Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated

../


Source. E. Glazer, Erdős Problem 501 after adding ω2\omega_2 random reals, draft rev10, Proposition 4.4 (homogeneous Borel reading), physical pp. 5--6, in the eight-page PDF held by its library source card, Glazer (2026). It uses Lemma 4.1 and Lemma 4.3.

Standing. This is an author-recorded reconstruction. It is not an independent review and changes no status and assigns no tier. Imported: the conventions (R1)--(R5) on the Lemma 4.1 page, and the fact that there are at most 2ℵ02^{\aleph_0} Borel maps between two standard Borel spaces (a Borel map is determined by the preimages of a countable base, and there are 2ℵ02^{\aleph_0} Borel sets; A. S. Kechris, Classical Descriptive Set Theory, Chapter 11).

Definitions

Let κ=ω2\kappa=\omega_2, Θ=κ×ω\Theta=\kappa\times\omega, and Dα={α}×ωD_\alpha=\{\alpha\}\times\omega for α<κ\alpha<\kappa; the blocks DαD_\alpha partition Θ\Theta. For a bijection π ⁣:P→P′\pi\colon P\to P' between coordinate sets and v∈2P′v\in2^{P'}, write v∘π∈2Pv\circ\pi\in2^P for the pullback; the source writes it π−1(v)\pi^{-1}(v). For a pair (P′,D′)(P',D') with D′⊆P′D'\subseteq P' countable and a fixed enumeration of D′D', its isomorphism type is determined by the cardinality of P′∖D′P'\setminus D', one of 0,1,2,…,ℵ00,1,2,\dots,\aleph_0; two pairs of the same type admit a bijection matching the enumerations of the distinguished subsets.

Statement

The following is provable in ZFC + CH. Let p∈B(Θ)p\in\mathbb B(\Theta), let XX be a standard Borel space, and for each α<κ\alpha<\kappa let w˙α\dot w_\alpha be a name for an element of XX. There are

  • a set J⊆κJ\subseteq\kappa of size κ\kappa;
  • a countable root RR supporting pp;
  • pairwise disjoint countable petals PαP_\alpha with Dα⊆PαD_\alpha\subseteq P_\alpha, for α∈J\alpha\in J;
  • a fixed countable pair D⊆PD\subseteq P with an enumeration ⟨dn:n<ω⟩\langle d_n:n<\omega\rangle of DD, and bijections πα ⁣:P→Pα\pi_\alpha\colon P\to P_\alpha with πα(dn)=(α,n)\pi_\alpha(d_n)=(\alpha,n);
  • a single Borel map F ⁣:2R×2P→XF\colon2^R\times2^P\to X

such that

⊩B(Θ)w˙α=F(G˙↾R, (G˙↾Pα)∘πα)(α∈J)\Vdash_{\mathbb B(\Theta)}\dot w_\alpha =F\bigl(\dot G\restriction R,\ (\dot G\restriction P_\alpha)\circ\pi_\alpha\bigr) \qquad(\alpha\in J)

(the source's (4.2)).

Proof

Supports. By (R1) choose a countable R0⊆ΘR_0\subseteq\Theta supporting pp. By Lemma 4.1, for each α<κ\alpha<\kappa choose a countable Sα⊆ΘS_\alpha\subseteq\Theta reading w˙α\dot w_\alpha through a Borel map Fα ⁣:2Sα→XF_\alpha\colon2^{S_\alpha}\to X, and enlarge SαS_\alpha to contain R0∪DαR_0\cup D_\alpha (a reading survives enlargement, as noted on the Lemma 4.1 page). The sets SαS_\alpha are countable. They need not be pairwise distinct, since two names may share a support: Lemma 4.3 as reconstructed applies to the sequence ⟨Sα:α<κ⟩\langle S_\alpha:\alpha<\kappa\rangle as it stands, and a member repeated inside the Δ\Delta-subsystem below equals the root RR, so its block lies in RR and its index is discarded under "Blocks inside petals".

