{"id":"04d5ae66-9924-4a96-8ad3-3133c8f31c2f","arxiv_id":"1908.05336","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Club guessing principles imply the existence of long towers of almost disjoint sets, yielding new non-saturation results for ideals extending the nonstationary ideal.","lead":"This paper studies when the collection of 'large' subsets of an uncountable infinite cardinal can be divided into many mutually almost disjoint pieces. The author proves that several set-theoretic club guessing principles force such divisions, in the form of long towers of sets, improving earlier results by Shelah, Gitik, and Rinot.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.9's recursive construction of g implicitly requires θ<τ; hypotheses allow θ≥τ, leaving the central theorem unproved in those cases.","rationale":"The reader's weakest-assumption diagnosis is correct and identifies exactly where the proof of Theorem 3.9 can break. I checked the surrounding hypotheses: κ=ν^+ and τ<Depth(κ^κ) do not relate the cardinality θ of the guessing sets to τ. In particular, when ν is singular and τ is a small regular cardinal like ω, θ can easily be a larger regular cardinal below ν, making the recursion g:θ^+→τ\\impossible. This is an internal gap in the proof, not a disagreement with external consensus. The theorem may still be true, and the conclusion for small τ might be derivable separately, but as written the claim is not established. I therefore leave the reader's conditional verdict unchanged.","tokens_in":34554,"tokens_out":8423,"duration_ms":95921,"concrete_test":"Re-run the proof of Theorem 3.9 in the case κ=ω_{ω+1}, ν=ω_ω, τ=ω, θ=ω1. At step ζ=ω, compute sup{g(ξ)*:ξ<ω}; since g is defined by iterating the operation β↦β*+1, this sup is already ω, so the recursive definition cannot produce an ordinal below τ. Then check whether any earlier part of the proof or Lemma 3.8 can force |A^i_δ|<τ on a J^+-set, e.g., via the regularity of τ or the bound τ<Depth(κ^κ). If no such forcing exists, Theorem 3.9 must either add the hypothesis θ<τ or supply a separate argument for τ≤ν.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing point is the definition of g:θ^+→τ\\(γ+1) in the proof of Theorem 3.9. The proof first chooses T∈J^+ and θ<ν such that |A^i_δ|=θ for all δ∈T; such a T exists by partitioning S_γγ* into the ν-many possible sizes, using κ-completeness of J and κ=ν^+. But ν is unrelated to τ. The recursive clause g(ζ)=(sup{g(ξ)*:ξ<ζ})+1 is only guaranteed to stay below τ when the sup of fewer than θ^+ ordinals below τ is below τ. Since τ is regular, this requires θ^+≤τ, i.e. θ<τ. Nothing in the hypotheses ♣cof,−_κ[J] or τ<Depth(κ^κ) gives θ<τ. Concretely, take κ=ω_{ω+1}, ν=ω_ω, τ=ω, and θ=ω1; at step ζ=ω, sup{g(ξ)*:ξ<ω} is already ω=τ, so g(ω) cannot be an ordinal in τ\\(γ+1). Lemma 3.8 does not bound |A^i_δ| in terms of τ, so there is no evident way to choose a smaller θ. The proof as written therefore depends on an unstated cardinality assumption that is essential for Theorem 3.9.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies club principles (diamond, club, and their variants) and their connection to the non-saturation of ideals extending the nonstationary ideal on a regular uncountable cardinal κ. The main new result is Theorem 3.9, which asserts that if κ = ν^+, J is a κ-complete ideal on κ satisfying ♣cof,−_κ[J], and τ is a regular cardinal less than Depth(κ^κ), then there exists either an ascending (J,I_κ)-tower of length τ or a descending (J,J)-tower of length τ. The paper also revisits several known results, proves variants, and extends results on generalized club guessing and on Gitik–Shelah style non-saturation.","tokens_in":34839,"tokens_out":7623,"duration_ms":72576,"significance":"If Theorem 3.9 and its corollaries (notably Proposition 3.14) are correct, they provide a uniform framework for deriving tower non-saturation of ideals from weak club principles, covering ideals that are not necessarily normal. The paper is also useful as a compendium of