---

Subject and independence

Role: independent reviewer in a fresh context, commissioned for refutation and given only the assignment text. The reviewer took no part in writing the page or any page in its folder, had not read the folder before this review, and consulted no other review.

Subject: path wiki/research/erdos_501/glazer_theorem_1_1_reconstruction.md as it stood at 2026-09-28T05:03:27Z, read in full as of that time; it is the Theorem 1.1 page.

Artifact: the eight-page PDF held by Glazer (2026), draft rev10, whose printed page numbers equal its physical page numbers. Read in full, from the text layer and from page images rendered at 130 dpi: p. 1 (abstract, Theorem 1.1, display (1.1), Corollary 1.2), p. 3 (Definition 3.1, Theorem 3.2 and display (3.5)), p. 6 (Theorem 5.1 and display (5.1), the opening of its proof), p. 7 (the end of the proof of Theorem 5.1 and the proof of Theorem 1.1) and p. 8 (Section 6: the proof of Corollary 1.2, the CH counterexample, the units F1--F6 and the references). Pages 2, 4 and 5 were read from the text layer only, for the definition of $\mathbb B(\Theta)$ (p. 4) and the coordinate set $\Theta=\kappa\times\omega$ of Proposition 4.4 (p. 5). The PDF prints "Proof of theorem 1.2" on p. 8 for the result labeled "Corollary 1.2" on p. 1, a cross-reference artifact; the page's label follows the statement.

Allowed material read: the Definitions and Statement sections of the Theorem 3.2 page and of the Theorem 5.1 page, and the Statement section of the CH counterexample page, all as of the same time and none of their proofs; the Statement paragraph of Problem 501; the provenance paragraph of the library card; docs/verification.md "Whole-claim report" and "Audit checklist", docs/evidence.md "Source fidelity" and docs/math_authoring.md. No evidence folder, no other review, nothing among the private working files, and no web search.

Exposures: two, both by over-wide reads. (1) The library card was printed whole, so its Read status, Overview, Companion formalization and Relation to E501 sections, including the acceptance paragraph, reached the reviewer; of these, only the sentence describing the formalized positive model was used, to confirm the page's Boundary paragraph, as noted under Findings. (2) The problem page's Status and Source paragraphs, which sit directly under its Statement paragraph, reached the reviewer; nothing from them was used.

Restatement

Conventions. $\lambda^*$ is Lebesgue outer measure on $\mathbb R$. For a family $\mathcal A=(A_y)_{y\in\mathbb R}$ of subsets of $\mathbb R$, $\mathrm{Free}_\omega(\mathcal A)$ says that some infinite $X\subseteq\mathbb R$ has $x\notin A_y$ for all distinct $x,y\in X$. $\mathrm{Prof}(\mathcal A)$ says that a profile certificate exists: a standard Borel probability space $(\Omega,\nu)$, a set $Z\subseteq\Omega$ with $\nu^*(Z)=1$, Borel maps $x_m\colon\Omega\to[m,m+1)$ with Lebesgue distribution, and Borel codes $c_m$ with $\lambda(U(c_m(z)))<1$ for every $z\in\Omega$, such that $A_{x_m(z)}\subseteq U(c_m(z))$ for every $z\in Z$ and every $m\in\mathbb Z$. For a coordinate set $\Theta$, $\mathbb B(\Theta)$ is the measure algebra of the completed product measure on $2^\Theta$, and $\mathbb B_{\omega_2}$ is $\mathbb B(\omega_2\times\omega)$, the algebra fixed in the source's proof of Theorem 5.1; "$\mathbb B\Vdash\varphi$" means that the top element forces $\varphi$. $P$ is the sentence: for every family $(A_y)_{y\in\mathbb R}$ of bounded subsets of $\mathbb R$ with $\lambda^*(A_y)<1$ for every $y$, $\mathrm{Free}_\omega(\mathcal A)$ holds; this is the first question of Problem 501 answered positively.

