You are on page 1of 9

THE DEPTH OF ULTRAPRODUCTS OF BOOLEAN ALGEBRAS SH853

Saharon Shelah The Hebrew University of Jerusalem Einstein Institute of Mathematics Edmond J. Safra Campus, Givat Ram Jerusalem 91904, Israel Department of Mathematics Hill Center-Busch Campus Rutgers, The State University of New Jersey 110 Frelinghuysen Road Piscataway, NJ 08854-8019 USA
Abstract. We show that if is a compact cardinal then the depth of ultraproducts of less than many Boolean Algebras is at most plus the ultraproduct of the depths of those Boolean Algebras.

revision:2005-05-03

modified:2005-05-04

I would like to thank Alice Leonhardt for the beautiful typing. This research was supported by the United States-Israel Binational Science Foundation First Typed - 03/Nov/14 Latest Revision - 05/May/3 Typeset by AMS-TEX

(853)

SAHARON SHELAH

0 Introduction Monk has looked systematically at cardinal invariants of Boolean Algebras. In particular, he has looked at the relations between inv( Bi /D) and inv(Bi )/D,
i< i<

i.e., the invariant of the ultraproducts of a sequence of Boolean Algebras vis the ultraproducts of the sequence of the invariants of those Boolean Algebras for various cardinal invariants inv of Boolean Algebras. That is: is it always true that inv( Bi /D) (inv(Bi /D)? is it consistently always true? Is it always true
i< i<

that
i<

inv(Bi )/D inv(


i<

Bi /D)? is it consistenly always true? See more on

this in Monk [Mo96]. Roslanowski Shelah [RoSh 534] deals with specic inv and with more on kinds of cardinal invariants and their relationship with ultraproducts. Monk [Mo90a], [Mo96], in his list of open problems raises the question for the central cardinal invariants, most of them have been solved by now; see Magidor Shelah [MgSh 433], Peterson [Pe97], Shelah [Sh 345], [Sh 462], [Sh 479], [Sh 589, 4], [Sh 620], [Sh 641], [Sh 703], Shelah and Spinas [ShSi 677]. We here throw some light on problem 12 of [Mo96], pg.287 and will be continued in [Sh:F683]. We thank the referee for many helpful comments.
modified:2005-05-04

0.1 Denition. For a Boolean Algebra B let (a) Depth(B) = sup{: in B there is an increasing sequence of length } (b) Depth+ (B) = sup{ + : in B there is an increasing sequence of length }.

0.2 Remark. So Depth+ (B) = + Depth(B) = and if Depth+ (B) is a limit cardinal then Depth+ (B) = Depth(B).

(853)

revision:2005-05-03

THE DEPTH OF ULTRAPRODUCTS OF BOOLEAN ALGEBRAS SH853

1 Above a compact cardinal The following claim gives severe restrictions on any try to build a ZFC example for Depth( B )/D > Depth(B )/D if V is near L, see [Sh 652] for compli< <

mentary to 1. 1.1 Claim. 1) Assume (a) < (b) is a compact cardinal (c) D is an ultralter on (d) = cf() such that ( < )(|| < ) (e) Bi (i < ) is a Boolean Algebra with Depth+ (Bi ) (f ) B =
i<

Bi /D.

Then Depth+ (B) . 2) Instead ( < )|| < ) it suces that ( < )(| /D| < = cf()). 3) We can weaken clause (e) (for parts (1) and (2)) to (g) {i < : Bi is a Boolean Algebra with Depth+ (Bi ) } D.
modified:2005-05-04

Proof. 1) Toward contradiction assume that a : < is an increasing sequence in B. So let a = a : i < /D, so for < , A, =: {i < : Bi |= a < a } i i i D. Let E be a -complete uniform ultralter on . For each < let A be such that the set { : < < and A, = A } is a member of E so an unbounded subset of (exist as = cf() > 2 ). We choose C as follows

revision:2005-05-03

C =: { < : is a limit ordinal and if u is bounded of cardinality then = sup(Su )} where Su =: { < : > sup(u) and ( u)(A, = A )}.

