Wiki
Wiki

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

Updated

../


Source. S. Lee, Relative independence of Erdős problem #501, second version dated 2026-06-01, Lemma 2.1 (statement, physical p. 2; proof, pp. 4--5), in the six-page PDF held by its library source card, Lee (2026). Its one input is the section inequality reconstructed in Lemma 3.1.

Standing. This is an author-recorded reconstruction. It is not an independent review and changes no status and assigns no tier.

Definitions

mm, m∗m^* and the upper integral are as on the Lemma 3.1 page. Let ν ⁣:P(R)→[0,∞]\nu\colon\mathcal P(\mathbb R)\to[0,\infty] be a countably additive measure extending Lebesgue measure. Two facts about ν\nu are used.

  • (1) ν(S)≤m∗(S)\nu(S)\le m^*(S) for every S⊆RS\subseteq\mathbb R (the source's (1)): if S⊆⋃iJiS\subseteq\bigcup_iJ_i with open intervals JiJ_i, then ν(S)≤∑iν(Ji)=∑im(Ji)\nu(S)\le\sum_i\nu(J_i)=\sum_im(J_i), and m∗(S)m^*(S) is the infimum of the right side over all such covers. In particular m∗(S)<1m^*(S)<1 implies ν(S)<1\nu(S)<1.
  • The upper-integral identity ∫R‾a1S dm=a m∗(S)\overline{\int_{\mathbb R}}a1_S\,dm=a\,m^*(S) for a≥0a\ge0 and S⊆RS\subseteq\mathbb R: a measurable T⊇ST\supseteq S gives the majorant a1Ta1_T with integral a m(T)a\,m(T), and every measurable majorant h≥a1Sh\ge a1_S has T={h≥a}⊇ST=\{h\ge a\}\supseteq S measurable and ∫h≥a m(T)≥a m∗(S)\int h\ge a\,m(T)\ge a\,m^*(S) when a>0a>0; for a=0a=0 both sides vanish.

For a family (Ay)y∈R(A_y)_{y\in\mathbb R} of subsets of R\mathbb R and x∈Rx\in\mathbb R, put Bx={y∈R:x∈Ay}B_x=\{y\in\mathbb R:x\in A_y\}, the set of indices whose sets contain xx.

Statement

Let ν\nu be as above and let (Ay)y∈R(A_y)_{y\in\mathbb R} satisfy m∗(Ay)<1m^*(A_y)<1 for every y∈Ry\in\mathbb R. If C⊆RC\subseteq\mathbb R and ν(C)=∞\nu(C)=\infty, then there is an x∈Cx\in C with ν(C∖Bx)=∞\nu(C\setminus B_x)=\infty.

Proof

Suppose, toward a contradiction, that

ν(C∖Bx)<∞for every x∈C\nu(C\setminus B_x)<\infty\qquad\text{for every }x\in C

(the source's (12)).

A window of measure above one. For N≥1N\ge1 put CN=C∩[−N,N]C_N=C\cap[-N,N]. The sets CNC_N increase to CC, so continuity from below gives ν(CN)→ν(C)=∞\nu(C_N)\to\nu(C)=\infty (the source's (13)). Fix NN with D:=CND:=C_N satisfying ν(D)>1\nu(D)>1. For k≥1k\ge1 define

Dk={x∈D:ν(C∖Bx)≤k}D_k=\{x\in D:\nu(C\setminus B_x)\le k\}

(the source's (14)). By the assumption (12), D=⋃k≥1DkD=\bigcup_{k\ge1}D_k, and the DkD_k increase, so continuity from below gives k≥1k\ge1 with ν(Dk)>1\nu(D_k)>1. By (1) and Dk⊆[−N,N]D_k\subseteq[-N,N],

1<ν(Dk)≤m∗(Dk)≤2N<∞1<\nu(D_k)\le m^*(D_k)\le2N<\infty

(the source's (15)).

