{"id":"76b71520-b452-4533-af38-11673876d682","arxiv_id":"2608.13523","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper asserts a ZFC proof that every comeager hereditary family of compact sets in a Polish space contains a dense hereditary G_delta subfamily, but a key set membership claim is false as stated.","lead":"This paper claims to solve an open question from Matheron and Zeleny: every comeager hereditary family of compact subsets of a Polish space contains a dense hereditary G_delta subfamily. The result is new and significant, but the proof contains a false lemma about the sets used to cover the complement.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 2.4 is false as stated: C = U \\ V need not be closed, so C ∉ H. However the proof's density argument establishes closure(C) ∈ H, so the flaw is repairable by concluding C ∈ H_ext instead.","rationale":"I agree with the reader that Proposition 2.4's statement 'C ∈ H' is false and that the provided counterexample is correct. The reader is also right that this is a load-bearing step in the proof as written, because step (3) invokes the membership of C_{r,s} in H. However, the reader's REJECT verdict overstates the impact. The proof of Proposition 2.4, when the missing closure bars are restored, actually proves closure(C) ∈ H: the contradiction shows that closure(C) ∩ P(n) has empty interior in P(n) for every n. Since C ⊆ closure(C), this yields C ∈ H_ext, and H_ext is closed under countable unions and subsets. The theorem's covering argument only needs B ⊆ ∪ C_{r,s} (which is proved independently) and each C_{r,s} ∈ H_ext; then B ∈ H_ext and Theorem 1.2 gives the result. Thus the error is a genuine but local gap, best resolved by a conditional acceptance requiring the author to replace the false conclusion of Proposition 2.4 with the correct H_ext conclusion and adjust step (3) accordingly. I found no other load-bearing flaw: Lemma 2.1, Lemma 2.2, Lemma 2.3, the density claim (2), and the Baire-category argument in Theorem 1.1 are coherent as written.","tokens_in":3306,"tokens_out":20556,"duration_ms":197131,"concrete_test":"Verify the counterexample: for X = [0,1], E = {0} ∪ {1/n}, O = {K : K ⊂ (0,1)}, F = {K : K ∩ E ≠ ∅}, check that {1/n} ∈ C for all n, {1/n} → {0} in the Vietoris topology, and {0} ∉ C; this confirms C is not closed. Then verify the repair: compute closure(C) = F, check F ∩ P(n) is nowhere dense in P(n) for each n, so closure(C) ∈ H and hence C ∈ H_ext. Finally, rerun Theorem 1.1 with Proposition 2.4 weakened to 'C_{r,s} ∈ H_ext' and confirm B ⊆ ∪ C_{r,s} still yields B ∈ H_ext via the σ-ideal property.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The load-bearing concern is Proposition 2.4's claim that C = U \\ V belongs to H. Since H consists only of closed subsets of K(X), C must be closed. It need not be: take X = [0,1], E = {0} ∪ {1/n : n ≥ 1}, O = {K : K ⊂ (0,1)}, F = {K : K ∩ E ≠ ∅}. Then C = {K ⊂ (0,1) : K ∩ E ≠ ∅}, and the sets {1/n} lie in C while their Vietoris limit {0} is not in C (no member of O contains 0). Hence C ∉ H and step (3) of Theorem 1.1 fails as written. This is a real gap, but not a fatal one. The proof of Proposition 2.4, with the intended closure bars, shows that closure(C) ∩ P(n) is nowhere dense in P(n) for every n, so closure(C) ∈ H. Therefore C ∈ H_ext because C ⊆ closure(C). Replacing the conclusion 'C ∈ H' by 'C ∈ H_ext' in Proposition 2.4 and in step (3) preserves the proof: B ⊆ ∪ C_{r,s} still holds, and H_ext is a σ-ideal, so B ∈ H_ext. The central construction is sound after this correction.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims an affirmative solution in ZFC to a problem of Matheron and Zelený: every comeager hereditary family I of compact subsets of a Polish space X contains a dense hereditary G_delta subfamily of the Vietoris hyperspace K(X). The proof uses the Matheron–Zelený characterization (Theorem 1.2) that a family is “big” exactly when its complement belongs to the sigma-ideal H_ext generated by a certain family H of closed nowhere-dense-in-P(n) sets. The author defines U_s = downarrow O_s and V_{r,s} = downarrow(O_s \\ F_r) for a countable base (O_s) and closed nowhere dense sets F_r covering the meager complement B of I, and Proposition 2.4 asserts each C_{r,s} = U_s \\ V_{r,s} lies in H. A Baire-category argument then places B in H_ext, so Theorem 1.2 yields the desired dense hereditary G_delta subset.","tokens_in":3585,"tokens_out":5702,"duration_ms":56358,"significance":"If the proof is correct, the result is substantial: it removes the coanalytic