
) |.
Without loss of generality we assume

k
[n
2k
, n
2k+1
) |.
Now dene a play A
0
; a
0
A
1
; a
1
A
2
; . . . ) of the game G(|) by induction as follows:
A
0
is given.
Given A
j
, let a
j
:= A
j
[n
2j
, n
2j+1
) and let A
j+1
be s response to a
j
.
438 6 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
Then as a
0
, . . . , a
j1
n
2j
, we have X A
j
n
2j
for all j. Therefore for all j we
have X [n
2j
, n
2j+1
) (A
j
n
2j
) [n
2j
, n
2j+1
) = A
j
[n
2j
, n
2j+1
) = a
j
. So

j
a
j

X

j
[n
2j
, n
2j+1
) |.
Thus player IN wins the play A
0
; a
0
A
1
; a
1
A
2
; . . .) in which player NOTIN obeyed
.
1.3 Denition: We say that a forcing notion Q preserves P-points, if for every P-point
ultralter | on , [
Q
| generates an ultralter, i.e. [
Q
x P() u | (u
x or u x).
[9] dened the following forcing notion:
1.4 Denition: Rational perfect set forcing, RP is dened as the set of trees p
<
satisfying
(1) for all p, [succ
p
()[ 1,
0
(See 0.19)
(2) for all p there is p with and [succ
p
()[ =
0
.
We let p q i p q.
Then the following hold:
1.5 Lemma:
(1) RP preserves P-points. ([9, 4.1])
(2) RP adds an unbounded function. ([9, 2])
(3) RP is proper. (This is implicit in [9]. See also 2.16)
The next lemma can be found, e.g., in [7, VII ?? and Exercise H2]:
1.6 Fact: If Q is a forcing notion satisfying the
2
-cc, then
(1) If [
Q
c

:
V
2

V
2
, then there is a function c :
2

2
such that
[
Q
<
2
: c

() < c().
(2) [
Q

V
2
=
2
.
(3) For every stationary S
2
, [
Q
S is stationary on
2
.
The following fact is from [14, V 4.4]:
1.7 Fact: Assume P

, Q

: <
2
) is an iteration of proper forcing notions Q

. Then
for every
2
of conality > , [

<
V

, or in other words: no new

reals appear in limit stages of conality > .
As a consequence, [

2
If X

, [X[
1
, then there is <
2
such that X V

.
We also recall the following facts about iteration of proper forcing notions:
1.8 Lemma: Assume CH, and let P

, Q

: <
2
) be a countable support iteration
such that for all <
2
, [

is a proper forcing notion of size c.

Then
(1) <
2
: [

c =
1
. (see [14, III 4.1])
438 7 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
(2) [

2
c
2
. (This follows from 1.7 and (1))
(3) For all
2
, P

is proper [14, III 3.2] and satises the

2
-cc. (See [14,
III 4.1])
(4) [

V
1
=
1
. (See [14, III 1.6])
In [2, 4.1] the following is proved:
1.9 Lemma: Assume P

, Q

: <
2
) is as in 1.8, and for all <
2
:
[

preserves P-points.
Then for all
2
, P

preserves P-points.
1.10 Denition: We say that a forcing notion Q is

-bounding, if the set of old
functions is a dominating family in the generic extension by Q, or equivalently,
[
Q
f

g

V nf(n) < g(n)
[14, V 4.3] proves:
1.11 Lemma: Assume P

, Q

: <
2
) is as in 1.8, and for all <
2
:
[

is

-bounding and -proper.
Then for all
2
, P

is

-bounding.
(We may even replace -proper by proper, see [14], [4])
The following is trivial to check:
1.12 Fact: Assume Q is a forcing notion that preserves P-points or is

-bounding. Then
[
Q
There are no Cohen reals over V
1.13 Denition: A forcing notion P is strongly

-bounding, if there is a sequence

n
: n ) of binary reexive relations on P such that for all n :
(1) p
n
q p q.
(2) p
n+1
q p
n
q.
(3) If p
0

0
p
1

1
p
2

3
, then there is a q such that n p
n+1

n
q.
(4) If p [

is an ordinal, and n , then there exists q

n
p and a nite
set A Ord such that Q[

A.
1.14 Denition: (1) If P

, Q

: < ) is an iteration of strongly

-bounding forcing
notions, F nite, n , p, q P

, we say that p
F,n
q i p q and
F q[[ p()
n
q().
(2) A sequence p
n
, F
n
) : n ) is called a fusion sequence if F
n
: n ) is an increas-
ing family of nite subsets of , p
n
: n ) is an increasing family of conditions in P

,
n p
n

n,F
n
p
n+1
and

n
dom(p
n
)

n
F
n
.
Note that 1.13 is not a literally a strengthening of Baumgarters Axiom A (see [1]), as
we do not require that the relations
n
are transitive, and in (2) we only require p
n+1

n
q
rather than p
n+1

n+1
q. Nevertheless, the same proof as in [1] shows the following fact:
438 8 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
1.15 Fact:
(1) If the sequence p
n
, F
n
) : n ) is a fusion sequence, then there exists a
condition q P

such that for all n , p

n+1

F
n
,n
q.
(2) If

is a P

-name of an ordinal, n , F P

nite, then for all p there

exists a condition q
n,F
p and a nite set A of ordinals such that q[

A.
(3) If X

is a P

-name of a countable set of ordinals, n , F P

nite,
then for all p there exists a condition q
n,F
p and a countable set A of
ordinals such that q[ X

A.
The next fact is also well known:
1.16 Fact: Let B be the random real forcing. Then B is strongly

-bounding.
[Proof: Conditions in B are Borel subsets of [0, 1] of positive measure, p q i p q.
We let p
n
q i p q and (p q) 10
n1
(p), where is the Lebesgue measure.
Then if p
0

