Coloring ordinals by reals
We study combinatorial principles we call Homogeneity Principle HP(κ) and Injectivity Principle IP(κ,λ) for regular κ>\aleph_1 and λ\leqκwhich are formulated in terms of coloring the ordinals <κby reals.
math.LO↗
arXiv subjects
Jörg Brendle
Publications and source records attributed to Jörg Brendle.
Ultrafilters on omega --- their ideals and their cardinal characteristics
For a free ultrafilter U on omega we study several cardinal characteristics which describe part of the combinatorial structure of U. We provide various consistency results; e.g. we show how to force simultaneously many characters and many pi --characters. We also investigate two ideals on the Baire space omega^omega naturally related to U and calculate cardinal coefficients of these ideals in terms of cardinal characteristics of the underlying ultrafilter.
math.LO↗
Solovay-type characterizations for forcing-algebras
We give characterizations for the (in ZFC unprovable) sentences ``Every $\Sb{1}{2}$--set is measurable" and ``Every $\Db{1}{2}$--set is measurable" for various notions of measurability derived from well--known forcing partial orderings.
math.LO↗
Strolling through Paradise
With each of the usual tree forcings I (e.g., I = Sacks forcing S, Laver forcing L, Miller forcing M, Mathias forcing R, etc.) we associate a sigma--ideal i^0 on the reals as follows: A \in i^0 iff for all T \in I there is S \leq T (i.e. S is stronger than T or, equivalently, S is a subtree of T) such that A \cap [S] = \emptyset, where [S] denotes the set of branches through S. So, s^0 is the ideal of Marczewski null sets, r^0 is the ideal of Ramsey null sets (nowhere Ramsey sets) etc. We show (in ZFC) that whenever i^0, j^0 are two such ideals, then i^0 \sem j^0 \neq \emptyset. E.g., for I=S and J=R this gives a Marczewski null set which is not Ramsey, extending earlier partial results by Aniszczyk, Frankiewicz, Plewik, Brown and Corazza and answering a question of the latter. In case I=M and J=L this gives a Miller null set which is not Laver null; this answers a question addressed by Spinas. We also investigate the question which pairs of the ideals considered are orthogonal and which are not. Furthermore we include Mycielski's ideal P_2 in our discussion.
math.LO↗
Mutually Generics and Perfect Free Subsets
We discuss ways of adjoining perfect sets of mutually generic random reals. In particular, we show that if V \sub W are models of ZFC and W contains a dominating real over V, then W[r], where r is random over W, contains a perfect tree of mutually random reals over V. This result improves an earlier result by Bartoszynski and Judah and sheds new light on an old Theorem of Mycielski's. We also investigate the existence of perfect free subsets for projective functions f: (omega^omega)^n \to omega^omega. In particular, we prove that every projective function has a perfect free subset in the Cohen real model, while in the Sacks real model the following holds: every Delta^1_2-- function f: omega^omega \to omega^omega has a perfect free subset and there is a Delta^1_2--function f: (omega^omega)^2 \to omega^omega without perfect free subsets. This is connected with recent results of Mildenberger. Finally, we show that the existence of a superperfect tree of Cohen reals over a model V of ZFC implies the existence of a dominating real over V, thus answering a question addressed by Spinas.
math.LO↗
Forcing tightness in products of fans
The theta--fan F_theta is the quotient space obtained by identifying the non--isolated points of the product theta \times (omega+1) with a single point \infty. (Here, theta has the discrete topology and omega+1 has the order topology.) We define several notions of forcing that allow us to manipulate the tightness t of products of fans. Some consequences include: t (F_theta \times F_omega) = theta does not imply the existence of a (theta,omega)--gap, new examples of first countable <theta--collectionwise Hausdorff (cwH) spaces that are not \leq theta--cwH for singular cardinals theta, and for cardinals lambda \leq theta with cf(theta) \geq omega_1 and lambda regular, a first countable <theta--cwH not \leq theta--cwH space that can be made cwH by removing a closed discrete set of cardinality lambda. We also prove two theorems that characterize tightness of products of fans in terms of families of integer-- valued functions.
math.LO↗
Evasion and prediction II
A subgroup G \leq Z^omega exhibits the Specker phenomenon if every homomorphism G \to Z maps almost all unit vectors to 0. We give several combinatorial characterizations of the cardinal se, the size of the smallest G \leq Z^omega exhibiting the Specker phenomenon. We also prove the consistency of b < e, where b is the unbounding number and e the evasion number introduced recently by Blass. Finally we show that e \geq \min { aleph_0--e , cov(M) }, where aleph_0--e is the aleph_0--evasion number (the uniformity of the evasion ideal) and cov(M) is the covering number of the meager ideal. All these results can be dualized. They answer several questions addressed by Blass.
math.LO↗
Nicely generated and chaotic ideals
We show that if the real line is the disjoint union of kappa meager sets such that every meager set is contained in a countable union of them, then kappa = omega_1. This answers a question addressed by J. Cichon. We also prove two theorems saying roughly that any attempt to produce the isomorphism type of the meager ideal in the Cohen real and the random real extensions must fail. All our results hold for "meager" replaced by "null", as well.
