---

Subject and independence

The reviewer worked in a fresh context from the commissioning assignment alone, took no part in writing the page, and had no contact with its author. The charge was refutation.

Frozen subject: path wiki/research/erdos_1219/lemma_1_1_reconstruction.md as it stood at 2026-09-28T05:03:27Z, read in full as of that time.

Artifact: the scan held under library/set_theory/shelah_1975_notes_partition_calculus/, twenty pages, printed pp. 1257--1276 = PDF pp. 1--20, stored with a rotation flag and without a text layer (a text extraction of PDF pp. 2--4 returns only the archive stamp). Page images were rendered from the scan at 150 dots per inch for PDF pp. 1--5; PDF pp. 2--4 (printed pp. 1258--1260) were read in full, and crops at 250 dots per inch were rendered and read for the lemma statement and Remark (p. 1258), the definition of $\operatorname{tf}$, the two type counts, the exceptional set $C_\alpha$ and the recursion (p. 1259), and the closing lines of the proof (p. 1260); two further crops at 400 dots per inch were read for the printed first type count and the printed application of the hypothesis on p. 1259, the witnesses of F4 and F6. Depth: every sentence and every displayed formula of the statement, the Remark and the proof of Lemma 1.1 was read clause by clause against the page; on p. 1260 the proof of Theorem 1.2 was read only far enough to see that it cites clause (1B) and defines its property $P_\alpha$ with $|B_{\alpha,0}|=|B_{\alpha,1}|=\mu(\alpha)$, which the page's "not used downstream" and "as it does in the application" sentences rest on. PDF pp. 1 and 5 were rendered but not read.

Allowed material read: the Statement section of wiki/research/erdos_1219/theorem_1_2_reconstruction.md as of that time; the statement section of wiki/problems/set_theory/E1219/_index.md as of that time (the part above its Current assessment heading), with its one status line masked; docs/verification.md sections "Whole-claim report" and "Audit checklist"; docs/evidence.md section "Source fidelity"; docs/math_authoring.md in full.

Exposures: two, both disclosed here. First, the library card _index.md of the Shelah source was read in full as of that time instead of its provenance paragraph only; the surplus was its read-status paragraph, a Contents bullet summarizing the structure of the lemma's proof, its Compiled scope, and a Bears-on paragraph that contains one sentence on the problem page's status. Second, the result page theorem_1_2.md on that card was read in full instead of its Statement section only; the surplus was its Proof pointer, Dependencies and Bears-on sections and its sentences saying that nothing there is independently reviewed. Neither surplus is a review of the page, and every derivation below was made from the page images and the page itself. No folder _index.md, no evidence folder, no other review, no workspace file and no web search was consulted.

Restatement

Conventions. All cardinals are von Neumann cardinals with the axiom of choice; a regular cardinal is an infinite cardinal equal to its own cofinality; cardinal sums, products and powers are meant throughout, the empty product being $1$.

Data. $\kappa$ is a regular cardinal. $\langle\lambda_i:i<\kappa\rangle$ are regular cardinals, strictly increasing in $i$. $\langle\mu(i):i<\kappa\rangle$ are cardinals and $\chi$ is a cardinal. $A_i$ ($i<\kappa$) are sets with $|A_i|=\lambda_i$, not assumed disjoint, and $A=\bigcup_{i<\kappa}A_i$. For each $i<\chi$, $F_i:A^{n_i}\to\chi$ with $1\le n_i<\omega$. Growth: for every $j<\kappa$, $\lambda^j=\prod_{i<j}\lambda_i^{\mu(i)}<\lambda_j$; and $2^{\chi+\kappa}<\lambda_0$, hence $2^{\chi+\kappa}<\lambda_j$ for all $j$. For every $\alpha<\kappa$ a property $P_\alpha$ of pairs $(\langle B_i:i\le\alpha\rangle,\langle a_i:\alpha<i<\kappa\rangle)$ with $B_i\subseteq A_i$ and $a_i\in A_i$ is given, subject to (H): for every $\alpha<\kappa$, every $\langle B_i:i<\alpha\rangle$ with $B_i\subseteq A_i$ and $|B_i|\le\mu(i)$, every $\langle a_i:\alpha<i<\kappa\rangle$ with $a_i\in A_i$, and every $C\subseteq A_\alpha$ with $|C|=\lambda_\alpha$, some $B_\alpha\subseteq C$ with $|B_\alpha|\le\mu(\alpha)$ satisfies $P_\alpha(\langle B_i:i\le\alpha\rangle,\langle a_i:\alpha<i<\kappa\rangle)$.