0
p
1

1
, letting q :=

n
p
n
we have for all n, all k n, (p
k
p
k+1
)
10
k1
(p
k
) 10
k1
(p
n
), so (p
n
q) 10
n1
+ 10
n2
+ 2 10
n1
(p
n
).
In particular, (q) 0.8 (p
0
), so q is a condition, and q
n1
p
n
for all n > 0.
Given a name

, an integer n and a condition p such that p[

is an ordinal, let A be
the set of all ordinals such that [[

= ]] p has positive measure ([[]] is the boolean

value of the statement , i.e. the union of all conditions forcing ). Since

A
([[

=
]] p) = (p) there is a nite subset F A such that letting q := p

A
[[

= ]] we
have (q) (1 10
n1
)(p). So q
n
p and q[

F.]
We will also need the following lemma from [17, 5, Theorem 9]:
1.17 Lemma: Every stationary S
2
can be written as a union of
2
many disjoint
stationary sets.
Finally, we will need the following easy fact (which is true for any forcing notion Q)
1.18 Fact: If f

is a Q-name for a function from to , [

Q
f

/ V , and r
0
, r
1
are any two
conditions in Q, then there are l , j
0
,= j
1
, r

0
r
0
, r

1
r
1
such that r

0
[ f

(l) = j
0
,
r

1
[ f

(l) = j
1
.
[Proof: There are a function f
0
and a sequence r
0
= r
0
r
1
of conditions in Q such
that for all n, r
n
[ f

[n = f
0
[n. Since r
1
[ f

/ V , r
1
[ l f

(l) ,= f
0
(l). There is a condition
r

1
r
1
such that for some l and some j
1
,= f
0
(l), r

1
[ f

(l) = j
1
. Let j
0
:= f
0
(l), and
let r

0
:= r
l+1
.]
438 9 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
2 H-perfect trees
In this section we describe a forcing notion PT
H
that we will use in an iteration in the
next section. We will prove the following properties of PT
H
:
(a) PT
H
is proper and

-bounding.
(b) PT
H
preserves P-points.
(c) PT
H
does not increase strong measure zero sets dened in the ground
model.
(d) PT
H
makes the reals of the ground model (and hence, by (c), the union of
all strong measure zero sets dened in the ground model) a strong measure
zero set.
2.1 Denition: For each function H with domain satisfying n 1 < [H(n)[ < ,
we dene the forcing PT
H
, the set of H-perfect trees to be the set of all p satisfying
(A) p
<
is a tree.
(B) p l dom() : (l) H(l).
(C) p : [succ
p
()[ 1, [H ([[)[.
(D) p p : , [succ
p
()[ = [H([[)[.
2.2 Denition:
(1) For p PT
H
, we let the set of splitting nodes of p be
split(p) := p : [succ
p
()[ > 1
(2) The height of a node p PT
H
is the number of splitting nodes strictly
below :
ht
p
() := [ : split(p)[
(Note that ht
p
() [[.)
(3) For p PT
H
, k , we let the kth splitting level of p be the set of splitting
nodes of height k.
split
k
(p) := split(p) : ht
p
() = k
(Note that split
0
(p) = stem(p).)
(4) For u , we let
split
u
(p) :=

ku
split
k
(p)
2.3 Remarks:
(i) Since H(n) is nite, (3) just means that either has a unique successor

i, or succ
p
() = H([[).)
(ii) Letting H

(n) = [H(n)[, clearly PT

H
is isomorphic to PT
H
(and the
obvious isomorphism respects the functions ht
p
(), p, k) split
k
(p),
etc)
438 10 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
2.4 Remark: If we let H(n) = for all n, then 2.1(A)(D) dene RP, rational perfect
set forcing. The denitions in 2.2 make sense also for this forcing. Since we will not use
the fact that H(n) is nite before 2.12, 2.52.11 will be true also for RP.
2.5 Fact: Let p, q PT
H
, n , ,
<
. Then
(1) If p, then ht
p
() ht
p
(). If moreover split(p), then
ht
p
() < ht
p
().
(2) If b p is a branch, then b split
n
(p) ,= .
(3) If p q, then for all n, q split
n
(p) ,= .
(4) If p and ht
p
() n then p, and split
n
(p).
(5) If
0
,=
1
are elements of split
n
(p), then
0
,
1
, and
1
,
0
.
Proof: (1) is immediate form the denition of ht.
For (2), it is enough to see that b split(p) is innite. (Then ordering b by inclusion, the
nth element of b split(p) will be in split
n1
(p).)
So assume that bsplit(p) is nite. Recall that each bsplit(p) has a unique successor
in p. By 2.1(C), b cannot have a last element, so b is innite. Hence there is
0
b such
that
b :
0
[succ
p
()[ = 1.
A trivial induction on [[ shows that this implies
p :
0
b.
Hence
p :
0
[succ
p
()[ = 1.
This contradicts 2.1(D).
To prove (3), let b be any branch of q. b is also a branch of p, so (2) shows that qsplit
n
(p)
b split
n
(p) ,= .
Proof of (4): Let b be a branch of p containing . By (2) there is b split
n
(p). If
, then ht
p
() > ht
p
() = n, which is impossible. Hence .
(5) follows easily from (1).
2.6 Denition: For p, q PT
H
, n , we let
(1) p q (q is stronger than p) i q p.
(2) p
n
q i p q and split
n
(p) q. (So also split
k
(p) q for all k < n.)
2.7 Fact: If p
n
q, n > 0, then stem(p) = stem(q).
2.8 Fact: Assume p, q PT
H
, n , p
n
q.
(0) For all q, ht
q
() ht
p
().
(1) For all k n, split
k
(p) split(q).
(2) For all k < n, split
k
(p) = split
k
(q).
438 11 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
(3) If p
n
q
n
r, then p
n
r.
Proof: (0) is clear.
(1): Let split
k
(p) for some k < n, then by 2.5(4) there is a , split
n
(p) q,
so q.
(2): Let split
k
(p), then split(q). Clearly ht
q
() ht
p
() = k. Using (1) induc-
tively, we also get ht
q
() k.
(3): Let split
n
(p). So q, ht
q
() ht
p
() = n. By 2.5(4), there is split
n
(q),
. As r, r.
2.9 Denition and Fact: If p
0