related principles and contains many small observations. However, the proof of the central theorem contains a gap (see Major Comment 1), so the significance of the paper is conditional on repairing that argument.","major_comments":[{"comment":"The recursive construction of g : θ^+ → τ \\(γ+1) implicitly requires θ < τ. Since τ is regular, the supremum of fewer than τ ordinals below τ is below τ; for ζ < θ^+ this requires θ^+ ≤ τ, i.e. θ < τ. The hypotheses only give θ < ν, with no relation between ν and τ (τ < Depth(κ^κ) may be larger or smaller than ν). Concretely, take κ = ω_{ω+1}, ν = ω_ω, τ = ω, and θ = ω_1; at step ζ = ω the supremum sup{g(ξ)^* : ξ < ω} is already ω = τ, so g(ω) cannot be defined. The proof therefore depends on an unstated cardinality assumption that is load-bearing for Theorem 3.9, and Proposition 3.14 inherits the gap.","section":"§3.1, Theorem 3.9 proof"},{"comment":"The same issue appears in the proof of Theorem 4.7: the function g : θ^+ → τ \\(γ+1) is defined with θ < σ, but nothing ensures θ < τ. Since τ is regular and θ^+ may exceed τ, the recursion can break before reaching θ^+. Thus Theorem 4.7 is not proved as stated.","section":"§4, Theorem 4.7 proof"},{"comment":"The proof of Proposition 3.15(i) begins with the sentence 'We modify the proof (which we do not understand) of Theorem 3.5 in [36]'. This is an explicit admission that the author cannot vouch for the correctness of the argument being modified. Since Proposition 3.15 is used in Proposition 3.17 and in the proof of Fact 2.18, this is a missing-support concern in a chain leading to a stated proof. The author should either supply a fully understood proof or explicitly mark the result as dependent on [36].","section":"§3.3, Proposition 3.15"}],"minor_comments":[{"comment":"The abstract contains the typo 'n onsaturation'; it should read 'non-saturation'.","section":"Abstract"},{"comment":"The definition of ♣_κ[J] contains an extra closing parenthesis: '♣κ[J])' should be '♣κ[J]'.","section":"Definition 2.21"},{"comment":"In the proof, 'fix i < j < κ' should be 'fix i < j < ρ', since the almost disjoint family is indexed by ρ.","section":"Observation 2.27 proof"},{"comment":"The text 'fg(ξ(j)' appears to be a typo for 'f_{g(ξ)}(j)', and the following inequality has mismatched parentheses, making the claim hard to read.","section":"Theorem 4.7, Claim 4"},{"comment":"Several proofs are only sketched or left to the reader, including the descending case in Theorem 2.31, part (i) of Theorem 4.7, and part (ii) of Observation 4.8; at least brief indications would improve accessibility.","section":"Various proofs"}],"recommendation":"major_revision","confidential_remarks":"The paper is rich in combinatorial content, but the central Theorem 3.9 has a genuine gap: the recursive definition of g requires θ < τ, a condition not present in the hypotheses. The admission in Proposition 3.15 that a source proof is not understood is also unusual and should be resolved before publication. I recommend major revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, here's my read on Matet's 'Towers and clubs'. The paper is exactly what it says: a systematic attempt to turn known non-saturation results for the nonstationary ideal into tower existence results for arbitrary κ-complete ideals extending NS, including non-normal ones. The main new statements—Theorem 3.9, Proposition 3.14, Observation 4.8, Theorem 5.12—are genuine generalizations of results by Rinot, Gitik-Shelah, and Džamonja, and the tower formulations are a useful unifying device. The proofs are mostly detailed adaptations of known arguments rather than a new framework, but that's fine; the paper is a solid piece of combinatorial set theory.