---

Subject and independence

Role. An independent reviewer in a fresh context, commissioned for refutation and given only the assignment. The reviewer took no part in writing the page, its inputs or any page of its folder, read no other review, consulted no material outside the repository and ran no search.

Subject. Path wiki/research/erdos_501/glazer_lemma_2_2_reconstruction.md as it stood at 2026-09-28T05:03:27Z, read in full as of that time.

Artifact. The eight-page PDF glazer_2026_erdos_problem_501_after_adding_random_reals.pdf under the library card Glazer (2026), the author's draft rev10; its printed page numbers coincide with its physical pages. Physical pages 2, 3 and 4 were read in full, in the text layer and in page images rendered at 130 dots per inch and read line by line: page 3 for Lemma 2.2, its displays (2.4) and (2.5) and its proof; page 2 for the Section 2 setting, the sections $E_t$ and $E^s$ and the hypotheses of Lemma 2.1, which the statement imports; page 4 for the recursion (3.9)–(3.10) of Theorem 3.2 that the page's Boundary paragraph describes. Page 1 was read in the text layer only, for the title and the paper's decomposition (1.1).

Allowed material read. The input page Lemma 2.1 as of the same time (see the exposures); the library card's provenance paragraph; the Statement paragraph of Problem 501; docs/verification.md "Whole-claim report" and "Audit checklist", with the canonical failure-mode list that the checklist instances; docs/evidence.md "Source fidelity"; docs/math_authoring.md in full.

Exposures. Four, none bearing on the mathematics reviewed:

1. The whole library card was read, not only its provenance paragraph. The card carries a "Read status" paragraph, an overview of the paper's results, the companion formalization's records and a "Relation to E501" section with the acceptance record of 2026-09-27, the site label for the problem and the words "the result behind the page-level status". None of it concerns the proof of Lemma 2.2, and none of it was used.
2. The Lemma 2.1 input page was read in full, including its Standing paragraph (which declares the same author-recorded standing that the reviewed page declares), its Proof and its Boundary. Only its Definitions and Statement were used; the Lemma 2.2 proof consumes Lemma 2.1's hypotheses and definitions, not its conclusion.
3. A structural scan of the problem page printed the first line of its Status paragraph, which begins "Not disprovable"; the paragraph and the frontmatter were not read.
4. The file names of the research folder were listed. Its _index.md, its evidence folders, every other page in it and every other review were not opened.

Restatement

Work in ZFC. Let $(S,\Sigma,\mu)$ be a $\sigma$-finite measure space, let $E\subseteq S^2$ belong to the product $\sigma$-algebra $\Sigma\otimes\Sigma$, and for $t,s\in S$ write

$$E_t=\{s\in S:(t,s)\in E\},\qquad E^s=\{t\in S:(t,s)\in E\},$$

both members of $\Sigma$. Assume the hypotheses of the source's Lemma 2.1: $\mu(S)=\infty$, a constant $K<\infty$ (necessarily $K\ge0$, since $S$ is nonempty), and the column bound $\mu(E^s)\le K$ for every $s\in S$ (the source's (2.1)). Assume in addition a map $x\colon S\to\mathbb R$ that is $\Sigma$-measurable for the Borel sets of $\mathbb R$ and whose every fiber is null: for every real $a$, $\mu(x^{-1}(\{a\}))=0$ (the source's (2.4)). Then for every $C\in\Sigma$ with $\mu(C)=\infty$ and for every single point $t$ of

$$Q(C)=\{t\in C:\mu(C\setminus E_t)=\infty\}$$

(the source's (2.2)), the set

$$C'=C\setminus\bigl(E_t\cup E^t\cup x^{-1}(\{x(t)\})\bigr)$$

(the source's (2.5)) belongs to $\Sigma$ and satisfies $\mu(C')=\infty$.