1
p
1

2
p
2

3
are conditions in PT
H
, then we call
the sequence p
n
: n < ) a fusion sequence. If p
n
: n < ) is a fusion sequence, then
(1) p

:=

n
p
n
is in PT
H
(2) For all n: p
n

n+1
p

.
2.10 Fact:
(1) If p PT
H
, then p
[]
PT
H
, and p p
[]
. (See 0.19.)
(2) If p q are conditions in PT
H
, q, then p
[]
q
[]
.
2.11 Fact: If for all split
n
(p), q

p
[]
is a condition in PT
H
, then
(1) q :=

split
n
(p)
q

is in PT
H
,
(2) q
n
p
(3) for all split
n
(p), q
[]
= q

.
2.12 Fact: If n , p PT
H
, then split
n
(p) is nite.
Proof: This is the rst time that we use the fact that each H(n) is a nite set: Assume
that the conclusion is not true, so for some n and p, split
n
(p) is innite. Then also
T := [k : split
n
(p), k [[ p
is innite. As T is a nitely splitting tree, there has to be an innite branch b T. By
2.5(2), there is b T, ht
p
() > n. This is a contradiction to 2.5(1).
2.13 Fact: PT
H
is strongly

-bounding, i.e.:
If

is a PT
H
-name for an ordinal, p PT
H
, n , then there exists a nite set A of
ordinals and a condition q PT
H
, p
n
q, and q[ A.
Proof: Let C := split
n
(p). C is nite. For each node C, let q

p
[]
be a condition
such that for some ordinal

. Now let
q :=

C
q

and A :=

: C
Since any extension of q must be compatible with some q
[]
(for some C), q[

A.
438 12 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
2.14 Corollary: PT
H
is proper (and indeed satises axiom A, so is -proper for any
<
1
) and

-bounding. Moreover, if n , p PT
H
,

a name for a set of ordinals,

then there exists a condition q
n
p such that
(1) If p [

is nite, then there is a nite set A such that q[

A.
(2) If p [

is countable, then there is a countable set A such that q[

A.
Proof: Use 2.13 and 2.9.
Similarly to 2.13 we can show:
2.15 Fact: Assume that

is a RP-name for an ordinal, p RP, n .

Then there exists a countable set A of ordinals and a condition q PT
H
, p
n
q, and
q[ A.
Proof: Same as the proof of 2.13, except that now the set C and hence also the set A may
be countable.
2.16 Fact: RP is proper (and saties axiom A). Proof: By 2.15 and 2.9.
2.17 Denition: Let G PT
H
be a V -generic lter. Then we let g

be the PT
H
-name
dened by
g

:=

pG
p
We may write g

H
or g

PT
H
for this name g

. If PT
H
is the th iterand Q

in an iteration,
we write g

for g

H
.
2.18 Fact:
PT
H
forces that
(0) g

is a function with domain ,

(1) n g

(n) H(n).
(2) For all f V , if n f(n) H(n) then

n f(n) = g

(n).
Furthermore, for all p PT
H
,
(3) p[ g

[n : n is a branch through p.
(4) p[ kng

[n split
k
(p)
Proof: (0) and (2) are straightforward density arguments. (1) and (3) follow immedaitely
from the denition of g

. (4) follows from (3) and 2.5(2), applied in V

PT
H
.
2.19 Remark: Since Unif (o) is equivalent to
for every H : , for every F [

n
H(n)]
<c
, there exists
f

such that for every f F there are innitely many n
satisfying f(n) = f

(n),
2.18(2) shows that if we have c =
2
and Martins Axiom for the forcing notions PT
H
(for
all H), then we also have Unif (o). (In fact the easy implication of this equivalence
is sucient.) This can be achieved by a countable support iteration of length
2
of forcing
notions PT
H
, with the usual bookkeeping argument (as in [16]).
438 13 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
We will show a stronger result in 3.3: If P := P

2
is the limit of a countable support
iteration P

, Q

: <
2
), where many Q

are of the form PT

H

for some H

, then
some bookkeeping argument can ensure that V
P
[= Add(o).
Since PT
H
is

-bounding, it does not add Cohen reals. The same is true for a countable
support iteration of forcings of the form PT
H
. However, in 3.9 we will have to consider a
forcing iteration in which some forcing notions are of the form PT
H
, but others do add an
unbounded real. To establish that even these iterations do not add Cohen reals, we will
need the fact that the forcing notion PT
H
preserves many ultralters.
2.20 Denition: Let Q be a forcing notion, x

a Q-name, p Q, p[ x

. We say that
x

is an interpretation of x

(above p), if for all n there is a condition p

n
p such
that p
n
[ x

n = x

n.
2.21 Fact: Assume Q, p, x

are as in 2.20. Then

(1) There exists x

such that x

is an interpretation of x

above p.
(2) If p p

and x

is an interpretation of x

above p

, then x

is an interpretation
of x

above p.
2.22 Lemma: PT
H
preserves P-points, i.e.: If | V is a P-point ultralter on , then
[
PT
H
| generates an ultralter.
Proof: Assume that the conclusion is false. Then there is a PT
H
-name

for a subset of
and a condition p
0
such that
p
0
[
PT
H
x | : [x

[ = [( x)