\n\nThe soft spots, though, are real. The stress-test on Theorem 3.9 holds up. In the proof, after choosing θ<ν with |A^i_δ|=θ for δ in T, the recursion defines g:θ^+→τ\\(γ+1). For this to make sense, every sup of fewer than θ^+ ordinals below τ must be below τ. Since τ is regular, that requires θ^+≤τ, i.e. θ<τ. Nothing in the hypotheses gives this: θ is bounded by ν, and τ is only bounded by Depth(κ^κ). The concrete counterexample with κ=ω_{ω+1}, τ=ω, θ=ω1 shows the construction fails at step ω. So Theorem 3.9, the central new result, is unproved as written in the case θ≥τ. This is not a minor typo; the recursion is the heart of the proof.\n\nSecond, Proposition 3.15 explicitly says the proof modifies a proof 'which we do not understand.' That is an honest admission, but it also means that step is not verified. The author may be right, but a referee should ask for a self-contained argument there.\n\nThe rest of the paper is in much better shape. The facts cited from the author's own prior work are used as background, not in a circular way. The generalizations in Section 4 and the Gitik-Shelah material in Section 5 look plausible and are carefully argued. I did not spot other load-bearing gaps.\n\nBottom line: this paper deserves a serious referee, not a desk reject, but it should not be accepted until Theorem 3.9 is either repaired (by adding a condition like θ<τ or by a different construction) and Proposition 3.15 is clarified. For a reading group, I'd say maybe—the ideas are worth discussing, but you'd want to flag the gap. I wouldn't cite it in the next year until the central theorem is settled.","headline":"Matet's tower theorems are a useful generalization but Theorem 3.9 has a genuine cardinality gap that needs fixing before the centerpiece holds.","tokens_in":35355,"tokens_out":2516,"would_cite":false,"duration_ms":24629,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["03E05"],"pacs":[],"model":"deepseek-v4-flash","headline":"Club guessing produces long ascending or descending towers in every ideal extending the nonstationary ideal over a successor cardinal.","keywords":["club principle","saturated ideal","tower","diamond principle","club guessing","nonstationary ideal","depth of the function space","set theory"],"falsifier":"Build a model with $\\kappa=\\nu^+$, a $\\kappa$-complete ideal $J$ extending the nonstationary ideal, $\\clubsuit^{\\mathrm{cof},-}_{\\kappa}[J]$, and a regular $\\tau<\\mathrm{Depth}(\\kappa^\\kappa)$ for which $J$ is $\\tau$-saturated; then neither tower of length $\\tau$ can exist and Theorem 3.9 fails. Concretely, inspect the proof's choice of $T$ and $\\theta$: if every $J^+$-set carries guesses of size at least $\\tau$, the recursion $g:\\theta^+\\to\\tau$ cannot be defined, so exhibiting such an ideal with $\\mathrm{Depth}(\\kappa^\\kappa)>\\tau$ would pinpoint the gap.","tokens_in":34353,"feed_emoji":"♣️","tokens_out":17715,"duration_ms":146112,"temperature":0.7,"pith_summary":"This paper tries to show that weak club-guessing principles force every $\\kappa$-complete ideal extending the nonstationary ideal on a regular uncountable cardinal $\\kappa$ to be non-saturated, and to witness this non-saturation by explicit towers of positive sets. The central theorem states that when $\\kappa=\\nu^+$ and the cofinal club-guessing principle $\\clubsuit^{\\mathrm{cof},-}_{\\kappa}[J]$ holds, every regular $\\tau<\\mathrm{Depth}(\\kappa^\\kappa)$ admits either an ascending $(J,I_\\kappa)$-tower or a descending $(J,J)$-tower of length $\\tau$. Tower existence is a direct certificate that the ideal cannot be $\\tau$-saturated, and the result covers ideals that are not normal. The same mechanism is extended to guessing generalized clubs and to full stationary reflection, and several known diamond and club results are reproved in passing.","feed_headline":"Club guessing forces long towers in ideals","feed_subtitle":"For $\\kappa=\\nu^+$, any such ideal gains a tower at every regular length below $\\mathrm{Depth}(\\kappa^\\kappa)$.","key_machinery":"The carrier of the argument is the cofinal club-guessing principle $\\clubsuit^{\\mathrm{cof},-}_{\\kappa}[J]$, which supplies small sets $B^i_\\delta\\subseteq\\delta$ for $i<\\delta$ such that for every $W\\in[\\kappa]^\\kappa$, the collection of $\\delta$ where some $B^i_\\delta$ has cofinal intersection