Scope qualifications and conventions. The conclusion is pointwise in $t$: it holds for each $t\in Q(C)$, with no almost-every exception. The fiber removed is that of the value $x(t)$ itself. Measures take values in $[0,\infty]$. "Measurable" for $E$ means $\Sigma\otimes\Sigma$-measurable, the reading fixed in the Definitions of the Lemma 2.1 page and inherited here; "measurable" for $x$ means $\Sigma$-measurable into the Borel sets. The hypothesis $\mu(S)=\infty$ is carried from Lemma 2.1 and not used; the positivity conclusion of Lemma 2.1 is not used either, only the definition of $Q(C)$.

Checklist

• Quantifiers and scope. Pass. The source and the page quantify identically: every $a\in\mathbb R$ in (2.4), every measurable $C$ of infinite measure, every $t\in Q(C)$, with the conclusion for each such $t$ and no exceptional set. The reconstructed proof never passes to an almost-every statement, and the boundary value $\mu(C')=\infty$ is the exact conclusion, not a limit.
• Circularity. Pass. The proof uses the definition of $Q(C)$, the column bound, the null-fiber hypothesis and measurability; it does not use the conclusion of Lemma 2.1 or any statement equivalent to its own.
• Model and convention changes. Pass. The objects are the source's: the same $E$, the same sections with the same orientation, the same $Q(C)$ and the same $C'$. The one specialization, "measurable" read as $\Sigma$-measurable, is the standard meaning; F2 asks the page to record it. Under it the section fact applies to every $t$.
• Finite and statistical overreach. Inapplicable: no finite cases, averages or samples occur.
• Uniformity. Pass. The only constant is $K$, which the column bound supplies uniformly in $s$ and which the proof uses at the one point $s=t$; no limit, sum or integral is exchanged.
• Extremal conclusions. Inapplicable: no infimum, supremum, attained value or sharpness is claimed; the conclusion is the exact value $\mu(C')=\infty$, rederived below.
• Consequences and composition. Pass. The single "hence", from $\mu(C\setminus E_t)\le\mu(C')+\mu(N)$ to $\mu(C')=\infty$, was rederived. The consumed clauses are supplied at their actual strength: $t\in Q(C)$ gives $\mu(C\setminus E_t)=\infty$ by definition, (2.1) at $s=t$ gives $\mu(E^t)\le K<\infty$, and (2.4) at $a=x(t)$ gives a null fiber. The Boundary sentence was checked against page 4 (F4).
• Computation. Inapplicable: the page carries no computation.
• Reproduction. Inapplicable: the page states no rerun command and no coverage claim.
• Source and verdict fidelity. Pass with notes. The statement, the labels (2.4) and (2.5), the lemma's label and name and the physical page were checked against the page image; the source's cross-reference text "theorem 2.1" is read as Lemma 2.1 (F2); the Standing paragraph claims author-recorded standing and nothing more.

Weakest steps

1. Measurability of $C'$ (the supplied step). The class of sets $A\subseteq S^2$ all of whose sections $A_t$ and $A^s$ lie in $\Sigma$ contains every measurable rectangle $B\times D$ (its sections are $D$ or $\emptyset$, and $B$ or $\emptyset$) and is closed under complements and countable unions, because taking a section commutes with both. So it contains $\Sigma\otimes\Sigma$, and $E_t,E^t\in\Sigma$ for every $t$. The singleton $\{x(t)\}$ is closed in $\mathbb R$, so $x^{-1}(\{x(t)\})\in\Sigma$ by the measurability of $x$. Since $C\in\Sigma$ and $\Sigma$ is closed under finite unions and differences, $C'\in\Sigma$. This composes with the rest by making $\mu(C\setminus E_t)$, $\mu(E^t)$, the measure of the fiber and $\mu(C')$ defined; the source's proof presupposes all four.

2. Finiteness of the removed part. Put $F=x^{-1}(\{x(t)\})$ and $N=E^t\cup F$. Monotonicity and finite subadditivity give $\mu(N)\le\mu(E^t)+\mu(F)\le K+0=K<\infty$, using (2.1) at $s=t$, (2.4) at $a=x(t)$ and $K<\infty$ from Lemma 2.1's hypotheses. Nothing else about $N$ is needed; in particular $N$ need not be disjoint from $C$ or from $E_t$.