[ =
0
.
For each p PT
H
we choose a set (p) such that
(p) is an interpretation of

above p.
If there is an interpretation of

above p that is an element of |, then

(p) |.
Note that if (p) | and p p

, then also (p

) |, since (by 2.21(2)) we could have

chosen (p

) := (p). Hence either for all p (p) |, or for some p

1
p
0
, all p p
1
,
(p) / |. In the second case we let

be a name for the complement of

, and let
(p) = (p). Then (p) | for all p p
1
. Also, (p) is an interpretation of

above
p.
So wlog for all p p
1
, (p) | for some p
1
PT
H
, p
1
p
0
.
We will show that there is a condition q p
1
and a set a | such that q[ a

.
Recall that as | is a P-point, player NOTIN does not have a winning strategy in the
P-point game for | (see 1.2).
We now dene a strategy for player NOTIN. On the side, player NOTIN will construct a
fusion sequence p
n
: n < ) and a sequence m
n
: n < ) of natural numbers.
438 14 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
p
0
is given.
Given p
n
, we let
A
n
=

split
n+1
(p
n
)
(p
n
[]
)
This set is in |. Player IN responds with a nite set a
n
A
n
. Let m
n
:= 1 + max(a
n
).
For each split
n+1
(p
n
) there is a condition q

p
n
[]
forcing

m
n
= (p
n
[]
) m
n
,
so in particular
q

[ a
n

m
n
Let p
n+1
=

split
n+1
(p
n
)
q

.
Then
() p
n+1

n+1
p
n
and p
n+1
[ a
n

This is a well-dened strategy for player NOTIN. As it is not a winning strategy, there
is a play in which IN wins. During this play, we have constructed a fusion sequence
p
n
: n < ). Letting a :=

n
a
n
, q :=

n
p
n
, we have that a |, p
0
q PT
H
(by 2.9),
and q[ a

(by ()), a contradiction to our assumption.

The following facts will be needed for the proof that if we iterate forcing notions PT
H
with
carefully chosen functions H, then we will get a model where the ideal of strong measure
zero sets is c-additive.
2.23 Fact and Denition: Assume p PT
H
, u is innite, v = u. Then
we can dene a stronger condition q by trimming p at each node in split
v
(p). (See
2.2(4).) Formally, let = i

: split
v
(p)) be a sequence satisfying i

H([[) for all

split
v
(p).
Then
p

:= p : n dom() : If [n split
v
(p), then (n) = i
|n

is a condition in PT
H
Proof: Let q := p

. q satises (A)(B) of the denition 2.1 of PT

H
. The denition of p

immediately implies:
(1) If split
v
(p) q, then succ
q
() = i

.
(2) If split
u
(p) q, then succ
q
() = succ
p
() = H([[).
(3) If q split(p), then psplit(p), so succ
q
() = succ
p
() is a singleton.
Note that split(p) = split
u
(p) split
v
(p), so (1)(3) cover all possible cases for q.
So q also satises 2.1(C).
From (1)(3) we can also conclude:
(4) For all q: succ
q
() ,= .
438 15 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
To show that q PT
H
, we still have to check condition 2.1(D). So let q. Since u is
innite, there is k u, k > [[. By (4), there is an innite branch b q containing . By
2.5(2) there is b, ht
p
() = k. Then , and split(q).
2.24 Fact: p

[ g

& split
v
(p) g

([[) = i

(where g

is a name for the

generic branch dened in 2.18).
Proof: p

[ g

and succ
p

() = i

.
To simplify notation, we will now assume that for all n, H(n) . (If H(n) are just
arbitrary nite sets as in 2.1, then we could prove analogous statements, replacing 0 and
1 by any two elements 0
n
,= 1
n
of H(n).)
2.25 Denition: Let f

be a PT
H
-name for a function from to . We say that f

splits
on p, k if for all split
k
(p) there are l and j
1
,= j
0
such that
p
[

0]
[ f

(l) = j
0
p
[

1]
[ f

(l) = j
1
2.26 Remark: If f

splits on p, k, and q
k+1
p, then f

splits on q, k.
(Proof: split
k
(p) = split
k
(q), and for split
k
(p), p
[

i]
q
[

i]
.)
2.27 Lemma: If p[ f

/ V , k , then there is q
k+1
p such that f

splits on q, k.
Proof: For split
k
(p), i 0, 1 we let
i
be the unique element of split
k+1
(p) satisfying

i
i
.
By 1.18, for each split
k
(p) we can nd conditions q

0
p
[
0
]
, q

1
p
[
1
]
and integers
l

, j
,0
,= j
,1
such that q

0
[ f

(l

) = j
0
, q

1
[ f

(l

) = j
1
. If split
k+1
(p) is not of the
form
0
or
1
for any split
k
(p), then let q

:= p
[]
.
By 2.11, q :=

split
k+1
(p)
q

is a condition, q
k+1
p, and q

= q
[]
for all split
k+1
(p).
We nish the proof of 2.27 by showing that f

splits on q, k: Let split

k
(p) = split
k
(q).
Then q
[

0]
= q
[
0
]
= q

0
, so q
[

0]
[ f

(l

) = j
,0
. Similarly, q
[

1]
[ f

(l

) = j
,1
.
2.28 Lemma: If p[ f

/ V , then there is q p, f

splits on q, k for all k.