A larger window. For every integer M>NM>N, ν(CM)≤ν([−M,M])=2M<∞\nu(C_M)\le\nu([-M,M])=2M<\infty. By (13) and (15), choose an integer M>NM>N with

ν(CM)>k m∗(Dk)m∗(Dk)−1\nu(C_M)>\frac{k\,m^*(D_k)}{m^*(D_k)-1}

(the source's (16)); the right side is a finite positive number by (15). Since m∗(Dk)/(m∗(Dk)−1)>1m^*(D_k)/(m^*(D_k)-1)>1, in particular ν(CM)>k\nu(C_M)>k.

Lower bound on the vertical sections. Let x∈Dkx\in D_k. Then ν(C∖Bx)≤k\nu(C\setminus B_x)\le k; as CMC_M has finite ν\nu-measure and CM∖Bx⊆C∖BxC_M\setminus B_x\subseteq C\setminus B_x,

ν(Bx∩CM)=ν(CM)−ν(CM∖Bx)≥ν(CM)−k\nu(B_x\cap C_M)=\nu(C_M)-\nu(C_M\setminus B_x)\ge\nu(C_M)-k

(the source's (17)). Define

H={(x,y)∈Dk×CM:x∈Ay}⊆R2.H=\{(x,y)\in D_k\times C_M:x\in A_y\}\subseteq\mathbb R^2.

For x∈Dkx\in D_k, Hx={y∈CM:x∈Ay}=Bx∩CMH_x=\{y\in C_M:x\in A_y\}=B_x\cap C_M, and Hx=∅H_x=\varnothing for x∉Dkx\notin D_k. So ν(Hx)≥(ν(CM)−k)1Dk(x)\nu(H_x)\ge(\nu(C_M)-k)1_{D_k}(x) for every x∈Rx\in\mathbb R, and by monotonicity of the upper integral and the identity for a1Sa1_S with a=ν(CM)−k≥0a=\nu(C_M)-k\ge0,

∫R‾ν(Hx) dm(x)≥(ν(CM)−k) m∗(Dk)\overline{\int_{\mathbb R}}\nu(H_x)\,dm(x)\ge(\nu(C_M)-k)\,m^*(D_k)

(the source's (18)).

Upper bound on the horizontal sections. For y∈CMy\in C_M, Hy=Ay∩DkH^y=A_y\cap D_k, and Hy=∅H^y=\varnothing for y∉CMy\notin C_M. Since m∗(Ay)<1m^*(A_y)<1 for every yy, m∗(Hy)≤1CM(y)m^*(H^y)\le1_{C_M}(y) for every y∈Ry\in\mathbb R. The function y↦m∗(Hy)y\mapsto m^*(H^y) is ν\nu-measurable, ν\nu being defined on all subsets, so

∫Rm∗(Hy) dν(y)≤∫R1CM dν=ν(CM)\int_{\mathbb R}m^*(H^y)\,d\nu(y)\le\int_{\mathbb R}1_{C_M}\,d\nu=\nu(C_M)

(the source's (19)).

Contradiction. The section inequality (11) of Lemma 3.1 applied to HH, with (18) and (19), gives

(ν(CM)−k) m∗(Dk)≤ν(CM).(\nu(C_M)-k)\,m^*(D_k)\le\nu(C_M).

Dividing by ν(CM)\nu(C_M), which is positive and finite,

(1−kν(CM))m∗(Dk)≤1.\Bigl(1-\frac{k}{\nu(C_M)}\Bigr)m^*(D_k)\le1.

But (16) is equivalent, after multiplying by m∗(Dk)−1>0m^*(D_k)-1>0 and dividing by ν(CM)\nu(C_M), to

(1−kν(CM))m∗(Dk)>1,\Bigl(1-\frac{k}{\nu(C_M)}\Bigr)m^*(D_k)>1,

a contradiction. Hence (12) fails: some x∈Cx\in C has ν(C∖Bx)=∞\nu(C\setminus B_x)=\infty.

Boundary. Only the values ν(C∖Bx)\nu(C\setminus B_x), ν(CM)\nu(C_M) and the outer measures m∗(Dk)m^*(D_k), m∗(Ay)m^*(A_y) enter; no measurability of the sets AyA_y or BxB_x for Lebesgue measure is assumed. The lemma drives the recursion in Theorem 1.1.