(853)

SAHARON SHELAH

As = cf() > 2 = |D|, for some A D the set S =: { < : cf() > and A = A } is a stationary subset of . As we have assumed = cf() and ( < )(|| < ), clearly C is a club of . Let { : < } C, increases continuous with and +1 S. For each < the family A = {Su +1 \ : u [+1 ] } is a downward + -directed family of non-empty subsets of [ , +1 ) hence there is a + -complete lter E on [ , +1 ) extending A . For < and i < let W,i =: { : < +1 and i A,+1 } and let + B =: {i < : W,i E }. As E is + -complete clearly W =: {[ , +1 )\W,i : i \B } E hence there is W ; if i A,+1 then { : < +1 and + i A,+1 } E , so A,+1 is a subset of B and belongs to D hence B D. So B A D is non-empty. So for each for some i+1 A we have
+ { : < +1 and i+1 A,+1 } E .

We can nd i A such that Y = { < : is an even ordinal and i+1 = i } has cardinality , and let Z = {+1 : Y } so Z [] . Now ()0 Y A+1 = A [why? as +1 S] ()1 i A D [trivial; note if < , ||2 < we can have E is (2 )+ -complete lter so we have B+1 instead of i so we can weaken D ultralter to: D P() upward closed and the intersection of any two non-empty] ()2 if < are from Z then i A, [why? let = +1 , = +1 so < ; let U1 := { : < < +1 , A, = A (= A+1 )}
modified:2005-05-04

revision:2005-05-03

so U1 = S{+1 } [ , +1 ) A E and let


+ U2 := { : < +1 , i A,+1 } E .

(853)

THE DEPTH OF ULTRAPRODUCTS OF BOOLEAN ALGEBRAS SH853

[Why? As this is how i+1 is dened.] + So for any < from Z as U1 E and U2 E clearly there is U1 U2 hence ( = +1 < < +1 = and) for i = i we have Bi |= ai +1 < a i (because U1 ) and Bi |= a < ai +1 (because U2 ) so together Bi |= ai +1 < i ai +1 but = +1 , = +1 so we have gotten Bi |= a < a so we are done. i i 2) We change the choice of the club C. By the assumption, for each < let f /D : < be a list of the members of /D without repetitions, so < . Let C = { :(i) (ii) (iii) < is a limit ordinal if < then < if < and < and A = Ai : i < D and there is [, ) such that i < Af (i), = Ai then there is
(, ) such that i < Af (i), = Ai }.

Clearly C is a club of . The only additional point in the proof is () if 1 < 2 are from C and A2 = A then there is i A such that: for every S 1 there is [1 , 2 ) satisfying A, = A i A,2 . [Why () holds? If not, then for every i A there is i S 1 satisfying [1 , 2 ) Ai , = A i A,2 . Let f be dened by: f (i) = i , if / i A , f (i) = 0 otherwise, so for some < 1 we have f = f 1 mod D hence A =: {i A : f (i) = f 1 (i)} D. As < and D is -complete there is 1 (2 , ) such that i < Af 1 (i),1 = Af 1 (i) hence by the choice of C there is 2 (1 , 2 ) such that i < Af 1 (i), = Af 1 (i), = Af 1 (i) . But
2 1

modified:2005-05-04

revision:2005-05-03

i A f 1 (i) = f (i) = i S Ai ,2 = Af 1 (i),2 = Af 1 (i) = A so i A Ai ,2 = A . Now A2 ,2 D hence there is i A A1 ,2 and for it we get contradiction.] Of course, the set of such i s belongs to D. 3) Obvious. 1.1

1.2 Conclusion: Let be a compact cardinal. If < and D is an ultralter on , Bi is a Boolean Algebra for i < then () (a) (b) if D is a regular ultralter then Depth(
i<

Bi /D) +
i<

Depth(Bi )/D

this holds if = 0 .

(853)

SAHARON SHELAH

Proof. If this fails, let = ( +


i<

Depth(Bi )/D)+ , so is a regular cardinal > Bi /D), so by 1.1 we get


i< 1.2

and ( < )[| /D| < ] - see below and Depth( a contradiction.

1.3 Remark. 1) Actually we prove that if is a compact cardinal, < = cf() and c : []2 then we can nd an increasing sequence : < of ordinals < and i, j < such that for every < < for some satisfying < < we have c{ , } = i, c{, } = j (the result follows using c : []2 D). 2) We use i rather than some B D in order to help clarify what we need. 3) Note that if D is a normal ultralter on > 0 and i : i < is increasing continuous with limit , i < j < i+1 then = i /D but /D >
ji i<

. This is essentially the only reason for the undesirable extra assumption D is regular in 1.2. Note 1.4 Claim. 1) In 1.1 instead (, ] is a compact cardinal it suces to demand: + ,2 , where
modified:2005-05-04