Proof: By 2.27, 2.26 and 2.9 (using a fusion argument).
2.29 Lemma: Assume Q is a strongly

-bounding forcing notion. Let f

be a Q-name
for a function, p a condition, n , p[ f

/ V . Then there exists a natural numer k such

that
() for all
k
2 there is a condition q
n
p, q[ f

/ [].
We will write k
p,n
or k
f

,p,n
for the least such k. Note that for any k k
p,n
, () will also
hold.
Proof: Assume that this is false. So for some f

, n
0
, p
0
,
() k
k

k
2 : (q
n
0
p
0
q[ f

/ [
k
])
438 16 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
Let
T :=
k
[l : l k, k
T is a nitely branching tree (

2) of innite height, so it must have an innite branch.
Let f

2 be such that f

[j : j T.
Since f

V but p
0
[
Q
f

/ V , there exists a name m

of a natural number such that

p
0
[ f

[m

,= f

[m

. Let q
n
0
p
0
be such that for some m

, q[ m

< m

.
Claim: For some k, q[ f

/ [
k
]. This will contradict ().
Proof of the claim: We have q[ f

[m

,= f

[m

. Since f

[m

T, there is a k m

such
that f

[m

=
k
[m

. Hence q[ f

[m

,= f

[m

=
k
[m

, so q[ f

/ [
k
[m

]. But then also

q[ f

/ [
k
].
This nishes the proof of the claim and hence of the lemma.
2.30 Lemma: Assume that Q is a strongly

-bounding forcing notion, H is a dominating
family in V , and =
h
: h H) has index H. Then
[
Q

hH

k
[
h
(k)] V
Proof: Assume that for some condition p and some Q-name f

,
p[ f

/ V & f

hH

n
[
h
(n)].
We will dene a tree of conditions such that along every branch we have a fusion sequence.
Specically, we will dene an innite sequence l
n
: n ) of natural numbers, and for
each n a nite sequence
p(
0
, . . . ,
n1
) :
0


0
2, . . . ,
n1

l
n1
2)
of conditions satisfying
(0) p() = p
(1) For all n:
0


0
2, . . . ,
n1

l
n1
2 : l
n
k
p(
0
,...,
n1
),n
.
(2) For all n:
0


0
2, . . . ,
n1

l
n1
2
n

l
n
2
(a) p(
0
, . . . ,
n1
)
n
p(
0
, . . . ,
n1
,
n
).
(b) p(
0
, . . . ,
n1
,
n
)[ f

/ [
n
].
Given p(
0
, . . . ,
n1
) for all
0


0
2, . . . ,
n1

l
n1
2, we can nd l
n
satisfying condition
(1). The by the denition of k
p(
0
,...,
n1
),n
we can (for all
n

l
n
2) nd p(
0
, . . . ,
n1
,
n
).
Now let h H be a function such that for all n, h(n) > l
n
. Dene a sequence
n
: n )
by
n
:=
h
(n)[l
n
, and let p
n
:= p(
0
, . . . ,
n
). Then p p
0

0
p
1

1
, so there
exists a condition q extending all p
n
. So for all n, q[ f

/ [
n
]. But then also for all n,
q[ f

/ [
h
(n)], a contradiction.
438 17 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
Lemma 2.30 will be needed later to show that if we iterate focings of the form PT
H
together
with random real forcing, after
2
many steps we obtain no strong measure zero sets of
size
2
. The proof (in 3.4) would be much easier if we could omit strongly from the
hypothesis of 2.30, i.e., if we could answer the following question positively:
2.31 Open Problem: Assume H

is a dominating family (or even wlog H =

),
and has index H. Let Q be an

-bounding forcing notion. Does this imply
[
Q

hH

n
[
h
(n)] V ?
2.32 Fact: Assume h

: 0, H

(n) =
h

(n)
2. Let H

be a dominating
family, and let have index H. Let g

be the name of the generic function added by

PT
H
.
Then
[
PT
H

h H

k
[
h
(k)]

n
[ g

(n)]
Proof: Assume not, then there is a condition p such that
() p[ h H

k
[
h
(k)] ,

n
[ g

(n)]
Let h H be a function such that k split
2k+1
(p) h

([[) h(k). This function

h will be a witness contradicting ().
For split
2k+1
(p) let i

succ
p
() = H

([[) =
h

(||)
2 be dened by i

:=
h
(k)[h

([[).
(Note that
h
(k)
h(k)
2 and h(k) h

([[).)
Let := i

: split
2k+1
(p), k ) and let q := p

.
Then q[ nk ( g

[n split
2k+1
(p) g

(n) = i
g

|n

h
(k)) by 2.24.
Since also q[ kn g

[n split
2k+1
(p), we get q[ kn[
k
(k)] [ g

(n)]. This contra-

dicts ().
3 Two models of Add(o).
Recall that S
1
2
:= <
2
: cf() =
1
.
3.1 Lemma: Let P

, Q

: <
2
) be an iteration of proper forcing noitions as in 1.8,
p P

2
, A

a P

2
-name. If p[ A

is a strong measure zero set, then there is a closed

unbounded set C
2
and a sequence

: C S
1
2
) such that each

is a P

-name,
and
p[

has index

V

and A

n
[
h
(n)]
Proof: Let c

be a P

2
-name for a function from
2
to
2
such that for all <
2
,
[

2
h

V

h
V
c

()
: n
h
(n)
h(n)
2 &A

n
[
h
(n)]
438 18 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
(Why does c

exist? Working in V [G

2
], note that there are only
1
many functions in

, and for each such h there is a

h
as required in

<
2
V

, by 1.7.)
As P

2
satises the
2
-cc, by 1.6(1) we can nd a function c V such that [

2
c

() <
c(). Let
C := : < c() <
The set C is closed unbounded. In V , we can assign to each P

-name h

(for < C)
a P

-name

h
such that
[

2
n

h
(n)
h

(n)
2 &A

n
[

h
(n)]
Now in V [G

] we can choose for each h

an < and a P

-name h

such that
h = h

[G

]. Then we let
h
:= (

h
)[G

]. Thus we found a sequence =

h
: h V

) V

as required.
3.2 Lemma:
Assume P

, Q

: <
2
) is a countable support iteration of proper forcing notions, where
for each ordinal S
1
2
[