Delta-system. Apply Lemma 4.3 to ⟨Sα:α<κ⟩\langle S_\alpha:\alpha<\kappa\rangle: there are J0⊆κJ_0\subseteq\kappa of size κ\kappa and a countable RR with Sα∩Sβ=RS_\alpha\cap S_\beta=R for distinct α,β∈J0\alpha,\beta\in J_0. Since R0⊆SαR_0\subseteq S_\alpha for all α\alpha, R0⊆RR_0\subseteq R, so RR supports pp. Put Pα=Sα∖RP_\alpha=S_\alpha\setminus R for α∈J0\alpha\in J_0; for distinct α,β∈J0\alpha,\beta\in J_0, Pα∩Pβ=(Sα∩Sβ)∖R=∅P_\alpha\cap P_\beta=(S_\alpha\cap S_\beta)\setminus R=\varnothing.

Blocks inside petals. The countable set RR meets only countably many of the pairwise disjoint blocks DαD_\alpha. Let J1J_1 be J0J_0 with those indices removed; ∣J1∣=κ|J_1|=\kappa, and for α∈J1\alpha\in J_1, Dα∩R=∅D_\alpha\cap R=\varnothing, so Dα⊆Sα∖R=PαD_\alpha\subseteq S_\alpha\setminus R=P_\alpha.

One isomorphism type. The pairs (Pα,Dα)(P_\alpha,D_\alpha), with DαD_\alpha enumerated as ⟨(α,n):n<ω⟩\langle(\alpha,n):n<\omega\rangle, fall into countably many isomorphism types. A set of size ω2\omega_2 is not a countable union of sets of size at most ω1\omega_1, so some type contains κ\kappa many indices; let J2⊆J1J_2\subseteq J_1 be those indices. Fix a countable pair D⊆PD\subseteq P of that type with an enumeration ⟨dn⟩\langle d_n\rangle of DD, and for α∈J2\alpha\in J_2 fix a bijection πα ⁣:P→Pα\pi_\alpha\colon P\to P_\alpha with πα(dn)=(α,n)\pi_\alpha(d_n)=(\alpha,n).

Pulling back. For α∈J2\alpha\in J_2, Sα=R⊔PαS_\alpha=R\sqcup P_\alpha, so 2Sα2^{S_\alpha} is identified with 2R×2Pα2^R\times2^{P_\alpha}. Define F~α ⁣:2R×2P→X\tilde F_\alpha\colon2^R\times2^P\to X by

F~α(u,v)=Fα(u∪(v∘πα−1)),\tilde F_\alpha(u,v)=F_\alpha\bigl(u\cup(v\circ\pi_\alpha^{-1})\bigr),

a Borel map, being FαF_\alpha composed with the homeomorphism (u,v)↦u∪(v∘πα−1)(u,v)\mapsto u\cup(v\circ\pi_\alpha^{-1}) of 2R×2P2^R\times2^P onto 2Sα2^{S_\alpha}. Since G˙↾Sα=(G˙↾R)∪(G˙↾Pα)\dot G\restriction S_\alpha=(\dot G\restriction R)\cup(\dot G\restriction P_\alpha) and G˙↾Pα=((G˙↾Pα)∘πα)∘πα−1\dot G\restriction P_\alpha=((\dot G\restriction P_\alpha)\circ\pi_\alpha)\circ\pi_\alpha^{-1},

⊩w˙α=Fα(G˙↾Sα)=F~α(G˙↾R,(G˙↾Pα)∘πα).\Vdash\dot w_\alpha=F_\alpha(\dot G\restriction S_\alpha) =\tilde F_\alpha\bigl(\dot G\restriction R, (\dot G\restriction P_\alpha)\circ\pi_\alpha\bigr).

Counting. The maps F~α\tilde F_\alpha (α∈J2\alpha\in J_2) all belong to the set of Borel maps 2R×2P→X2^R\times2^P\to X, which has size at most 2ℵ0=ℵ12^{\aleph_0}=\aleph_1 under CH. Since ∣J2∣=ℵ2|J_2|=\aleph_2, some Borel map FF equals F~α\tilde F_\alpha for every α\alpha in a set J⊆J2J\subseteq J_2 of size κ\kappa. This JJ, RR, (Pα)α∈J(P_\alpha)_{\alpha\in J}, D⊆PD\subseteq P, (πα)α∈J(\pi_\alpha)_{\alpha\in J} and FF satisfy the statement.

Boundary. CH enters twice: through Lemma 4.3 and through the count of Borel maps. The proposition is applied in Theorem 5.1 with X=OZX=\mathcal O^{\mathbb Z}.