Conclusion. There exist $a^*_i\in A_i$ and $B_i\subseteq A_i$ with $|B_i|\le\mu(i)$ for all $i<\kappa$ such that:

(1A) for all $\alpha<\kappa$, $i<\chi$, $b,b'\in B_\alpha$ and every $\bar a\in(\bigcup_{j<\alpha}B_j)^{n_i-1}$: $F_i(b,\bar a)=F_i(b',\bar a)$;

(1B) for all $\alpha<\beta<\kappa$, $i<\chi$ with $n_i\ge2$, $b,b'\in B_\alpha$, $c,c'\in B_\beta$ and every $\bar a\in(\bigcup_{j<\alpha}B_j)^{n_i-2}$: $F_i(b,c,\bar a)=F_i(b',c',\bar a)=F_i(b',a^*_\beta,\bar a)$;

(2) for all $\alpha<\kappa$, $P_\alpha(\langle B_i:i\le\alpha\rangle,\langle a^*_i:\alpha<i<\kappa\rangle)$;

(3) if moreover every $n_i=3$, $2^{\chi+\kappa}<\operatorname{cf}\mu(i)$ for every $i$, and every $P_\alpha$ is preserved when each $B_i$ ($i\le\alpha$) is replaced by a subset of the same cardinality, then the $B_i$ can be taken so that also $F_i(a,b,c)=F_i(a',b',c')$ for all $a,a'\in B_\alpha$, $b,b'\in B_\beta$, $c,c'\in B_\gamma$, $\alpha<\beta<\gamma<\kappa$.

The page proves (1A), (1B) and (2) as stated. It proves (3) under one further reading, disclosed in its Standing paragraph and at the head of its clause (3) section: the sets produced by the recursion have $|B_\alpha|=\mu(\alpha)$, as they do when $P_\alpha$ forces that size.

Checklist

Canonical failure modes.

• "Almost all" upgraded to "all": absent. Every "for every admissible sequence" on the page is earned by choosing $a^*_\alpha$ outside the union $C_\alpha$ over all admissible sequences at once.
• Induction that presupposes termination: absent. The recursion runs over the well-ordered $\kappa$ and each stage uses only earlier stages and the pre-chosen $a^*_j$.
• Probabilistic or averaging heuristics as proofs: absent; the counting is exact cardinal arithmetic.
• Circular use of an equivalent statement: absent; the second thinning at stage $\alpha$ uses $a^*_j$ for $j>\alpha$, but these were fixed before the recursion, not by it.
• Exceptional sets dropped: absent; $C_\alpha$ is bounded explicitly and $a^*_\alpha$ is taken outside it.
• Finite verification cited as more: inapplicable; nothing is verified by instances.
• Convergence of a relaxed system standing in for the objects: inapplicable.

Named patterns.