= PT
H

for some P

-name H

. We will write g

for the
generic function added by Q

.
Assume H

is a name for a dominating family (

( 0)) in V

2
, and
[

2
For all h H

, S
h
:= <
2
:cf() =
1
, Q

= PT
H
V

is stationary (where H(n) =

h(n)
2).
Let G

2
P

2
be V -generic, then in V [G

2
], a set A IR is a strong measure zero set i
there is a closed unbounded set C
2
such that for every C S
1
2
, A

n
[g

(n)].
Proof: First we prove the easy direction. Assume that for some club C, for all C S
1
2
,
A

n
[g

(n)]. Then for every h V

2


( 0) there is a =
h
C S
h
S
1
2
.
So Q

h
= (PT
H
)
V

h
, where H(n) =
h(n)
2. Since g

h
(n)
h(n)
2, and A

[g

h
(n)] for
arbitrary h, A is a strong measure zero set.
Now for the reverse implication: In V

2
, let A be a strong measure zero set. By the previous
lemma, there is a club set C
2
and a sequence

: C S
1
2
) such that each V

is a sequence with index

V

and V

2
[= A X

. By 2.32 we have for all C S

1
2
:
V
+1
[= h V

n
[
h

(n)]

n
[g

(n)]
So x h
0
V

witnessing this. This inclusion is absolute, so also

V

2
[=

n
[
h
0

(n)]

n
[g

(n)]
438 19 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
Thus
V

2
[= A X

n
[
h
0

(n)]

n
[g

(n)]
and we are done.
3.3 Corollary: Assume P

2
is as above. Then [
P

2
Add(o).
Proof: Let A
i
: i
1
) be a family of strong measure zero sets in V

2
. To each i we can
associate a closed unbounded set C
i
as in 3.2. Let C :=

i
1
C
i
, then also C is closed
unbounded, and for all C S
1
2
,

i
1
A
i

n
[g

(n)]. Again by 3.2,

i
1
A
i
is a strong
measure zero set.
Our rst goal is to show that Unif (o) does not guarantee the existence of a strong measure
zero set of size c. Clearly the model for this should satisfy d =
1
(if c =
2
), so we will
construct a countable support iteration of

-bounding forcing notions.
3.4 Theorem: If ZFC is consistent, then
ZFC + c =
2
+ o = [IR]

1
+ no real is Cohen over L
+ there is a generalized Sierpinski set
is consistent.
Proof: We will start with a ground model V
0
satisfying V = L. Let H :=

( 0) L =
h

: <
1
, and let H

(n) =
h

(n)
2.
Let S

: <
1
) be a family of disjoint stationary sets <
2
: cf() =
1
.
Construct a countable support iteration P

, Q

: <
2
) satisfying
(1) For all even <
2
:
[
P

For some h : 0, letting H(n) =

h(n)
2, Q

= PT
H
.
(2) If S

, then [

= PT
H

.
(3) For all odd <
2
:
[
P

= random real forcing.

By 1.11 (or as a consequence of 1.15), P

2
is

-bounding, so [

2
H is a dominating
family. By 1.8(3) and 1.6 the assumptions of 3.3 are satised, so [

2
Add(o). Also,
[

2
c =
2
and there are no Cohen reals over L. Letting X be the set of random reals
added at odd stages, X is a generalized Sierpinski set: Any null set H V

2
is covered
by some G

null set H

that coded in some intermediate model. As coboundedly many

elements of X are random over this model, [H X[ [H

X[
1
.
To conclude the proof of 3.4, we have to show
V

2
[= If X IR is of strong measure zero, then [X[ < c.
438 20 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
Since H is a dominating family, by 0.4 it is enough to show that in V

2
the following holds:
If has index H, then [X

[
1
.
We will show: If V

has index H, then X

V

. (This is sucient, by 1.7.)

Assume to the contrary that G

2
is a generic lter, V

, and in V [G

2
] there is > ,
f V

<
V

, f X

. So also
V [G

] [= f V

<
V

and f X

Let f

be a P

-name,

a P

-name, and let p P

be a condition forcing all this. cannot

be a successor ordinal, by 2.30. So is a limit ordinal, and cf() must be , otherwise we
would have

2 V

=

2

<
V

.
So we have reduced the problem to the following lemma:
3.5 Lemma: Let P

, Q

: < ) be a countable support iteration of forcings where each

Q

(for even ) is of the form PT

H

for some (P

-name) H

, and Q

is random real
forcing for odd . Let be a limit ordinal of countable conality, and let f

be a
P

-name of a function in

2 such that [

< f

/ V

.
Let H V
0
be a dominating family of functions, and assume that has index H.
Then [

/

hH

n
[
h
(n)].
For notational simplicity, we again assume that for all even , [

: (rather
than H

: 2
<
).
Before we prove this lemma, we need the following two denitions (which make sense for
any countable support iteration P

, Q

: <
2
)).
3.6 Denition and Fact: For p P

, < , p[[ p() r

, we dene p r

as
follows: (p r

)() = p() for ,= , and (p r

)() = r

.
Then p r

, p r

p, and (p r

)[ = p[, so in particular p[[ p r

/G

.
Furthermore, p[[ (p r

)() = r

.
Also, if p() = r

, then p r

= p.
3.7 Denition and Fact: If p P

, A a countable subset of , and p[ r

/G

& r

p & dom( r

) A, then we dene p r

as follows:
For < , (p r

)() = p(). For and A, (p r

)() = r().
Again, p r

, p r

p, and (p r

)[ = p[, so in particular p[[ p r

/G

.
Also, if p
1
p
2
, then p
1
r

p
2
r

.
3.8 Proof of 3.5: cf() = , so we can nd an increasing sequence
n
: n < ) of even
ordinals converging to . Assume there is a condition p forcing that f

hH

n
[
h
(n)].
438 21 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
We will dene sequences p
n
: n < ),
F
n
: n < ),

n
: n < ),
s

n
: n )
p
i
n
: n , i 0, 1),
such that the following hold: For each n, p
n
, p
0
n
, p
1
n
are conditions in P