Theorem 1.1. Let $M$ be a model of ZFC + CH, let $\kappa$ be the $\omega_2$ of $M$, and let $G$ be a filter on the measure algebra $\mathbb B(\kappa\times\omega)$ as computed in $M$ that is generic over $M$. Then in $M[G]$: for every function assigning to each real $y$ of $M[G]$ a subset $A_y$ of the reals of $M[G]$, if the outer measure of $A_y$, computed in $M[G]$, is below $1$ for every $y$, then some infinite set $X$ of reals of $M[G]$ satisfies $x\notin A_y$ for all distinct $x,y\in X$. No boundedness of the $A_y$ is assumed.

Corollary 1.2. If ZFC is consistent, then ZFC + $P$ is consistent and ZFC + $\neg P$ is consistent.

Checklist

Canonical failure modes:

• "Almost all" upgraded to "all". Not found. The page has no almost-everywhere sentence; the outer-measure-one set $Z$ lives inside the inputs and is not manipulated here.
• Induction presupposing termination. Inapplicable: no induction or recursion on the page.
• Probabilistic or averaging heuristic as proof. Inapplicable: no probabilistic argument on the page.
• Circular use of an equivalent statement. Not found. Theorem 1.1 is deduced from Theorem 5.1 and Theorem 3.2, neither of which is equivalent to it; Corollary 1.2 is deduced from the syntactic form of that assembly, from CH $\to\neg P$, and from Gödel's theorem.
• Exceptional sets dropped from density arguments. Inapplicable: no density argument.
• Finite verification cited beyond base cases. Inapplicable: nothing is verified by cases.
• Relaxed or averaged system standing in for the objects. Inapplicable: no relaxation.

Named patterns:

• Model-class transport instead of entailment. Passes. The page classifies no extension by the form of its axioms; Corollary 1.2 is exactly the entailment question for $P$ over ZFC, answered by two models, a forcing extension and $L$, relative to $\mathrm{Con}(\mathrm{ZFC})$.
• Uniformity over an infinite family from finitely many instances. Inapplicable: no constants or bounds.
• Extremal claims audited in their own units. Inapplicable: no extremal sentence.
• Consequence sentences are claim surfaces. Passes with F1. Each was attacked on its own: "Hence $M[G]\models\mathrm{Free}_\omega(\mathcal A)$" holds; "in particular every family of bounded such sets does, which is $P$" holds, $P$ being the restriction to bounded families; "So $L^N\models\mathrm{ZFC}+\neg P$" holds for every model $N$ of ZFC; "Together these give the corollary" holds, since $\mathrm{Con}(\mathrm{ZFC})\to\mathrm{Con}(\mathrm{ZFC}+P)$ and $\mathrm{Con}(\mathrm{ZFC})\to\mathrm{Con}(\mathrm{ZFC}+\neg P)$ are what independence relative to $\mathrm{Con}(\mathrm{ZFC})$ means; "yields a model of ZFC in which every family ... has an infinite independent set" holds only when a filter generic over $L^N$ exists (F1).
• Carry hypotheses actually used. Passes with F1. CH in $M$ is carried into the application of Theorem 5.1; the syntactic route of the corollary carries nothing beyond $\mathrm{Con}(\mathrm{ZFC})$; the model-theoretic gloss uses the existence of a generic filter over $L^N$ without carrying it (F1).
• A composition inherits its unproved premises. Passes. Theorem 1.1 inherits the standing of the Theorem 3.2 and Theorem 5.1 pages and Corollary 1.2 inherits that of the CH counterexample page; the page's Standing paragraph says author-recorded and claims nothing more, names its two imports, and states that it assigns no tier.
• Reproducibility notes are claims. Inapplicable: no rerun line, count or harness statement on the page.
• Verifier quotations are claims. Inapplicable: the page quotes no verifier ruling. Its Boundary sentence characterizes the library card's description of the companion formalization, not a verdict; the reviewer confirmed it only through the disclosed exposure (the card records that the formalized positive model is the Boolean-valued model of the random algebra with $\mathfrak c^+$ coordinates, not the paper's $\omega_2$ random reals over a CH ground).
• Verdict words spelled in full. Inapplicable to the page, which carries no verdict word; this report writes refutation-failed in full.
• Certified-bracket functions fail loudly. Inapplicable: no numerics.
• A harness leg with no failing input is decoration. Inapplicable: no harness.
• A gate that reads caches is defective. Inapplicable: no gate or evidence program.

