{"id":"007cb3b9-6c31-4e55-b579-3cfc6a5181c0","arxiv_id":"2411.17568","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every abstract logic has a weak compactness cardinal iff Ord is essentially faint, and this is compatible with having no strongly inaccessible cardinals.","lead":"This paper characterizes when every abstract logic has a weak compactness cardinal: it is exactly equivalent to a new combinatorial scheme about the ordinals, called essential faintness. It then shows this situation can coexist with the absence of strongly inaccessible cardinals, while still forcing unboundedly many of them inside the inner model HOD.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.14's forward direction rests on an unproved preservation of subtlety in the factor model M; this is the load-bearing gap, not the HOD absoluteness issue.","rationale":"The reader's weakest assumption targets Corollary 4.5's claim that 'Ord is essentially subtle' is downward absolute to HOD. This particular concern does not land: any definable instance of the scheme in HOD is also a definable instance in V, and the witness (α,β,A) produced in V automatically lies in HOD because A∈E_β⊆P(β)^HOD. So that part of Corollary 4.5 is sound. However, the reader's broader caution about the complex forcing proof of Theorem 1.14 is justified. The step 'Then δ is a subtle cardinal in M' is asserted without proving that the factor model M has no new δ-sequences, and subtlety is exactly a statement about all such sequences. This is not a mere technicality: if δ fails to be subtle in M, then Lemma 3.4 cannot be invoked, and the generic extension may lack the proper class of weakly shrewd cardinals that is needed to conclude that 'Ord is essentially faint' holds and that no inaccessible cardinals exist. The central equivalence Theorem 1.6 is not affected by this gap, but Theorem 1.14—the headline separation result—depends on this preservation claim. Therefore the reader's CONDITIONAL verdict is appropriate, though for a different reason than the one stated.","tokens_in":34516,"tokens_out":45753,"duration_ms":429397,"concrete_test":"Formalize the factor analysis in the proof of Theorem 1.14: state precisely what M is (e.g., the Easton product of the stages below δ) and verify whether P(δ)^M⊆V, together with V=L, implies that every δ-indexed sequence in M is an element of V. Concretely: take a single subtle cardinal δ in L, force with the Easton product Q of stages below δ, and check whether δ remains subtle in M[G0] by testing the defining property against all Q-names for δ-sequences. If a Q-name forces a new δ-sequence that is not coded by a ground-model subset of δ, the preservation of subtlety fails and Theorem 1.14's forward direction collapses.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the proof of Theorem 1.14 (Section 4), after the 'standard factor analysis' one obtains an inner model M with V⊆M, P(δ)^M⊆V and a generic G0×G1 over M such that V[G]=M[G0,G1]. The next sentence asserts: 'Then δ is a subtle cardinal in M.' This is load-bearing: Lemma 3.4 is then applied inside M to produce an inaccessible weakly shrewd κ in (R(δ),δ), and the rest of the argument lifts κ to V[G]. Subtlety of δ, however, quantifies over all δ-indexed sequences ⟨A_α | α<δ⟩ in M with A_α⊆α. The condition P(δ)^M⊆V only places subsets of δ inside V; a δ-sequence is a function δ→V, not a subset of δ. In the ground model V=L one can of course code such a sequence by a subset of δ using the canonical well-ordering, but the paper does not say this, and it is not automatic: a <δ-closed forcing (which the Easton product below δ is) can add new subsets of δ and hence new δ-sequences. Since M is an inner model of the class-forcing extension, it is not self-evident that all M-sequences of length δ belong to V. If δ is not subtle in M, the construction yields no weakly shrewd cardinal below δ, and the claim that V[G] has a proper class of weakly shrewd cardinals (and hence that ZFC+φ proves 'Ord is essentially faint') is unsupported. This is exactly the step that carries the main consistency result Theorem 1.14.