,
n
is an even
ordinal < , F
n
is a nite subset of
n
,
n
is an integer, and s

n
is a P

n
-name for an
element of
<
. (We let p
0
= p, F
0
= ,
0
= 0, p
1
0
= p
0
0
= p
0
, s
0
= ). For all n > 0 we
will have:
(1) p
n1

F
n
,n
p
n
.
(2) F
n

n
, F
n1
F
n+1
,

k
dom(p
k
)

k
F
k
.
(3)
n1
F
n
.
(4) p
n
[
n
[ s

n
= stem(p
n
(
n
)) = stem(p
n1
(
n
))
(5) For i 0, 1, p
i
n
= p
n
p
n
(
n
)
[s
n

i]
.
(6) p
n
[
n
[

n
l <
n
j
0
,= j
1
i 0, 1 : p
i
n
[

n
,
f

(l) = j
i
.
Note that (5) implies:
(5) p
n
[
n
[ p
i
n
P

/G

n
,
and (6) implies
(6) For all

n
2: p
n
[
n
[ i 0, 1 : p
i
n
[

n
,
f

[
n
,=
[Proof of (6) (6): In V

n
, let i 0, 1 be such that (l) ,= j
i
, where l is as in (6).]
Finally, let q =

n
p
n
. Then q[
n
[ stem(p
n
(
n
)) = stem(q(
n
)) = s

n
by (1), (3) and (4)
and 2.7. Let h

H be a function such that for all n,

n
< h

(n). So for all n,

h

(n)[
n
is a well-dened member of

n
2.
For each n, let i

n
be a P

n
-name of an element of 0, 1 such that
(6) p
n
[
n
[ p
i

n
n
[ f

[
n
,=
h

(n)[
n
Now dene a condition q

as follows: For /
n
: n , q

() = q(), and
q

(
n
) = q()
[s
n

n
]
(This is a P

n
-name.)
Claim: q

q p (this is clear) and q

[ f

/

n
[
h

(n)].
To prove this claim, let G

be a generic lter containing q

, and assume f := f

[G

] is
in [
h

(n)]. Let i
n
:= i

n
[G

n
]. Now q G

implies p
n
G

, so in particular p
n
[
n
G

n
.
Note that stem(q(
n
)) = stem(p
n
(
n
)) = s
n
, so q

&p
n
G

implies p
i
n
n
G

. Also,
by (6) we have q

[ f

n
[s
f

(n)], a contradiction.
This nishes the proof of 3.5 modulo the construction of the sequences p
n
, F
n
, etc.
First we x enumerations dom(r) =
m
r
: m for all r P

. We will write
m
n
for

m
p
n
.
438 22 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
Assume p
n1
is given. Let F
n
:=
n
(F
n1

m
k
: k < n, m < n
n1
). This will
take care of (2) and (3).
To dene p
n
, rst work in V [G

n
], where p
n1
[
n
G

n
.
We let s
n
:= stem(p
n1
(
n
)).
We let r
0
:= p
n1
p
n1
(
n
)
[s
n

0]
, and r
1
:= p
n1
p
n1
(
n
)
[s
n

1]
.
By 1.18, we can nd l, j
0
,= j
1
and r

0
, r

1
such that r

i
r
i
, and r

i
[ f

(l) = j
i
.
We now dene a condition r P

/G

n
as follows:
r[
n
= p
n1
[
n
.
r(
n
) = r

0
(
n
) r

1
(
n
)

p
n1
(
n
)
[s
n

i]
: i succ
p
n1
(
n
)
(s
n
) 0, 1.
(So stem(r(
n
)) = s
n
.)
If dom(p
n1
) dom(r

0
) dom(r

1
) and >
n
, we let r() be a P

-name
such that
p
n1
[
n
[

n
[

n
,
For i in 0, 1: If s

i g

n
, then r() = r

i
(),
and if g

n
extends neither s

0 nor s

1,
then r() = p
n1
()
(We write g

n
for g
Q

n
, the branch added by the forcing Q

n
.)
This is a condition in P

/G

n
. Note that we have the following:
(i) stem(p
n
(
n
)) = s
n
= stem(p
n1
(
n
))
(ii) For i 0, 1, r r

i
(
n
) r

i
.
(iii) r

p
n1
.
Coming back to V , we can nd names r

, . . . , such that the above is forced by p

n1
[
n
.
Now let r be a condition in P

n
satisfying the following:
(a) r
F
n
,n
p
n1
[
n
.
(b) For some countable set A , r[ dom( r

) A.
(c) For some
n
, r[ l

<
n
.
We can nd a condition r satisfying (a)(c) by 1.15.
Finally, let p
n
:= r r

. So p
n
[
n
= r.
And let p
i
n
be dened by (5).
Why does this work?
First we check (1): p
n1
[
n

F
n
,n
p
n
[
n
by (a), and p
n1
p
n
, because p
n
[
n
[ p
n
=
r r

p
n1
(by (iii)). So p
n1

F
n
,n
p
n
.
(2) and (3) are clear.
Proof of (4): p
n
[
n
[ stem(p
n
(
n
)) = stem(( r r

)(
n
)) = stem( r

(
n
)) = s

n
.
(6): Let G

n
be a generic lter containing p
n
[
n
. Work in V [G

n
]. We write r for r

[G

n
],
etc.
438 23 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
We want to show p
0
n
[