• Model-class transport instead of entailment: inapplicable; no axiom system or certificate class is classified.
• Uniformity over an infinite family asserted from finitely many instances: passes; the bounds (A), (B), (C) are proved for every $\alpha$ with the dependence on $\alpha$ explicit ($\lambda^\alpha$, $\lambda_\alpha$), and the bound in the Claim is stated to be independent of $\bar B$.
• Extremal claims audited in the claim's own units: inapplicable; the page makes no sharpness, infimum or attainment claim.
• Consequence sentences are claim surfaces: checked one by one. "so that $2^{\chi+\kappa}<\lambda_j$ for every $j$" (monotone $\lambda$), "so the product is below $\lambda_\alpha$" (infinite $\lambda_\alpha$), "So $C_\alpha$ is a union of fewer than $\lambda_\alpha$ sets" (product of the two counts), "so some fiber has size $\lambda_\alpha$" (regularity), "So the value $F_i(a,b,c)$ depends only on $i$, $\beta$, $\gamma$ and $a$" (re-derived in W3) all hold. The frontmatter desc's "so that a value depends only on the blocks of its arguments" overstates (1B); see F3.
• Carry hypotheses actually used by a quantified argument: passes with one suggestion. Regularity of $\lambda_\alpha$, $\kappa\ge\aleph_0$, (H) and the heredity of $P_\alpha$ are stated where used; the reading $|B_\alpha|=\mu(\alpha)$ that clause (3) needs is stated in the proof section and in Standing but not at the statement of (3); see F1.
• A composition inherits its unproved premises: passes. The page consumes no local claim. Its only premises are ZFC cardinal arithmetic, listed under Premises. Clause (3) inherits the reading just named, and the page says so.
• Reproducibility notes are claims: inapplicable; the page contains no rerun line, count of passing checks or harness statement.
• Verifier quotations are claims: inapplicable; the page quotes no verifier and its Standing says it is not an independent review.
• Verdict words spelled in full: inapplicable to the page, which carries no verdict; this report writes its verdict words 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 instead of re-running: inapplicable; no gate.

Weakest steps

W1, the exceptional set and ($\ast$). Fix $\alpha<\kappa$ and write $\sigma=\sum_{i<\alpha}\mu(i)$. For an admissible $\bar B$ of length $\alpha$, $|E(\bar B)|\le\sigma$, so the patterns over $E(\bar B)$ number at most $\chi\cdot(\sigma+\aleph_0)\le\chi+\kappa+\sigma=:\theta_\alpha$, using $\kappa\ge\aleph_0$; a type is a function from the patterns into $\chi$, so the types realized in $A_\alpha$ number at most $\chi^{\theta_\alpha}\le2^{\theta_\alpha}=2^{\chi+\kappa}\cdot\prod_{i<\alpha}2^{\mu(i)}\le2^{\chi+\kappa}\cdot\lambda^\alpha=:\rho_\alpha$, by $2^{\sum\mu(i)}=\prod2^{\mu(i)}$ and $2\le\lambda_i$. Since $2^{\chi+\kappa}<\lambda_0\le\lambda_\alpha$ and $\lambda^\alpha<\lambda_\alpha$ with $\lambda_\alpha$ infinite, $\rho_\alpha<\lambda_\alpha$. The admissible sequences number at most $\prod_{i<\alpha}\lambda_i^{\mu(i)}=\lambda^\alpha$, because a nonempty subset of $A_i$ of size at most $\mu(i)\ge1$ is the range of a map $\mu(i)\to A_i$ and $\lambda_i^{\mu(i)}$ is infinite. Now $a\in C_\alpha$ if and only if $a$ lies in some $S(\bar B,a)$ of size below $\lambda_\alpha$, and any $a'$ in such a set has $S(\bar B,a')=S(\bar B,a)$, so $C_\alpha$ is exactly the union of the small fibers, at most $\lambda^\alpha\cdot\rho_\alpha<\lambda_\alpha$ sets each of size below $\lambda_\alpha$; regularity of $\lambda_\alpha$ gives $|C_\alpha|<\lambda_\alpha=|A_\alpha|$. Choosing $a^*_\alpha\in A_\alpha\setminus C_\alpha$ gives, for every admissible $\bar B$, $|S(\bar B,a^*_\alpha)|\ge\lambda_\alpha$, hence equality since $S(\bar B,a^*_\alpha)\subseteq A_\alpha$. Composition: ($\ast$) is what the first thinning needs at stage $\alpha$ whatever the earlier $B_i$ turned out to be, which is why all $a^*_j$ are fixed before the recursion; the source does the same (p. 1259, "Choose $a^*_i\in A_i-C_i$ for each $i<\kappa$" precedes "Now define inductively").