definability assumption and the additional set-theoretic hypothesis of Matheron and Zelený, answering their Problem 2.1 in full generality in ZFC. The strategy is attractive and elementary, relying only on the published Matheron–Zelený criterion, basic hyperspace topology, and a Baire-category covering argument. The paper is short, clearly written, and the key density argument is correct in its main line. These strengths make the flaw discussed below particularly worth repairing rather than grounds for outright rejection.","major_comments":[{"comment":"The proposition asserts that C = U \\ V belongs to H, but H by definition contains only closed subsets of K(X). The proof shows that for every n, C ∩ P(n) is nowhere dense in P(n), but it does not show that C is closed, and in general C need not be closed. For a concrete counterexample, take X = [0,1], E = {0} ∪ {1/n : n ≥ 1}, O = {K ∈ K(X) : K ⊂ (0,1)}, and F = {K ∈ K(X) : K ∩ E ≠ ∅}. Then C = {K ⊂ (0,1) : K ∩ E ≠ ∅}, and the sets {1/n} lie in C while their Vietoris limit {0} does not, since no member of O contains 0. Therefore C ∉ H, and step (3) of the proof of Theorem 1.1 is not justified as written.","section":"Proposition 2.4, page 4"},{"comment":"The gap in Proposition 2.4 is load-bearing for the main theorem because the conclusion B ∈ H_ext relies on each C_{r,s} being in H. The good news is that the defect is repairable within the manuscript's scope: the density argument actually establishes that C ∩ P(n) is nowhere dense in P(n) for every n, and because P(n) is closed, the closure of C also has this property, so closure(C) ∈ H. Since C ⊆ closure(C), this gives C ∈ H_ext. Replacing the conclusion “C ∈ H” by “C ∈ H_ext” in Proposition 2.4 and in step (3) preserves the proof, because H_ext is a sigma-ideal and the inclusion B ⊆ ∪_{r,s} C_{r,s} is unchanged. The authors should make this correction explicitly.","section":"Proof of Theorem 1.1, step (3), page 5"}],"minor_comments":[{"comment":"The title and abstract contain spacing artifacts (“HEREDIT AR Y”, “F AMILIES”, “COMP ACT”) that should be corrected for publication.","section":"Title and abstract"},{"comment":"After the density argument, the text says “therefore C ∈ H”; this should instead say that the closure of C belongs to H, with the consequent adjustment to H_ext.","section":"Proposition 2.4, final sentence"},{"comment":"The sentence “It follows from (4) and (3) that B ⊆ ∪_{r,s} C_{r,s}” merely repeats (4); the intended meaning is that combined with C_{r,s} ∈ H (or H_ext), this inclusion gives B ∈ H_ext. Rewording would avoid confusion.","section":"Proof of Theorem 1.1, after equation (4)"}],"recommendation":"major_revision","confidential_remarks":"The paper's central idea is sound and the flaw in Proposition 2.4 is local and repairable. I recommend major revision rather than rejection because the proposed fix is straightforward and does not change the overall structure of the argument. The authors should verify that the corrected statement of Proposition 2.4 with H_ext is used consistently, and should double-check the closure argument for C ∩ P(n) in the counterexample-free case."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Ula, here's my honest read.\n\nThe paper looks like it solves the Matheron-Zelený problem in ZFC: every comeager hereditary family of compact subsets of a Polish space contains a dense hereditary G_delta subfamily. That is a genuine 20-year-old question, previously known only under coanalytic definability or the extra hypothesis omega_1^{L[x]} < omega_1. The argument is short and elegant: reduce to the Matheron-Zelený ideal-theoretic characterization, then cover the meager complement by sets C_{r,s} built from a countable basis and nowhere dense closed sets. Lemmas 2.1 through 2.3 are correct. Lemma 2.1 (hereditary closure of an open set is open) is nicely done. Lemma 2.3 is the standard approximation trick. The density argument in Proposition 2.4 is basically right.\n\nThe soft spot is real but repairable. Proposition 2.4 asserts C = U \\ V is in H, but H is defined as a family of closed sets, and C need not be closed. The reader's counterexample is valid: with X = [0,1], E = {0} ∪ {1/n}, O = {K : K ⊂ (0,1)}, and F = {K : K ∩ E ≠ ∅}, the sets {1/n} are in C while their limit {0} is not. So the proposition as stated is false. But the stress-test is right that this is not fatal. The proof really shows C ∩ P(n) is nowhere dense for each n. A small modification, applying the argument to the closure of C, gives closure(C) ∈ H, and hence C ∈ H_ext directly (since C ⊆ closure(C)). Step (3) of the main proof can then say C_{r,s} ∈ H_ext instead of H. Because H_ext is a σ-ideal, the final B ∈ H_ext still follows. So the main theorem survives intact after a modest revision.