(l) = j
0
. (p
1
n
[

(l) = j
1
is similar.) As r

0
[

(l) = j
0
, it
is enough to see p
0
n
r

0
.
First we note that p
0
n
p
n
p
n1
. Also p
0
n
= p
n
p
n
(
n
)
[s
n

0]
p
n
= r r r.
Finally, p
0
n
(
n
) = r(
n
)
[s
n

0]
= r

0
(
n
).
So p
0
n
= p
0
n
r

0
(
n
) r r

0
(
n
) r

0
, and we are done.
Our next model will satisfy
() Unif(o) + d = c =
2
.
This in itself is very easy, as it is achieved by adding
2
Cohen reals to L. (Also Miller
[10] showed that Unif(o) + c =
2
+ b =
1
is consistent.)
Our result says that we can obtain a model for () (and indeed, satisfying Add(o)) without
adding Cohen reals. In particular, () does not imply Cov(/).
3.9 Theorem: Con(ZFC) implies
Con(ZFC + c = d =
2
> b + Add(o) + no real is Cohen over L)
Proof (sketch): We will build our model by a countable support iteration of length
2
where
at each stage we either use a forcing of the form PT
H
, or rational perfect set forcing. A
standard bookkeeping argument ensures that the hypothesis of 3.3 is satised, so we get
[

2
Add(o). Using rational perfect set forcing on a conal set yields [

2
d = c =
2
.
Since all P-point ultralters from V
0
are preserved, no Cohen reals are added.
Proof (detailed version): Let <
2
: cf() =
1

<
2
S

, where S

: <
2
) is a
family of disjoint stationary sets. Let :
2

1

2
be a bijection. We may assume
that S
(,)
> .
First we claim that there is a countable support iteration P

, Q

: <
2
) and a sequence
of names H

: <
2
) : <
1
) such that
(1) For all <
2
, all <
1
, H

is a P

-name.
(2) For all <
2
, [

: <
1
=

( 0, 1).
(3) For all <
2
: If /

<
2
S

, then [

= RP.
(4) For all <
2
, all <
1
, all S
(,)
: [

= PT
H

.
Proof of the rst claim: By induction on we can rst dene P

, then H

: <
1
) (by
1.8(1)), then Q

(by (3) or (4), depending on whether

<
2
S

or not).
Our second claim is that letting H

be a name for all functions from to 0, 1 in

V [G

2
], the assumptions of 3.3 are satised, namely:
(a) [

2
H H

<
2
S

S
H
.
(b) [

2
<
2
S

is stationary.
(b) follows from 1.8(3) and 1.6, and (a) follows from
[

2
For all H H

there is <
2
and <
1
such that H = H

.
438 24 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
which in turn is a consequence of 1.7.
So by 3.3, V

2
[= Add(o).
Let G

2
P

2
be a generic lter, V

2
= V [G

2
].
Again by 1.7, every H

V

2
of size
1
is a subset of some V

, <
2
, so H
cannot be a dominating family, as rational perfect set forcing Q
+1
will introduce a real
not bounded by any function in H V

V
+1
. Hence d = c =
2
.
Finally, any P-point ultralter from V is generates an ultralter in V

2
, so there are no
Cohen reals over V .
This ends the proof of 3.9.
REFERENCES.
[1] J. Baumgartner, Iterated forcing, in: Surveys in set theory (A. R. D. Mathias,
editor), London Mathematical Society Lecture Note Series, No. 8, Cambridge
University Press, Cambridge, 1983.
[2] A. Blass and S. Shelah, There may be simple P

1
- and P

2
-points, and the Rudin-
Keisler order may be downward directed, Annals of pure and applied logic, 33
(1987), pp.213243.
[3] P. Corazza, The generalized Borel Conjecture and strongly proper orders, Transac-
tions of the American Mathematical Society, 316 (1989), pp.115140. .
[4] M. Goldstern, H. Judah, On Shelahs preservation theorem, preprint.
[5] H. Judah, S Shelah, H. Woodin, The Borel Conjecture, Annals of pure and applied
logic, to appear.
[6] H. Judah, Strong measure zero sets and rapid lters, Journal of Symbolic logic, 53
(1988), pp.393402.
[7] K. Kunen, Set Theory: An Introduction to Independence Proofs,
[8] A. Miller, Mapping a set of reals onto the reals, Journal of Symbolic logic, 48 (1983),
pp.575584.
[9] A. Miller, Rational Perfect Set Forcing, in: Axiomatic Set Theory, Boulder, Co
1983, 143159, Contemporary Math 31, AMS, Providence RI 1984.
[10] A. Miller, Some properties of measure and category, Transactions of the American
Mathematical Society, 266,1 (1981), pp.93114.
[11] J. Pawlikowski, Power of transitive bases of measure and category, Proceedings of
the American Mathematical Society, 93 (1985), pp.719729.
438 25 January 1991
4
3
8


r
e
v
i
s
i
o
n
:
1
9
9
3
-
0
8
-
2
9







m
o
d
i
f
i
e
d
:
1
9
9
3
-
0
8
-
2
9


Goldstern, Judah, Shelah: Strong measure zero sets without Cohen reals
[12] Rothberger, Sur des families indenombrables de suites de nombres naturels et
les probl`emes concernant la proprietee C, Proc. Cambr. Philos. Soc., 37 (1941),
pp.109126.
[13] Rothberger, Eine Versch arfung der Eigenschaft C, Fundamenta Mathematicae, 30
(1938), pp.5055.
[14] S. Shelah, Proper Forcing, Lecture Notes in Mathematics Vol. 942, Springer Verlag.
[15] S. Shelah, Proper and Improper Forcing, to appear in Lecture Notes in Mathematics,
Springer Verlag.
[16] R. Solovay and S. Tennenbaum, Iterated Cohen extensions and Souslins problem,
Annals of Mathematics, 94 (1971), pp.201245.
[17] R. Solovay, Real valued measurable cardinals, in: Axiomatic Set Theory, Proc.
Symp. Pure Math. 13 I (D. Scott, ed.), pp.397428, AMS, Providence RI, 1971.
Martin GOLDSTERN goldstrn@bimacs.cs.biu.ac.il
Haim JUDAH judah@bimacs.cs.biu.ac.il
Department of Mathematics
Bar Ilan University
52900 Ramat Gan, Israel
Saharon SHELAH shelah@shum.huji.ac.il
Department of Mathematics
Givat Ram
Hebrew University of Jerusalem
438 26 January 1991