W2, the bridge through $a^*_\beta$ in (1B). At stage $\alpha$, $|D_\alpha|\le\sigma+\kappa$, so the type map over $D_\alpha$ takes fewer than $\lambda_\alpha$ values on $B^1_\alpha$ by the count of W1 with $\theta_\alpha$ unchanged; since $|B^1_\alpha|=\lambda_\alpha$ is regular, a fiber $B^2_\alpha$ of size $\lambda_\alpha$ exists, and (H) applied to $\langle B_i:i<\alpha\rangle$, $\langle a^*_i:\alpha<i<\kappa\rangle$ and $C=B^2_\alpha$ gives $B_\alpha\subseteq B^2_\alpha$ with $|B_\alpha|\le\mu(\alpha)$ and $P_\alpha$. Let $\alpha<\beta$, $b,b'\in B_\alpha$, $c,c'\in B_\beta$, $\bar a\in E_\alpha^{n_i-2}$. Since $B_\alpha\subseteq E_\beta$ and $E_\alpha\subseteq E_\beta$, $(b,x,\bar a)$ is a pattern over $E_\beta$, and $c$, $a^*_\beta$, $c'$ share their type over $E_\beta$ (T1 at $\beta$), so $F_i(b,c,\bar a)=F_i(b,a^*_\beta,\bar a)=F_i(b,c',\bar a)$, and the same with $b'$. Since $a^*_\beta\in D_\alpha$ and $E_\alpha\subseteq D_\alpha$, $(x,a^*_\beta,\bar a)$ is a pattern over $D_\alpha$, and $b$, $b'$ share their type over $D_\alpha$ (T2 at $\alpha$), so $F_i(b,a^*_\beta,\bar a)=F_i(b',a^*_\beta,\bar a)$. Chaining gives both equalities of (1B). The argument never compares $a^*_\beta$ with $c$ over $D_\alpha$, only over $E_\beta$, so no hypothesis on the type of $a^*_\beta$ over the other $a^*_j$ is needed. Composition: T1 at $\beta$ needs $E_\beta\supseteq B_\alpha$, available because stage $\beta$ follows stage $\alpha$; T2 at $\alpha$ needs $a^*_\beta\in D_\alpha$ for $\beta$ above $\alpha$, available because the $a^*_j$ precede the recursion. This matches p. 1259--1260 line by line.

W3, clause (3). Under the reading $|B_\alpha|=\mu(\alpha)$. For $\alpha<\beta<\gamma$, $a\in B_\alpha$, $b,b'\in B_\beta$, $c,c'\in B_\gamma$: $(a,b,x)$ is a pattern over $E_\gamma$ (as $a,b\in E_\gamma$), so T1 at $\gamma$ gives $F_i(a,b,c)=F_i(a,b,a^*_\gamma)$; $(a,x,a^*_\gamma)$ is a pattern over $D_\beta$ (as $a\in E_\beta$, $a^*_\gamma\in D_\beta$), so T2 at $\beta$ gives $F_i(a,b,a^*_\gamma)=F_i(a,b',a^*_\gamma)$; T1 at $\gamma$ again gives $F_i(a,b',a^*_\gamma)=F_i(a,b',c')$. So the value is a function $h_{i,\beta,\gamma}(a)$ of $a$ alone once $i,\beta,\gamma$ are fixed. The map $H_\alpha$ sending $a\in B_\alpha$ to $\langle h_{i,\beta,\gamma}(a):i<\chi,\alpha<\beta<\gamma<\kappa\rangle$ has at most $\chi^{\chi\cdot\kappa}$ values, which for $\chi\ge2$ is at most $2^{\chi\cdot\chi\cdot\kappa}=2^{\chi+\kappa}<\operatorname{cf}\mu(\alpha)$ because $\kappa$ is infinite. If every fiber of $H_\alpha$ had size below $\mu(\alpha)$, then $B_\alpha$ would be a union of fewer than $\operatorname{cf}\mu(\alpha)$ sets of size below $\mu(\alpha)$; the supremum of fewer than $\operatorname{cf}\mu(\alpha)$ cardinals below $\mu(\alpha)$ is below $\mu(\alpha)$, and its product with the number of fibers is below the infinite $\mu(\alpha)$, contradicting $|B_\alpha|=\mu(\alpha)$. So a fiber $B'_\alpha$ of size $\mu(\alpha)$ exists. Shrinking every $B_\alpha$ to $B'_\alpha$ preserves (1A) and (1B), which quantify universally over elements of the $B$'s and use only the unchanged $a^*_j$; preserves (2) by the heredity hypothesis, which allows all $B_i$ ($i\le\alpha$) to shrink at once to subsets of equal size; and yields (3) because for $a,a'\in B'_\alpha$, $b,b'\in B'_\beta$, $c,c'\in B'_\gamma$ the values $F_i(a,b,c)=h_{i,\beta,\gamma}(a)$ and $F_i(a',b',c')=h_{i,\beta,\gamma}(a')$ are computed in the original $B_\beta,B_\gamma$, which contain the new elements, and $H_\alpha(a)=H_\alpha(a')$. Composition: the shrinking must come after the whole recursion, because $h_{i,\beta,\gamma}$ depends on the fibers $t_\beta$, $t_\gamma$ chosen at later stages; nothing in (1A), (1B), (2) depends on which fibers are then chosen. The step that carries the reading is the size of the fiber: with $|B_\alpha|=\nu$ and $\operatorname{cf}\nu\le2^{\chi+\kappa}$ the fibers may all be small, and the page says as much ("The cofinality hypothesis has no force unless $|B_\alpha|=\mu(\alpha)$").