math.LO↗
Evasion and prediction --- the Specker phenomenon and Gross spaces
We study the set--theoretic combinatorics underlying the following two algebraic phenomena. (1) A subgroup G leq Z^omega exhibits the Specker phenomenon iff every homomorphism G to Z maps almost all unit vectors to 0. Let se be the size of the smallest G leq Z^omega exhibiting the Specker phenomenon. (2) Given an uncountably dimensional vector space E equipped with a symmetric bilinear form Phi over an at most countable field KK, (E,Phi) is strongly Gross iff for all countably dimensional U leq E, we have dim(U^perp) leq omega. Blass showed that the Specker phenomenon is closely related to a combinatorial phenomenon he called evading and predicting. We prove several additional results (both theorems of ZFC and independence proofs) about evading and predicting as well as se, and relate a Luzin--style property associated with evading to the existence of strong Gross spaces.
math.LO↗
Combinatorial properties of classical forcing notions
We discuss the effect of adding a single real (for various forcing notions adding reals) on cardinal invariants associated with the continuum (like the unbounding or the dominating number or the cardinals related to measure and category on the real line). For random and Cohen forcing, this question was investigated by Cicho'n and Pawlikowski; for Hechler forcing, by Judah, Shelah and myself. We show here: (1) adding an eventually different or a localization real adjoins a Luzin set of size continuum and a mad family of size omega_1; (2) Laver and Mathias forcing collapse the dominating number to omega_1 --- consequences: (A) CON(d=omega_1 + unif(L) = unif (M) = kappa = 2^omega) for any regular uncountable kappa; (B) Two Laver or Mathias reals added iteratively always force CH (even diamond); (C) Sigma^1_4-Mathias-absoluteness implies the Sigma^1_3- Ramsey property; (3) Miller's rational perfect set forcing preserves the axiom MA(sigma-centered).
math.LO↗
The additivity of porosity ideals
We show that several sigma-ideals related to porous sets have additivity omega_1 and cofinality 2^omega. This answers a question addressed by Miroslav Repick'y.
math.LO↗
Set-theoretic aspects of periodic $FC$-groups --- extraspecial p-groups and Kurepa trees
Given a group G, we let Z(G) denote its center, G' its commutator subgroup, and Phi (G) its Frattini subgroup (the intersection of all maximal proper subgroups of G). Given U leq G, we let N_G (U) stand for the normalizer of U in G. A group G is FC iff every element g in G has finitely many conjugates. A p-group E is called extraspecial iff Phi (E) = E' = Z(E) cong Z_p, the cyclic group with p elements. When generalizing a characterization of centre-by-finite groups due to B. H. Neumann, M. J. Tomkinson asked the following question. Is there an FC-group G with vert G / Z(G) vert = kappa but [G:N_G(U)] < kappa for all (abelian) subgroups U of G, where kappa is an uncountable cardinal. We consider this question for kappa = omega_1 and kappa = omega_2. It turns out that the answer is largely independent of ZFC, and that it differs greatly in the two cases. More explicitly, for kappa = omega_1, it is consistent with, and independent of, ZFC that there is a group G with vert G / Z(G) vert = omega_1 and [G:N_G (A)] leq omega for all abelian A leq G. We do not know whether the same statement is still consistent if we drop abelian. On the other hand, for kappa = omega_2, the non-existence of groups G with vert G / Z(G) vert = omega_2 and [G : N_G (A) ] leq omega_1 for all (abelian) A leq G is equiconsistent with the existence of an inaccessible cardinal. In particular, there is an extraspecial p-group with this property if there is a Kurepa tree.
math.LO↗
Combinatorial properties of Hechler forcing
In this work we use a notion of rank first introduced by James Baumgartner and Peter Dordal and later developed independently by the third author to show that adding a Hechler real has strong combinatorial consequences. We prove: 1) assuming omega_1^V = omega_1^L, there is no real in V[d] which is eventually different from the reals in L[d], where d is Hechler over V; 2) adding one Hechler real makes the invariants on the left-hand side of Cicho'n's diagram equal omega_1 and those on the right-hand side equal 2^omega and produces a maximal almost disjoint family of subsets of omega of size omega_1; 3) there is no perfect set of random reals over V in V[r][d], where r is random over V and d Hechler over V[r], thus answering a question of the first and second authors. As an intermediate step in the proof of 3) we show that given models M subseteq N of ZFC such that there is a perfect set of random reals in N over M, either there is a dominating real in N over M or mu (2^omega cap M) = 0 in N.
math.LO↗
Perfect sets of random reals
We discuss the relationship between perfect sets of random reals, dominating reals, and the product of two copies of the random algebra B. Recall that B is the algebra of Borel sets of 2^omega modulo the null sets. Also given two models M subseteq N of ZFC, we say that g in omega^omega cap N is a dominating real over M iff forall f in omega^omega cap M there is m in omega such that forall n geq m (g(n) > f(n)); and r in 2^omega cap N is random over M iff r avoids all Borel null sets coded in M iff r is determined by some filter which is B-generic over M. We show that there is a ccc partial order P which adds a perfect set of random reals without adding a dominating real, thus answering a question asked by the second author in joint work with T. Bartoszynski and S. Shelah some time ago. The method of the proof of this result yields also that B times B does not add a dominating real. By a different argument we show that B times B does not add a perfect set of random reals (this answers a question that A. Miller asked during the logic year at MSRI).
math.LO↗
Amoeba-absoluteness and projective measurability
We study the relationship between Amoeba forcing (the partial order which generically adds a measure one set of random reals) and projective measurability. Given a universe V of set theory and a forcing notion P in V we say that V is Sigma^1_n - P - absolute iff for every Sigma^1_n-sentence phi with parameters in V we have V models phi iff V^P models phi. We show that Sigma^1_4-Amoeba-absoluteness implies that forall a in omega^omega (omega_1^{L[a]} < omega_1^V), and hence Sigma^1_3-measurability. This answers a question of Haim Judah (private communication).
math.LO↗