with $W$ is $J^+$. Lemma 3.8 converts these into $\\nu$-sized pieces $A^i_\\delta$ of $\\kappa\\times\\kappa$ that capture $f\\upharpoonright Z$ for cofinal $Z\\subseteq\\delta$. Towers are sequences $\\langle A_\\alpha:\\alpha<\\tau\\rangle$ of $J^+$ sets whose successive differences lie in a prescribed family $Y$; an ascending $(J,I_\\kappa)$-tower means each new part is almost disjoint from the union of earlier parts, and a descending $(J,J)$-tower means the leftover part stays $J^+$. The proof tracks the guessed sets $S_\\alpha$ and the crossing sets $S_{\\alpha\\beta}$, and uses $\\mathrm{Depth}(\\kappa^\\kappa)$, the least length of an unavoidable increasing chain in $(\\kappa^\\kappa,<^{*})$, as the ceiling for $\\tau$.","core_discovery":"The paper's central claim is a uniform transfer: cofinal club guessing converts any strictly $<^{*}$-increasing chain of functions in $\\kappa^\\kappa$ of length $\\tau$ into either an ascending $(J,I_\\kappa)$-tower or a descending $(J,J)$-tower of length $\\tau$, for any $\\kappa$-complete ideal $J$ extending the nonstationary ideal. Theorem 3.9 states this for $\\kappa=\\nu^+$ under $\\clubsuit^{\\mathrm{cof},-}_{\\kappa}[J]$ and every regular $\\tau<\\mathrm{Depth}(\\kappa^\\kappa)$. The proof first turns club guesses into sets $A^i_\\delta\\subseteq\\kappa\\times\\kappa$ that capture restrictions of functions to cofinal subsets of $\\delta$, then defines $J^+$ sets $S_\\alpha$ where a function is guessed and $S_{\\alpha\\beta}$ where $\\kappa$-truncations of two functions cross. Assuming neither tower exists, a diagonal recursion on those sets produces a contradiction. The paper derives corollaries for ideals extending $NS_\\kappa|E^\\kappa_\\theta$ under cardinal arithmetic assumptions, for ideals with generalized-club guessing, and for stationary reflection.","pith_inferences":["If the missing assumption that the guessed sets have size below $\\tau$ can be proved from the stated hypotheses, then club guessing and $\\mathrm{Depth}(\\kappa^\\kappa)$ alone control non-saturation; if not, the theorem likely needs an explicit size bound on the guesses.","The same construction might yield ascending $(J,J)$-towers or descending $(J,I_\\kappa)$-towers by symmetry, which would answer the paper's Questions 6.1 and 6.2 affirmatively under club guessing.","The bound $\\tau<\\mathrm{Depth}(\\kappa^\\kappa)$ may be improvable to $\\tau<\\mathrm{Depth}(C_\\kappa)$, since Fact 2.47 shows the two depths differ by at most one; testing this would show whether the tower length is really tied to the function space.","A natural test is whether $\\clubsuit^{\\mathrm{cof},-}_{\\kappa}[J]$ alone yields a tower of length $\\nu^+$ even when $\\mathrm{Depth}(\\kappa^\\kappa)=\\nu^+$, which would isolate whether the depth parameter is essential."],"forward_implications":["Under $\\kappa=\\nu^+$ and $\\clubsuit^{\\mathrm{cof},-}_{\\kappa}[J]$, every regular $\\tau<\\mathrm{Depth}(\\kappa^\\kappa)$ is witnessed by a tower, so $J$ is not $I_\\kappa$-$\\tau$-saturated (by Observation 2.29).","For ideals extending $NS_\\kappa|E^\\kappa_\\theta$ with $\\mathrm{cf}(\\nu)\\neq\\theta$, Fact 3.12 supplies the club guessing and Proposition 3.14 yields the same tower dichotomy without extra hypotheses.","The non-normal case is covered, so ideals that are not subnormal still admit long towers whenever the relevant club principle holds.","Guessing generalized clubs (Theorem 4.7) and full stationary reflection (Theorem 5.12) each produce the same ascending-or-descending tower conclusion.","The slow-train propositions show that $\\clubsuit^{\\mathrm{cof},-}_{\\kappa}[J]$ together with $2^{<\\kappa}=\\kappa$ upgrades to diamond $\\lozenge_\\kappa[J]$, recovering and extending an earlier diamond theorem."],"supporting_citations":[{"why":"Supplies the tower construction that Theorem 3.9 modifies; its Lemma 1.5 is the template for Lemma 3.8.","marker":"[37]"},{"why":"Gives the diamond-from-guessing result (Fact 2.18) that the slow-train section reproves and