Strongest attack

The strongest attack aimed at clause (3). The page's Statement carries (3) as printed on p. 1258, whose only size constraint on the $B_i$ is the $|B_i|\le\mu(i)$ delivered by (H). The attack takes the instance in which $P_\alpha$ forces $|B_\alpha|=\aleph_0$ while $2^{\chi+\kappa}\ge\aleph_0<\operatorname{cf}\mu(\alpha)$: (H) can hold (every set of full size $\lambda_\alpha$ has a countable subset), the recursion produces countable $B_\alpha$, and the fiber step of W3 fails, since $H_\alpha$ may take up to $2^{\chi+\kappa}$ values on a countable set and every fiber may be finite. So the page's argument does not establish the printed clause in this instance. The attack does not refute the page: its Standing paragraph and the head of its clause (3) section state that (3) is proved under the reading $|B_\alpha|=\mu(\alpha)$, the reading is the one the source's cofinality hypothesis presupposes, and the clause is not consumed downstream (the proof of Theorem 1.2 on p. 1260 cites (1B) only). Whether the printed clause holds in the instance above by a different argument was not settled by this review; that is a question about the source, not about the page, and it yields the labeling finding F1 rather than a defect.

Three further attacks failed outright. Against (B): at $\alpha=0$, or when every $\mu(i)$ with $i<\alpha$ is $0$, the printed chain "$\le2^\chi\cdot\prod_{i<\alpha}2^{\mu(i)}\le\lambda^\alpha$" is false, but the page's (B) replaces it by $2^{\chi+\kappa}\cdot\lambda^\alpha$ and uses $2^{\chi+\kappa}<\lambda_0$, so the Claim survives, and the page records the printed slip. Against (1B): the bridge could have needed $a^*_\beta$ and $c$ to agree over $D_\alpha$, which nothing guarantees; but W2 shows it needs their agreement over $E_\beta$ only, which T1 gives. Against the degenerate parameters $\chi=0$, $\mu(i)=0$, $n_i=1$, $\alpha=0$ and overlapping $A_i$: (1B) is vacuous for $n_i=1$, (C) handles $\mu(i)=0$, $\alpha=0$ has the one empty admissible sequence, disjointness is never used, and only the intermediate chain of (A) misstates the case $\chi=0$ (F5), where the lemma has no content.

Premises

The page consumes no local claim and imports no theorem from outside Zermelo--Fraenkel set theory with choice. The cardinal-arithmetic facts it uses, each standard and each checked here, are:

• $2^{\sum_{i<\alpha}\mu(i)}=\prod_{i<\alpha}2^{\mu(i)}$ for cardinal sums and products over $\alpha$;
• for a regular $\lambda$, a union of fewer than $\lambda$ sets each of size below $\lambda$ has size below $\lambda$; used for $|C_\alpha|$ and for the fiber $B^2_\alpha$;
• for an infinite $\mu$, a union of fewer than $\operatorname{cf}\mu$ sets each of size below $\mu$ has size below $\mu$; used in clause (3);
• for an infinite $\lambda$, a product of two cardinals below $\lambda$ is below $\lambda$; used in (B) and the Claim;
• for infinite $\kappa$ and $\chi\ge1$, $\chi\cdot\kappa=\chi+\kappa$ and $|[\kappa]^2|=\kappa$; used in clause (3);
• the axiom of choice, used to pick $a^*_\alpha$, the fibers, and the surjections in (C).