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies weak compactness cardinals for all abstract logics and characterizes their existence through new combinatorial principles for the class of ordinals, 'Ord is essentially faint' and 'Ord is essentially subtle'. The main results are Theorem 1.6 (Ord is essentially faint iff every abstract logic has a weak compactness cardinal), Theorem 1.4 (the stationary version is equivalent to Ord being essentially subtle), Theorem 1.14 (an equiconsistency between a proper class of subtle cardinals and the failure of strongly inaccessible cardinals together with essential faintness), and Theorem 1.15 (essential faintness implies unboundedly many strongly inaccessible cardinals in HOD). The proofs use C(n)-weakly shrewd cardinals, forcing with Easton products, and coding arguments for abstract logics.","tokens_in":34865,"tokens_out":55002,"duration_ms":517939,"significance":"If the results are correct, they provide the first ZFC characterizations of weak compactness cardinals for arbitrary abstract logics, resolve a natural question raised by Boney–Dimopoulos–Gitman–Magidor, and sharply delimit the large-cardinal strength of the existence of such cardinals. The paper is careful and detailed, with full proofs of the main equivalences and explicit constructions of the relevant logics. Strong points include the parameter-free combinatorial formulations, the use of C(n)-weakly shrewd cardinals as a bridge between model-theoretic and set-theoretic properties, and the exact equiconsistency statements. The main reservations concern one load-bearing step in the consistency proof of Theorem 1.14 that is stated without sufficient justification.","major_comments":[{"comment":"The forward direction of Theorem 1.14 begins with the sentence 'Work in a model of ZFC+ V = L in which a proper class of subtle cardinals exists and no inaccessible cardinal is a limit of subtle cardinals.' The existence of such a model is not proved or cited, and it is not immediate from the consistency of ZFC + 'there is a proper class of subtle cardinals', since in L the class of all subtle cardinals may have inaccessible limit points. This assumption is load-bearing: the claim that V[G] has no inaccessible cardinals uses the inequality R(δ)<κ for an inaccessible κ in V[G], and if κ were an inaccessible limit of the chosen class S in the ground model, then sup(S∩δ)=κ and R(δ)>κ, so the argument would fail. Please supply a proof or reference that the required starting model exists, or modify the construction to use a sparse subclass of subtle cardinals with no inaccessible limit point.","section":"Section 4, proof of Theorem 1.14"},{"comment":"After the 'standard factor analysis', the proof asserts 'Then δ is a subtle cardinal in M'. This is true, but it is not automatic from P(δ)^M⊆V alone, and the step is not justified in the text. A δ-sequence ⟨A_α | α<δ⟩ with A_α⊆α is coded by a subset of δ×δ, hence by a subset of δ via a standard pairing function; since P(δ)^M⊆V, both the code and any M-club C⊆δ lie in V, so subtlety of δ in V applies to the decoded sequence and club. Please add this argument (or a reference) at this point.","section":"Section 4, proof of Theorem 1.14, factor analysis"}],"minor_comments":[{"comment":"The claim that 'Ord is essentially subtle' is downward absolute to HOD in the first case of Corollary 4.5 would benefit from a brief justification: HOD-definable class sequences and club classes are V-definable via the definition of HOD, so the scheme in V directly yields the scheme in HOD.","section":"Section 1, after Definition 1.3"},{"comment":"In the sentence 'Then δ is a subtle cardinal in M', the paper should also clarify that the relevant sequence in M is coded by a subset of δ, since the values A_α are subsets of α and hence the entire sequence is a subset of δ×δ.","section":"Section 4, proof of Theorem 1.14"},{"comment":"There are occasional typographical issues, such as 'comp actness' in the title, which should be corrected to 'compactness'.","section":"Title and throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and the central results appear plausible and significant. The main issue is the missing consistency argument for the ground model used in the proof of Theorem 1.14; this is a load-bearing step that should be fixed by adding a lemma or a reference. The other points are local and easily addressed. I do not see grounds for rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Good paper, worth careful reading. Lücke answers both questions left open in [5]: Ord essentially faint iff every abstract logic has a weak compactness cardinal, and the essential subtle version for stationary classes. That is clean, genuinely new material. The consistency analysis in Theorem 1.14 is substantial: starting from V=L with a proper class of subtle cardinals, he kills strong inaccessibles while preserving enough weak shrewdness to keep Ord essentially faint. The contrast with strong compactness is exactly the right kind of result. The new C(n)-weakly shrewd cardinals are a sensible tool, and the embedding characterization in Lemma 3.3 does the work.