,, if c : []2 then we can nd a stationary S and < such that for every u [S]< the set Su = { < : ( u)[c{, } = ]} is unbounded in . 2) If is supercompact < = cf() < < = cf() and Q = adding Cohen subsets of then in V, ,, holds (even ,1 , if < < in V). 1 In 1.4 we cannot get such results for > because for supercompact Laver indestructible and regular > > we can force { < :cf() > } to have a square preserving the supercompactness.

revision:2005-05-03

1.5 Claim. Assume = cf() > + and = cf(), and there is a square on S = { < : cf() } (see 1.6 below). Then (a) there is a sequence Bi : i < of Boolean Algebras such that () Depth+ (Bi ) () for any uniform ultralter D on , Depth+ (
i<

Bi /D) >

(853)

THE DEPTH OF ULTRAPRODUCTS OF BOOLEAN ALGEBRAS SH853

(b) the proof of [Sh 652, 5.1] can be carried. Where 1.6 Denition. For = cf() > 0 , S = sup(S) we say that S has a square when we can nd S + and C : S + such that (a) S\S + is not a stationary subset of (b) C is a closed subset of (c) C S C = C (d) we stipulate C = {} for S + . /

Proof of 1.5. As in [Sh 652, 5.1] using C = C : S + from 1.6 instead + acc(C ) : < . The only change being that in the proof of [Sh 652, Fact 5.3] in case 3 we have just cf() and let : < cf() be increasing continuous with limit . If cf() < we can nd () < such that 1 < 2 < 1 A2 ,() and let A, = if < () and A, = {A , : < cf()} if [(), ). 1.6

(853)

revision:2005-05-03

modified:2005-05-04

SAHARON SHELAH

REFERENCES. [MgSh 433] Menachem Magidor and Saharon Shelah. Length of Boolean algebras and ultraproducts. Mathematica Japonica, 48(2):301307, 1998. math.LO/9805145. [Mo90a] J. Donald Monk. Cardinal Invariants of Boolean Algebras. Lectures in Mathematics. ETH Zurich, Birkhauser Verlag, Basel Boston Berlin, 1990. J. Donald Monk. Cardinal Invariants of Boolean Algebras, volume 142 of Progress in Mathematics. Birkhuser Verlag, BaselBostonBerlin, a 1996. Douglas Peterson. Cardinal functions on ultraproducts of Boolean algebras. Journal of Symbolic Logic, 62:4359, 1997.

[Mo96]

[Pe97]

[RoSh 534] Andrzej Roslanowski and Saharon Shelah. Cardinal invariants of ultrapoducts of Boolean algebras. Fundamenta Mathematicae, 155:101151, 1998. math.LO/9703218. [Sh:F683] [Sh 345] Saharon Shelah. About the DEPTH+ of Boolean Algebra. Saharon Shelah. Products of regular cardinals and cardinal invariants of products of Boolean algebras. Israel Journal of Mathematics, 70:129187, 1990. Saharon Shelah. On Monks questions. Fundamenta Mathematicae, 151:119, 1996. math.LO/9601218. Saharon Shelah. -entangled linear orders and narrowness of products of Boolean algebras. Fundamenta Mathematicae, 153:199275, 1997. math.LO/9609216. Saharon Shelah. Special Subsets of cf() , Boolean Algebras and Maharam measure Algebras. Topology and its Applications, 99:135 235, 1999. 8th Prague Topological Symposium on General Topology and its Relations to Modern Analysis and Algebra, Part II (1996). math.LO/9804156. Saharon Shelah. Applications of PCF theory. Journal of Symbolic Logic, 65:16241674, 2000. Saharon Shelah. Constructing Boolean algebras for cardinal invariants. Algebra Universalis, 45:353373, 2001. math.LO/9712286. Saharon Shelah. More constructions for Boolean algebras. Archive for Mathematical Logic, 41:401441, 2002. math.LO/9605235.

modified:2005-05-04

[Sh 479] [Sh 462]

[Sh 620]

revision:2005-05-03

[Sh 589] [Sh 641] [Sh 652]

(853)

THE DEPTH OF ULTRAPRODUCTS OF BOOLEAN ALGEBRAS SH853

[Sh 703] [ShSi 677]

Saharon Shelah. On ultraproducts of Boolean Algebras and irr. Archive for Mathematical Logic, 42:569581, 2003. math.LO/0012171. Saharon Shelah and Otmar Spinas. On incomparability and related cardinal functions on ultraproducts of Boolean algebras. Mathematica Japonica, 52:345358, 2000. math.LO/9903116.

(853)

revision:2005-05-03

modified:2005-05-04

You might also like