The source is held as the scan described above and was read at the depth stated under Subject. Its interface as the page uses it: the statement of Lemma 1.1 with its hypothesis (H) and clauses (1A), (1B), (2), (3) (p. 1258); the Remark (p. 1258); the proof (pp. 1258--1260), including the definition of $\operatorname{tf}(\bar a,B)$, the two printed type counts, the count of sequences, the set $C_\alpha$, the choice of $a^*_i$, the recursion through $B^1_\alpha$, $B^2_\alpha$, $t_\alpha$ and $B_\alpha$, and the verification of (2), (1A), (1B); and the one-sentence instruction for (3). The cited input page, the reconstruction of Theorem 1.2, is a consumer of this page, not a premise; its Statement section was read only to confirm the page's "consumed by" sentence, and the source's own proof of Theorem 1.2 (p. 1260) confirms that the consumer uses (1B) only.

Explicit assumptions on the page: $\kappa$ infinite, which the convention "regular" already carries; $1\le n_i<\omega$, which (1A) presupposes; the $A_i$ need not be disjoint; and, for clause (3) only, the reading $|B_\alpha|=\mu(\alpha)$ discussed above.

Findings

F1. Severity: suggested. Location: the Statement, "(3) if every $F_i$ is three-place ...". Defect: the Statement presents (3) as printed, with no size requirement on the $B_i$ beyond $|B_i|\le\mu(i)$, while the page's proof establishes it only under the reading that the recursion's sets have $|B_\alpha|=\mu(\alpha)$; the reading is disclosed in Standing and at the head of the clause (3) section, but a reader of the Statement alone sees the printed clause claimed. Witness: p. 1258, clause (3), carries only "$2^{\chi+\kappa}<\operatorname{cf}[\mu(i)]$ for every $i$" and the heredity phrase; p. 1260 gives "replace the $B_\alpha$ by a subset of the same cardinality"; the fiber step in W3 needs $\operatorname{cf}|B_\alpha|>2^{\chi+\kappa}$. Proposed replacement: after the paragraph "The printed conclusion does not repeat ...", add "Clause (3) is proved below under the reading, stated in its section, that the recursion produces $|B_\alpha|=\mu(\alpha)$; the printed clause carries no such requirement."

F2. Severity: suggested. Location: the Statement, "The Remark after the statement (p. 1258) says that the lemma could be refined along the lines of the paper's [7], § 5, without application here." Defect: the Remark's second sentence is dropped without notice, although the Source paragraph claims the lemma "with its Remark". Witness: p. 1258, Remark: "We could refine the lemma along the lines of [7] § 5, but there is no application of it. We can assume that the range of $F_i$ is $2^\chi$." Proposed replacement: "... without application here, and that the range of the $F_i$ may be taken to be $2^\chi$ instead of $\chi$; the count (A) allows this, since a type is then one of at most $(2^\chi)^\theta=2^\theta$ functions."

F3. Severity: suggested. Location: the frontmatter desc, "so that a value depends only on the blocks of its arguments", and the title, "values depend only on block indices". Defect: (1B) fixes a value under exchange of the two leading arguments within $B_\alpha$ and $B_\beta$ only; the value still depends on the parameters $\bar a$ from the lower blocks, so a value of an $n_i$-place function with $n_i\ge3$ does not depend on block indices alone. The desc propagates into the generated index row, where it is the only text a reader sees. Witness: p. 1258, (1B), "$F_i(b,c,a_1,\ldots)=F_i(b',c',a_1,\ldots)$" with $a_1,\ldots$ fixed. Proposed replacement desc: "... so that a value with one argument from each of two blocks and the rest from lower blocks does not depend on which elements of the two blocks are used; for two-place functions it depends only on the two block indices."