Weakest steps

W1. From Theorem 5.1 in $M$ to $\mathrm{Prof}(\mathcal A)$ in $M[G]$. Write $\psi$ for $\forall\mathcal A\,[(\forall y\in\mathbb R\ \lambda^*(A_y)<1)\to\mathrm{Prof}(\mathcal A)]$. Theorem 5.1 is one sentence of set theory, "the top element of $\mathbb B(\omega_2\times\omega)$ forces $\psi$", and ZFC + CH proves it. Since $M\models\mathrm{ZFC}+\mathrm{CH}$, that sentence holds in $M$, where $\mathbb B(\omega_2\times\omega)$ is evaluated as $\mathbb B^M(\kappa\times\omega)$ with $\kappa=\omega_2^M$, the algebra of the theorem's hypothesis. The forcing theorem for the $M$-generic $G$ (the top element belongs to $G$) gives $M[G]\models\psi$. Instantiating $\psi$ in $M[G]$ at a family $\mathcal A\in M[G]$ whose outer measures, computed in $M[G]$, are all below one gives $\mathrm{Prof}(\mathcal A)$ in $M[G]$. Composition: $M[G]\models\mathrm{ZFC}$ by the generic model theorem, and ZFC proves $\forall\mathcal A\,[\mathrm{Prof}(\mathcal A)\to\mathrm{Free}_\omega(\mathcal A)]$ (Theorem 3.2), so this implication holds in $M[G]$ and $M[G]\models\mathrm{Free}_\omega(\mathcal A)$. The one place a hypothesis could slip is the identification of "the measure algebra adding $\kappa$ random reals" with $\mathbb B^M(\kappa\times\omega)$: the source's proof of Theorem 5.1 (p. 6) fixes exactly $\Theta=\kappa\times\omega$ and $\mathbb B=\mathbb B(\Theta)$, and for infinite $\kappa$ a bijection of coordinate sets makes $\mathbb B(\kappa)$ and $\mathbb B(\kappa\times\omega)$ isomorphic, so nothing slips.

W2. From the two theorems to $\mathrm{Con}(\mathrm{ZFC}+\mathrm{CH})\to\mathrm{Con}(\mathrm{ZFC}+P)$. First, $\mathrm{ZFC}+\mathrm{CH}\vdash\mathbb B_{\omega_2}\Vdash P$. By Theorem 5.1, $\mathrm{ZFC}+\mathrm{CH}\vdash\mathbb B_{\omega_2}\Vdash\psi$. By Theorem 3.2 and the forcing theorem (ZFC proves, for each of its theorems, that every complete Boolean algebra forces it), $\mathrm{ZFC}\vdash\mathbb B_{\omega_2}\Vdash\forall\mathcal A\,[\mathrm{Prof}(\mathcal A)\to\mathrm{Free}_\omega(\mathcal A)]$. Forced sentences are closed under logical consequence, and $\psi$ with this implication yields $\forall\mathcal A\,[(\forall y\ \lambda^*(A_y)<1)\to\mathrm{Free}_\omega(\mathcal A)]$, which yields $P$, since $P$ only adds the hypothesis that each $A_y$ is bounded. Second, suppose $\mathrm{ZFC}+P$ is inconsistent. Then $\mathrm{ZFC}\vdash\neg P$, so $\mathrm{ZFC}\vdash\mathbb B_{\omega_2}\Vdash\neg P$, and with the first part $\mathrm{ZFC}+\mathrm{CH}$ proves that $\mathbb B_{\omega_2}$ forces $P\wedge\neg P$, that is, that its top element equals its bottom element, while ZFC proves that the measure algebra of a probability measure is nontrivial. So $\mathrm{ZFC}+\mathrm{CH}$ is inconsistent; the contrapositive is the claim. Third, $\mathrm{Con}(\mathrm{ZFC})\to\mathrm{Con}(\mathrm{ZFC}+\mathrm{CH})$: ZFC proves every axiom of ZFC relativized to $L$ together with $\mathrm{CH}^L$ (Gödel), so a derivation of a contradiction from $\mathrm{ZFC}+\mathrm{CH}$ relativizes to one from ZFC. Composition: chaining the three gives $\mathrm{Con}(\mathrm{ZFC})\to\mathrm{Con}(\mathrm{ZFC}+P)$, the first half of Corollary 1.2. This is the deduction the page's "Formally" sentence states; it is complete on its own and does not use the model-theoretic gloss before it.