3. Transfer of infinite measure. From $C'=(C\setminus E_t)\setminus N$ (removing $E_t\cup N$ in one step or in two steps gives the same set),

$$C\setminus E_t=C'\cup\bigl((C\setminus E_t)\cap N\bigr)\subseteq C'\cup N,$$

so $\infty=\mu(C\setminus E_t)\le\mu(C')+\mu(N)\le\mu(C')+K$ in $[0,\infty]$. If $\mu(C')$ were finite the right side would be finite, contradicting $t\in Q(C)$; hence $\mu(C')=\infty$. This is exactly the source's sentence "Removing their union leaves infinite measure", and it is what Theorem 3.2's recursion needs to restart Lemma 2.1 on $C'$.

Strongest attack

The attack aimed at the measurability clause through the word "measurable". If $E$ were measurable only for the completion of $\mu\times\mu$, or $x$ only for the completion of $\mu$, then the sections $E_t$, $E^t$ and the fiber would be $\mu$-measurable for almost every $t$ but could fail to lie in $\Sigma$, or even to be $\mu$-measurable, at particular points; since the conclusion is asserted for every single $t\in Q(C)$, a bad $t$ would refute the measurability clause as the page states it. The attack fails against the page as written: the page fixes, through the Definitions of the Lemma 2.1 page, the reading $E\in\Sigma\otimes\Sigma$ with $x$ $\Sigma$-measurable, which is the standard meaning of "measurable" in a measure space $(S,\Sigma,\mu)$ and the one the source's own proof needs, since it evaluates $\mu(E^t)$ at the given $t$; under that reading the section fact of weakest step 1 holds at every $t$, not almost every $t$. In the source's only application (page 4) the graph $E$ and the map $x$ are Borel on $\mathbb Z\times\Omega$, so the question does not arise there either. A second attempt looked for an exceptional $t\in Q(C)$ with $\mu(C\setminus E_t)=\infty$ but $C'$ of finite measure; the inclusion $C\setminus E_t\subseteq C'\cup N$ with $\mu(N)\le K$ leaves no room for one. A third looked at the boundary $K=0$: then every column and the fiber are null, and the argument is unchanged.

Premises

• Lemma 2.1 (local input). Lemma 2.1 as of the same time, whole page read, Definitions and Statement used. Consumed interface: the setting $(S,\Sigma,\mu)$, the sections $E_t$ and $E^s$ with the source's orientation and their membership in $\Sigma$, the hypothesis list ($\mu(S)=\infty$, $K<\infty$, $E\in\Sigma\otimes\Sigma$, the column bound (2.1)) and the definition (2.2) of $Q(C)$. Not consumed: the conclusion $\mu(Q(C))>0$, so the soundness of Lemma 2.2 does not rest on Lemma 2.1's proof. Standing: the page declares itself author-recorded; no other standing text was read.
• The source (held). Glazer, draft rev10, Lemma 2.2 with displays (2.4) and (2.5) and its proof, physical page 3, read in full in the text layer and in the page image; page 2 for the setting and Lemma 2.1's hypotheses and page 4 for the recursion (3.9)–(3.10), read the same way. The page's statement matches the source clause by clause.
• Measurability of sections (standard, no held source). Every section of a $\Sigma\otimes\Sigma$-measurable set lies in $\Sigma$; rederived in weakest step 1. Stated in the Definitions of the Lemma 2.1 page; not named in the reviewed page's Standing paragraph (F3).
• Measurable maps (standard, no held source). The preimage of a Borel set, here a singleton, under a $\Sigma$-measurable map lies in $\Sigma$.
• Measure axioms (standard, no held source). Monotonicity and finite subadditivity of $\mu$ on $[0,\infty]$.
• Explicit assumptions. The readings of "measurable" recorded in the Restatement; $K<\infty$; ZFC as the ambient theory, as the source states.