extends.","marker":"[47]"},{"why":"Introduces the weak club principle lineage and the saturation arguments on which the cofinal club principle is built.","marker":"[7]"},{"why":"Provides the depth parameters $\\mathrm{Depth}(\\kappa^\\kappa)$ and $\\mathrm{Depth}(C_\\kappa)$ and the inequalities used in the tower-length ceiling.","marker":"[46]"},{"why":"Supplies the non-saturation theorem and full-reflection argument that Theorem 5.12 recasts in tower form.","marker":"[17]"},{"why":"Provides the generalized-club guessing method that Theorem 4.7 extends from successors of regulars to weakly inaccessible cardinals.","marker":"[39]"}],"fun_headline_variants":["Club guessing turns chains into long towers","Guessing clubs forces towers of any regular length","From club guessing to towers in ideals","Lengthy chains become ideal towers via club guessing","Club guessing yields uniform tower transfer"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of Theorem 3.9 secretly assumes that the guessed sets have size smaller than the tower length $\\tau$; the stated hypotheses only bound them by $\\nu$, and no argument is given that this smaller bound holds.","fun_headline_variants_meta":{"raw":{"variants":["Club guessing turns chains into long towers","Guessing clubs forces towers of any regular length","From club guessing to towers in ideals","Lengthy chains become ideal towers via club guessing","Club guessing yields uniform tower transfer"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000148,"raw_usage":{"total_tokens":1141,"prompt_tokens":846,"completion_tokens":295,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":462,"completion_tokens_details":{"reasoning_tokens":231}},"tokens_in":462,"tokens_out":295,"duration_ms":3264,"temperature":1.0,"reasoning_tokens":231,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:18:02.154748+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Build a model with $\\kappa=\\nu^+$, a $\\kappa$-complete ideal $J$ extending the nonstationary ideal, $\\clubsuit^{\\mathrm{cof},-}_{\\kappa}[J]$, and a regular $\\tau<\\mathrm{Depth}(\\kappa^\\kappa)$ for which $J$ is $\\tau$-saturated; then neither tower of length $\\tau$ can exist and Theorem 3.9 fails. Concretely, inspect the proof's choice of $T$ and $\\theta$: if every $J^+$-set carries guesses of size at least $\\tau$, the recursion $g:\\theta^+\\to\\tau$ cannot be defined, so exhibiting such an ideal with $\\mathrm{Depth}(\\kappa^\\kappa)>\\tau$ would pinpoint the gap.","supporting_citations":[{"cited_title":"RINOT, A relative of the approachability ideal, diamond and non- saturation, Journal of Symbolic Logic 75 (2010), 1035-1065","cited_arxiv_id":null,"evidence_quote":"Supplies the tower construction that Theorem 3.9 modifies; its Lemma 1.5 is the template for Lemma 3.8."},{"cited_title":"SHELAH, Diamonds, Proceedings of the American Mathematical Soci- ety 138 (2010), 2151-2161","cited_arxiv_id":null,"evidence_quote":"Gives the diamond-from-guessing result (Fact 2.18) that the slow-train section reproves and extends."},{"cited_title":"D ˇZAMONJA and S","cited_arxiv_id":null,"evidence_quote":"Introduces the weak club principle lineage and the saturation arguments on which the cofinal club principle is built."},{"cited_title":"SHELAH, Applications of PCF theory , Journal of Symbolic Logic 65 (2000), 1624-1674","cited_arxiv_id":null,"evidence_quote":"Provides the depth parameters $\\mathrm{Depth}(\\kappa^\\kappa)$ and $\\mathrm{Depth}(C_\\kappa)$ and the inequalities used in the tower-length ceiling."},{"cited_title":"GITIK and S","cited_arxiv_id":null,"evidence_quote":"Supplies the non-saturation theorem and full-reflection argument that Theorem 5.12 recasts in tower form."},{"cited_title":"RINOT, On guessing generalized clubs at the successors of regulars , Annals of Pure and Applied Logic 162 (2011), 566-577","cited_arxiv_id":null,"evidence_quote":"Provides the generalized-club guessing method that Theorem 4.7 extends from successors of regulars to weakly inaccessible cardinals."}],"review_version":1}