W3. The $\neg P$ half and the assembly. The CH counterexample page's Statement gives, under CH, a family of countable (so outer measure $0<1$) bounded sets with no infinite independent set, a counterexample to $P$; so $\mathrm{ZFC}+\mathrm{CH}\vdash\neg P$. Every model of $\mathrm{ZFC}+\mathrm{CH}$ is then a model of $\mathrm{ZFC}+\neg P$, so $\mathrm{Con}(\mathrm{ZFC}+\mathrm{CH})\to\mathrm{Con}(\mathrm{ZFC}+\neg P)$, and with Gödel's theorem $\mathrm{Con}(\mathrm{ZFC})\to\mathrm{Con}(\mathrm{ZFC}+\neg P)$. The page's model-theoretic sentence for this half is valid as written for every model $N$ of ZFC, not only countable ones: $L^N$ is a definable inner model of $N$ satisfying $\mathrm{ZFC}+\mathrm{CH}$, and a theorem of $\mathrm{ZFC}+\mathrm{CH}$ holds in it; unlike the $P$ half it needs no generic filter. Composition with W2: $\mathrm{Con}(\mathrm{ZFC})$ implies both $\mathrm{Con}(\mathrm{ZFC}+P)$ and $\mathrm{Con}(\mathrm{ZFC}+\neg P)$, which is Corollary 1.2 and is what the page's closing sentence calls independence relative to $\mathrm{Con}(\mathrm{ZFC})$.

Strongest attack

The strongest attack aimed at the corollary's model-theoretic paragraph: "let $N$ be a model of ZFC ... Adding $\omega_2$ random reals over it, in the sense of Theorem 1.1, yields a model of ZFC in which ...". Theorem 1.1 needs a filter $G$ generic over $M=L^N$. For an arbitrary model $N$ of ZFC, such a filter need not exist: $L^N$ may be uncountable, and a filter meeting all of its dense sets is not available in general. So the sentence, read as a deduction, uses a hypothesis that is not available, and the source's own proof (p. 8, "pass to a constructible universe ... and then add $\omega_2$ random reals") has the same informal shape. The attack fails to refute the page because its next sentence, "Formally, ...", carries the deduction syntactically, as re-derived in W2, with no generic filter and no model $N$: the two theorems give $\mathrm{ZFC}+\mathrm{CH}\vdash\mathbb B_{\omega_2}\Vdash P$, and the relative-consistency form of the forcing theorem gives $\mathrm{Con}(\mathrm{ZFC}+\mathrm{CH})\to\mathrm{Con}(\mathrm{ZFC}+P)$. The residue is a wording defect, filed as F1 (suggested).

Two further attacks failed outright. (a) The source's Theorem 1.1 says "the measure algebra adding $\kappa$ random reals" while the page says $\mathbb B(\kappa\times\omega)$; the source's proof of Theorem 5.1 fixes exactly that coordinate set, and the algebras on $\kappa$ and on $\kappa\times\omega$ coordinates are isomorphic, so the page's Theorem 1.1 is the source's (noted as F4). (b) Theorem 3.2 is written "for every family $\mathcal A$, $\mathrm{ZFC}\vdash\mathrm{Prof}(\mathcal A)\to\mathrm{Free}_\omega(\mathcal A)$" in both the source and the input page; a family is not a syntactic object, so the only reading is $\mathrm{ZFC}\vdash\forall\mathcal A\,[\ldots]$, which is the reading the page applies inside $M[G]$ and inside the forced theory; the source's display (1.1) omits the $\forall\mathcal A$ that its display (5.1) carries, and the page follows (5.1) through the Theorem 5.1 page. No quantifier changes.