Findings

F1. Severity: suggested. Location: "The sections $E_t$ and $E^t$ of the measurable set $E$ lie in $\Sigma$ ... So $C'\in\Sigma$." Defect: a supplied step not labeled as such. The source's statement asserts that $C'$ is measurable, and its three-sentence proof (page 3) argues only the measure; the page adds the justification of measurability without marking it. Witness: page 3, the proof of Lemma 2.2, which reads in full "The set $C\setminus E_t$ has infinite measure. By (2.1), $\mu(E^t)\le K$, and the last set in (2.5) is null. Removing their union leaves infinite measure." Proposed replacement: open the paragraph with "Measurability (supplied). The source asserts that $C'$ is measurable and its proof does not argue it. The sections ..." and leave the rest unchanged.

F2. Severity: note. Location: "In addition to the hypotheses of Lemma 2.1, let $x\colon S\to\mathbb R$ be $\Sigma$-measurable". Two readings are silent. The source's cross-reference prints "theorem 2.1", a label artifact (Lemma 2.1 is the only result numbered 2.1, and page 4 writes "theorems 2.1 and 2.2" for the two lemmas); and the source writes "measurable", which the page specializes to $\Sigma$-measurable. Both readings are correct. Witness: page 3, the first two lines of Lemma 2.2. Proposed replacement: "In addition to the hypotheses of Lemma 2.1 (the source's cross-reference prints "theorem 2.1"), let $x\colon S\to\mathbb R$ be measurable, read as $\Sigma$-measurable for the Borel sets of $\mathbb R$, with every fiber null:".

F3. Severity: note. Location: the Standing paragraph, "This is an author-recorded reconstruction. ... assigns no tier." The paragraph names no external input, while the proof rests on the measurability of the sections of a $\Sigma\otimes\Sigma$-measurable set and on finite subadditivity; the Lemma 2.1 page names its one external input in the same place. Witness: the page's Standing paragraph against its Proof. Proposed addition, after the last sentence: "The only external inputs are the measurability of the sections of a $\Sigma\otimes\Sigma$-measurable set, stated in the Definitions of Lemma 2.1, and finite subadditivity of $\mu$."

F4. Severity: note. Location: Boundary, "The lemma is the inductive step of the recursion in Theorem 3.2". On page 4 the step from $C_j$ to $C_{j+1}$ has two halves: Lemma 2.1 and $\nu^*(Z)=1$ choose $t_j$ through the set $H_j$ of (3.9), and Lemma 2.2 keeps $C_{j+1}$ of (3.10) measurable and infinite. Lemma 2.2 is the preservation half, as the page's own title says. The second clause of the sentence, on the fiber removal, is the source's own sentence on page 4 and was rederived: for $i<j$, $t_j\in C_{i+1}$ avoids the fiber of $x(t_i)$, so $x(t_j)\ne x(t_i)$. Witness: page 4, displays (3.9) and (3.10) and the sentence "The fiber removal makes the $y_j$ pairwise distinct." Proposed replacement: "The lemma is the preservation half of the inductive step of the recursion in Theorem 3.2 (display (3.10) there); Lemma 2.1 supplies the selection half; the fiber removal there is what makes the selected reals pairwise distinct."

Verdict

Source fidelity: faithful. The statement, its hypotheses, quantifiers, displays (2.4) and (2.5), the lemma label and the physical page all match the held draft. The argument as reconstructed: sound; every step was rederived above, and the one supplied step is correct. Required corrections: none; one suggested label (F1) and three notes (F2–F4).

Limitations. The review covers Lemma 2.2's statement and proof, the hypotheses it imports from Lemma 2.1 and the page's Boundary sentence, against physical pages 2–4 of the held draft. It does not examine the proof of Lemma 2.1, Theorem 3.2, the forcing sections, the companion Lean development or any standing or status text. The section-measurability fact and the measure axioms are taken as standard measure theory with no held source.

This focused review assigns no tier and changes no status.