F4. Severity: suggested. Location: Counting, "(A) For $B\subseteq A$ there are at most $2^{\chi+|B|+\aleph_0}$ types over $B$." Defect: the page silently corrects the printed first count, which lacks the $\aleph_0$ and is false for finite parameters; the page's Reading notes record two other printed slips but not this one, and the source-fidelity rule asks for an incorrect formula to be recorded explicitly. Witness: p. 1259, "Clearly $|\{\operatorname{tf}(\bar a,B):\bar a\in A\}|\le2^{|B|+\chi}$"; with $\chi=3$, three one-place functions into $3$ and $B=\emptyset$ there are $27$ types and $2^{|B|+\chi}=8$. Proposed replacement: add a reading note, "The printed first count '$\le2^{|B|+\chi}$' (p. 1259) omits the $\aleph_0$ that (A) carries and can fail for finite $\chi$ and $B$ (three one-place functions into $3$ over $B=\emptyset$ give $27$ types); the lemma is unaffected, since every bound it needs has $\kappa\ge\aleph_0$ in the exponent."

F5. Severity: note. Location: Counting (A), "so there are at most $\chi^\theta\le(2^\chi)^\theta=2^{\chi\cdot\theta}=2^\theta$ types". Defect: for $\chi=0$ the pattern set is empty and there is exactly one type, while $\chi^\theta=0$ and $2^{\chi\cdot\theta}=1\ne2^\theta$; the chain needs $\chi\ge1$. The final bound $2^\theta$ holds in every case and the lemma is empty at $\chi=0$. Witness: the page's own definition of a type as a function from the patterns into $\chi$. Proposed replacement: "so for $\chi\ge1$ there are at most $\chi^\theta\le(2^\chi)^\theta=2^{\chi\cdot\theta}=2^\theta$ types, and for $\chi=0$ exactly one."

F6. Severity: note. Location: Proof, "Choice of $B_\alpha$", and the Reading notes. Defect: the printed line the step transcribes has a misprint that the Reading notes do not record. Witness: p. 1259, "$P_\alpha(\langle B_i:i\le\alpha\rangle,\langle a^*_i:\alpha<i<a\rangle)$ holds", where the bound is $\kappa$. Proposed replacement: add to the Reading notes, "The printed application of the hypothesis (p. 1259) writes the second sequence as $\langle a^*_i:\alpha<i<a\rangle$; the bound is $\kappa$, as in the statement."

F7. Severity: note. Location: the Statement, "(1) ... every finite sequence $\bar a=a_1,a_2,\ldots$ of elements of $\bigcup_{j<\alpha}B_j$ of the length that fills the remaining places of $F_i$". Defect: the length differs between the two clauses, $n_i-1$ in (1A) and $n_i-2$ in (1B), and (1B) is vacuous for one-place $F_i$; the sentence can be read as naming one length for both. Witness: p. 1258, (1A) "$F_i(b,a_1,\ldots)$" and (1B) "$F_i(b,c,a_1,\ldots)$". Proposed replacement: "of length $n_i-1$ in (1A) and $n_i-2$ in (1B), so that (1B) says nothing for one-place $F_i$".

Verdict

Source fidelity: faithful with corrections. The hypotheses, quantifiers, clauses and locators of the reconstructed statement match p. 1258 of the held scan, the proof follows pp. 1258--1260 step by step, the supplied material (the pattern form of types, the counts (A)--(C), the regularity uses, the expansion of clause (3)) is marked as supplied, and the printed slips the page records are real. No correction is required; four are suggested (F1--F4) and three are notes (F5--F7).

The argument as reconstructed: sound. Clauses (1A), (1B) and (2) are established from the stated hypotheses without gap; clause (3) is established under the reading $|B_\alpha|=\mu(\alpha)$ that the page states, and is not consumed downstream.

Limitations. The review is noncomputational and rests on reading a scan by eye at 150, 250 and 400 dots per inch; every formula was cross-checked against the page's transcription and the internal logic of the proof. Whether the printed clause (3) holds without the page's reading was not settled. The use of the lemma by the reconstruction of Theorem 1.2 was not reviewed here beyond confirming, from the source's own proof, that it cites (1B) only.

This focused review assigns no tier and changes no status.