Premises

• Theorem 3.2 page (local, same folder, Definitions and Statement read as of that time, proof not read). Interface: for every family $\mathcal A$, $\mathrm{ZFC}\vdash\mathrm{Prof}(\mathcal A)\to\mathrm{Free}_\omega(\mathcal A)$, with Definition 3.1 as restated above; checked clause by clause against the source's p. 3, display (3.5) and Definition 3.1, and found the same. Standing: not read (standing text is outside this review's read set); the page under review describes the reconstruction as author-recorded.
• Theorem 5.1 page (local, Definitions and Statement read, proof not read). Interface: $\mathrm{ZFC}+\mathrm{CH}\vdash\mathbb B_{\omega_2}\Vdash\forall\mathcal A\,[(\forall y\in\mathbb R\ \lambda^*(A_y)<1)\to\mathrm{Prof}(\mathcal A)]$ with $\kappa=\omega_2$, $\Theta=\kappa\times\omega$ and $\mathbb B=\mathbb B(\Theta)$; checked against the source's p. 6, display (5.1) and the opening of its proof, and found the same. Standing: as above.
• CH counterexample page (local, Statement read). Interface: under CH there is a family $(A_y)_{y\in\mathbb R}$ with every $A_y$ countable and bounded and no infinite independent set, hence $\mathrm{CH}\to\neg P$; checked against the source's p. 8 and found the same. Standing: as above.
• The forcing theorem for complete Boolean algebras (imported; the page cites T. Jech, Set Theory, third millennium edition, Chapter 14; not held in this review's read set, so checked against general knowledge only). Interface used: for a complete Boolean algebra $\mathbb B$ in $M$ and an $M$-generic filter $G$, $M[G]\models\mathrm{ZFC}$, and a sentence forced by an element of $G$ holds in $M[G]$; provably in ZFC, every theorem of ZFC is forced by every complete Boolean algebra, forced sentences are closed under logical consequence, and the measure algebra is nontrivial; hence if $\mathrm{ZFC}+\Sigma\vdash\mathbb B\Vdash\varphi$ then $\mathrm{Con}(\mathrm{ZFC}+\Sigma)\to\mathrm{Con}(\mathrm{ZFC}+\varphi)$. Standing: named on the page as imported.
• Gödel's theorem (imported; Jech, Chapter 13; not held). Interface used: ZFC proves each of its axioms and CH relativized to $L$, so $\mathrm{Con}(\mathrm{ZFC})\to\mathrm{Con}(\mathrm{ZFC}+\mathrm{CH})$, and $L^N\models\mathrm{ZFC}+\mathrm{CH}$ for every model $N$ of ZFC. Standing: named on the page as imported.
• The source PDF (held; read as recorded above). Interface: Theorem 1.1 and Corollary 1.2 as restated above, the proofs at pp. 7--8.
• Explicit assumptions. $\mathbb B_{\omega_2}$ means $\mathbb B(\omega_2\times\omega)$, the source's proof convention; "$\mathbb B\Vdash\varphi$" means that the top element forces $\varphi$; $M$ ranges over the models for which "generic over $M$" and $M[G]$ make sense, exactly as in the source's Theorem 1.1; in the corollary's model-theoretic gloss, the completeness theorem supplies a set model $N$ from $\mathrm{Con}(\mathrm{ZFC})$. No batch acceptance order.

Findings

F1. Severity: suggested. Location: "Adding $\omega_2$ random reals over it, in the sense of Theorem 1.1, yields a model of ZFC". Defect: for an arbitrary model $N$ of ZFC no filter generic over $L^N$ for the measure algebra need exist, so this sentence, read as a deduction, uses a hypothesis that is not available; the page's next sentence supplies the valid syntactic route, but the paragraph does not mark the first sentences as the source's informal proof rather than the page's argument. Witness: the source's proof of Corollary 1.2 (p. 8) has the same informal shape, and its Theorem 1.1 (p. 1) hypothesizes "$G$ generic over $M$" without supplying one. Proposed replacement for the paragraph "Consistency of $P$": "The source's proof (p. 8) passes to a constructible universe, which satisfies CH, and adds $\omega_2$ random reals over it; read as a construction this needs a filter generic over $L^N$, which need not exist for an arbitrary model $N$ of ZFC. The deduction here is syntactic. Theorems 5.1 and 3.2, with the forcing theorem (a theorem of ZFC is forced by every complete Boolean algebra), give $\mathrm{ZFC}+\mathrm{CH}\vdash\mathbb B_{\omega_2}\Vdash P$, since the forced statement without boundedness implies $P$; the relative consistency the forcing theorem yields turns this into $\mathrm{Con}(\mathrm{ZFC}+\mathrm{CH})\to\mathrm{Con}(\mathrm{ZFC}+P)$; and $\mathrm{Con}(\mathrm{ZFC})\to\mathrm{Con}(\mathrm{ZFC}+\mathrm{CH})$ by the constructible universe."