\n\nI have no other serious concerns. The citation pattern is clean; it relies on the published Matheron-Zelený characterization and does not assume its target. No parameter fitting, no circularity. The writing is clear, maybe a bit terse, but fine.\n\nWho is this for? People working in descriptive set theory of hyperspaces, and anyone who likes clean Baire-category arguments. It deserves a serious referee: the result is important and the proof is repairable. My recommendation: send it to review, ask for the fix to Proposition 2.4 (conclude C ∈ H_ext, or restate H to include arbitrary subsets of H sets). If the author does that, this is an accept.","headline":"A genuinely new ZFC solution to Matheron-Zelený's problem, with one real but easily repairable gap in Proposition 2.4.","tokens_in":4128,"tokens_out":4346,"would_cite":true,"duration_ms":41530,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["03E15","54B20","54E52"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that every comeager hereditary family of compact subsets of a Polish space contains a dense hereditary G-delta subfamily, answering Matheron and Zelený's problem in ZFC.","keywords":["Vietoris hyperspace","hereditary family","comeager","dense hereditary G_delta","Baire category","Polish space","Matheron-Zelený problem","H_ext ideal"],"falsifier":"Check Proposition 2.4 with $X=[0,1]$, $E=\\{0\\}\\cup\\{1/n:n\\ge1\\}$, $F=\\{K\\in\\mathcal{K}(X):K\\cap E\\neq\\emptyset\\}$, and $O=\\{K:K\\subseteq(0,1)\\}$. Then $C=\\downarrow O\\setminus\\downarrow(O\\setminus F)=\\{K\\subseteq(0,1):K\\cap E\\neq\\emptyset\\}$, and the singleton compact sets $\\{1/n\\}$ lie in $C$ but converge in the Vietoris topology to $\\{0\\}\\notin C$; since $H$ contains only closed subsets of $\\mathcal{K}(X)$, this refutes Proposition 2.4 and thereby the proof's step (3).","tokens_in":3095,"feed_emoji":"♾️","tokens_out":18717,"duration_ms":159231,"temperature":0.7,"pith_summary":"The paper asserts an affirmative answer to Matheron and Zelený's Problem 2.1: for every Polish space $X$, every comeager hereditary family $\\mathcal{I}\\subseteq\\mathcal{K}(X)$ contains a dense hereditary $G_\\delta$ subfamily. This removes the definability restriction (the previous result covered coanalytic families) and the additional set-theoretic hypothesis that earlier work needed. Because a comeager family always contains a dense $G_\\delta$, the real content is that the dense $G_\\delta$ can be chosen hereditary. The proof works by placing the complement of such a family into the $\\sigma$-ideal $H_{\\mathrm{ext}}$ and then using the Matheron–Zelený characterization of bigness.","feed_headline":"Comeager hereditary families of compact sets are big in ZFC","feed_subtitle":"Matheron and Zelený's problem is settled without definability restrictions or extra set-theoretic hypotheses.","key_machinery":"The load-bearing object is the $\\sigma$-ideal $H_{\\mathrm{ext}}$: a set belongs to it if it can be covered by countably many closed sets $F\\subseteq\\mathcal{K}(X)$ whose intersections with the finite-cardinality layers $P(n)=\\{K\\in\\mathcal{K}(X):|K|\\le n\\}$ are nowhere dense in $P(n)$. The Matheron–Zelený characterization (Theorem 1.2) turns 'contains a dense hereditary $G_\\delta$' into the statement that the complement lies in $H_{\\mathrm{ext}}$. The proof's engine is Proposition 2.4, which asserts that differences of the form $\\downarrow O\\setminus\\downarrow(O\\setminus F)$ lie in $H$ for open $O$ and closed nowhere dense $F$; this is what lets the meager, upward-closed complement $B$ be written as a countable union of $H$-sets. The Baire category theorem then does the covering: for each $K\\in B$, the upper cone $\\uparrow K$ is a Baire space covered by the closed nowhere dense sets $F_r$, so one of the $F_r$ has nonempty relative interior inside $\\uparrow K$.","core_discovery":"On the paper's own terms, the central claim is Theorem 1.1: for every Polish space $X$ and every hereditary family $\\mathcal{I}\\subseteq\\mathcal{K}(X)$ that is comeager in the Vietoris topology, $\\mathcal{I}$ is big, meaning that it contains a dense hereditary $G_\\delta$ subset of $\\mathcal{K}(X)$. This answers Matheron and Zelený's Problem 2.1 in ZFC. The proof uses their criterion that a set $A\\subseteq\\mathcal{K}(X)$ is big exactly when its complement belongs to $H_{\\mathrm{ext}}$, the $\\sigma$-ideal generated by closed sets $F$ with $F\\cap P(n)$ nowhere dense in $P(n)$ for every $n$, where $P(n)=\\{K\\in\\mathcal{K}(X):|K|\\le n\\}$. The new step is Proposition 2.4, which supplies the needed $H$-membership for the sets $\\downarrow O\\setminus\\downarrow(O\\setminus F)$ when $O$ is open and $F$ is closed and nowhere dense; from that, the complement $B=\\mathcal{K}(X)\\setminus\\mathcal{I}$ is shown to lie in $H_{\\mathrm{ext}}$ by covering each upper cone $\\uparrow K$ with the closed nowhere dense pieces.","pith_inferences":["A repair of Proposition 2.4 would have to replace the non-closed set $\\downarrow O\\setminus\\downarrow(O\\setminus F)$ by a closed (or closed-hereditary) approximation and show the covering argument still places $B$ in $H_{\\mathrm{ext}}$; this is the most direct test of whether the theorem survives the gap.","The concrete failure is caused by finite sets accumulating to a limit compact set outside the family; a corrected definition of $H$ might need to close the candidate sets under Vietoris limits before forming the $\\sigma$-ideal.","If a repaired proof goes through, a similar upper-cone covering could be tried for other hereditary small-set ideals, such as compact sets of Lebesgue measure zero or compact sets of bounded cardinality, where an analogous 'comeager hereditary implies big' question can be posed.","One can probe the theorem on $X=[0,1]$ directly: build a comeager hereditary family whose every dense hereditary $G_\\delta$ subfamily would violate Proposition 2.4; if such a family exists, the theorem itself would fail, not just the proof."],"forward_implications":["Matheron and Zelený's Problem 2.1 and its restatement as Problem 6.9 are settled affirmatively in ZFC, with no definability or set-theoretic hypotheses.","For hereditary families, bigness and comeagerness coincide: every comeager hereditary family is big, and every big family is comeager because it contains a dense $G_\\delta$.","The complement of a comeager hereditary family belongs to $H_{\\mathrm{ext}}$, so it admits a countable cover by closed sets that are nowhere dense on every finite-cardinality layer $P(n)$.","The theorem holds for every Polish space uniformly, including spaces that are not locally compact, so local compactness plays no role."],"supporting_citations":[{"why":"Posed Problem 2.1, proved the coanalytic case, and supplied the characterization Theorem 1.2 that the paper uses.","marker":"[MZ05]"},{"why":"Restated the question as Problem 6.9 and provided the survey context the paper completes.","marker":"[MZ07]"}],"fun_headline_variants":["Comeager hereditary families always have dense G_delta subfamilies","Matheron-Zeleny problem answered in ZFC","ZFC proof: comeager hereditary families are big","Dense G_delta inside every comeager hereditary family","Settled in ZFC: comeager hereditary families are big"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof assumes that a certain difference of two hereditary open families of compact sets is itself closed in the space of compact subsets; the argument never proves this, and for $X=[0,1]$ with $O$ the families contained in $(0,1)$ and $F$ the families meeting $\\{0\\}\\cup\\{1/n:n\\ge1\\}$, that difference is not closed.","fun_headline_variants_meta":{"raw":{"variants":["Comeager hereditary families always have dense G_delta subfamilies","Matheron-Zeleny problem answered in ZFC","ZFC proof: comeager hereditary families are big","Dense G_delta inside every comeager hereditary family","Settled in ZFC: comeager hereditary families are big"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000288,"raw_usage":{"total_tokens":1648,"prompt_tokens":866,"completion_tokens":782,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":482,"completion_tokens_details":{"reasoning_tokens":712}},"tokens_in":482,"tokens_out":782,"duration_ms":6926,"temperature":1.0,"reasoning_tokens":712,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:10:13.937451+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check Proposition 2.4 with $X=[0,1]$, $E=\\{0\\}\\cup\\{1/n:n\\ge1\\}$, $F=\\{K\\in\\mathcal{K}(X):K\\cap E\\neq\\emptyset\\}$, and $O=\\{K:K\\subseteq(0,1)\\}$. Then $C=\\downarrow O\\setminus\\downarrow(O\\setminus F)=\\{K\\subseteq(0,1):K\\cap E\\neq\\emptyset\\}$, and the singleton compact sets $\\{1/n\\}$ lie in $C$ but converge in the Vietoris topology to $\\{0\\}\\notin C$; since $H$ contains only closed subsets of $\\mathcal{K}(X)$, this refutes Proposition 2.4 and thereby the proof's step (3).","supporting_citations":[],"review_version":1}