F2. Severity: note. Location: "Formally, Theorems 5.1 and 3.2 give $\mathrm{ZFC}+\mathrm{CH}\vdash\mathbb B_{\omega_2}\Vdash P$". Defect: the step from Theorem 3.2, a theorem of ZFC, to its being forced uses the imported forcing theorem (every theorem of ZFC is forced) and closure of forced sentences under consequence, which the sentence attributes to the two theorems alone; the import is named in the Standing paragraph, so nothing is unsupported. Witness: the source's display (1.1) (p. 1) and its proof of Theorem 1.1 (p. 7), which likewise leave the composition to the reader. Proposed replacement: the text under F1, which names the forcing theorem at this step.

F3. Severity: note. Location: "Boundedness is not assumed." Defect: the sentence sits inside the bold Theorem 1.1 statement, but it is a remark, not a clause of the source's theorem. Witness: the source's Theorem 1.1 (p. 1) ends at "satisfies $\mathrm{Free}_\omega(\mathcal A)$"; the remark comes from the abstract (p. 1, "in fact, boundedness is unnecessary"). Proposed replacement: end the statement at "satisfies $\mathrm{Free}_\omega(\mathcal A)$." and add after it: "The hypothesis is only $\lambda^*(A_y)<1$; boundedness, part of $P$, is not assumed (the source's abstract, p. 1)."

F4. Severity: note. Location: "the measure algebra $\mathbb B(\kappa\times\omega)$ adding $\kappa$ random reals". Defect: the source's theorem statement (p. 1) says only "the measure algebra adding $\kappa$ random reals"; the coordinate set $\kappa\times\omega$ is fixed in the proof of Theorem 5.1 (p. 6) and in Proposition 4.4 (p. 5). The statements agree, since the algebras on $\kappa$ and on $\kappa\times\omega$ coordinates are isomorphic, but the specification is the proof's, not the statement's. Proposed replacement: "for the measure algebra adding $\kappa$ random reals, $\mathbb B(\kappa\times\omega)$ in the coordinates of the source's proof of Theorem 5.1 (p. 6)".

Verdict

Source fidelity: faithful. The page's Theorem 1.1 and Corollary 1.2 match the source's statements on p. 1 clause by clause, with the coordinate specification of F4 taken from the source's proof; the definition of $P$ matches the abstract and the first question of Problem 501; the locators (statements p. 1, proof of Theorem 1.1 p. 7, proof of Corollary 1.2 and the counterexample in Section 6, p. 8, eight pages) are right; the Boundary paragraph's list of the forcing units matches F4--F6 on p. 8; the Standing paragraph claims author-recorded and nothing more.

The argument as reconstructed: sound. The proof of Theorem 1.1 composes Theorem 5.1 in $M$, the forcing theorem, the generic model theorem and Theorem 3.2 in $M[G]$ without gap (W1). The proof of Corollary 1.2 is carried by its syntactic sentence (W2) and the $\neg P$ half (W3); its model-theoretic gloss presupposes a generic filter that an arbitrary model need not have (F1), a wording defect that leaves the deduction intact.

Limitations: the input pages were read at their Statement sections only, so their proofs and standing are not vouched for here, and this verdict is conditional on their interfaces as restated; the two imports were checked against general knowledge, the cited textbook not being in the read set; the companion formalization was not examined; the two disclosed exposures were not used beyond the one confirmation noted. Refutation-failed. This focused review assigns no tier and changes no status.