\n\nWhere the paper earns real credit: the equivalences are proved by building logics from non-shrewd cardinals and then using reflection, and the forcing construction is standard enough to be checkable. Theorems 5.5 and 6.7 extend the scope to strict and weak Lowenheim-Skolem-Tarski numbers. The paper is clearly written and the main line of argument holds together.\n\nTwo places need extra justification, and neither looks fatal. In the proof of Theorem 1.14, after the factor analysis the paper asserts \"Then δ is a subtle cardinal in M.\" The stress-test worry about <δ-closed forcing adding δ-sequences does not land, because the ground model is L: any δ-indexed sequence in M can be coded by a subset of δ using the canonical L-well-order, P(δ)^M⊆V brings the code into V, and then the sequence itself is in V. M-clubs are also V-clubs. But the paper should say this; as written it is a genuine gap in exposition, not a load-bearing error. Second, in Corollary 4.5 the step \"if Ord is essentially subtle, then Ord is essentially subtle in HOD\" is asserted without proof. It is defensible: HOD-definable sequences and clubs are V-definable, the witnesses produced in V lie in HOD because the sets Eα do, and membership is absolute. A one-paragraph clarification would settle it.\n\nMinor issues: the abstract-logic machinery in Section 6 is long and the construction of the separating logics could be broken out more explicitly, but it is standard for this program. Citations to [2], [5], and [13] are appropriate and not padding.\n\nBottom line: this is a solid contribution that deserves a serious referee. With the two justifications added, I would be happy to accept.","headline":"Lücke proves the right ZFC characterizations and a sharp consistency contrast; the two suspicious steps in Theorem 1.14 and Corollary 4.5 are exposition gaps, not fatal flaws.","tokens_in":35431,"tokens_out":4872,"would_cite":true,"duration_ms":66315,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["03B16","03C55","03E45","03E55"],"pacs":[],"model":"deepseek-v4-flash","headline":"Weak compactness for every abstract logic is equivalent, over ZFC, to the scheme \"Ord is essentially faint\"; the scheme can hold without any strongly inaccessible cardinals, yet forces unboundedly many in HOD.","keywords":["abstract logic","weak compactness cardinal","Ord is essentially faint","C(n)-weakly shrewd cardinal","subtle cardinal","hereditarily ordinal definable sets","Löwenheim–Skolem–Tarski number","large cardinals"],"falsifier":"A model of ZFC in which \"Ord is essentially faint\" holds but some explicitly defined abstract logic has no weak compactness cardinal would falsify Theorem 1.6; the theorem says no such model exists. For the HOD conclusion, a model in which \"Ord is essentially subtle\" holds in $V$ but not in HOD would falsify the transfer assumption used in the first case of the proof of Corollary 4.5.","tokens_in":34314,"feed_emoji":"♾️","tokens_out":13202,"duration_ms":107864,"temperature":0.7,"pith_summary":"The paper establishes an exact structural characterization: in ZFC, every abstract logic has a weak compactness cardinal if and only if the class of ordinals satisfies a principle called \"Ord is essentially faint.\" It then shows this principle does not force the existence of strongly inaccessible cardinals: the consistency of a proper class of subtle cardinals is exactly the consistency of \"Ord is essentially faint\" together with the absence of strongly inaccessible cardinals. It also proves that \"Ord is essentially faint\" nevertheless forces unboundedly many ordinals to be strongly inaccessible inside the inner model HOD of all hereditarily ordinal definable sets. The upshot is that weak compactness for all abstract logics is a genuinely weaker global property than strong compactness, governed by a single combinatorial scheme rather than by a large-cardinal existence statement.","feed_headline":"Every abstract logic has weak compactness iff Ord is faint","feed_subtitle":"One combinatorial scheme about Ord decides compactness for every abstract logic, and no inaccessible cardinals are needed.","key_machinery":"Let $C(n)$ be the class of ordinals $\\alpha$ with $V_\\alpha$ a $\\Sigma_n$-elementary substructure of $V$. The central combinatorial object is the class version of faintness: \"Ord is essentially faint\" says that for every definable class sequence $\\langle E_\\alpha\\rangle$ with $\\varnothing\\neq E_\\alpha\\subseteq\\mathcal{P}(\\alpha)$, and every ordinal $\\xi$, there are $\\xi<\\alpha<\\beta$ and $A\\in E_\\beta$ with $A\\cap\\alpha\\in E_\\alpha$. The stronger scheme \"Ord is essentially subtle\" replaces the final \"every ordinal $\\xi$\" by \"every closed unbounded class $C$\", requiring $\\alpha<\\beta$ both in $C$. Between these schemes and abstract logics stand the $C(n)$-weakly shrewd cardinals, defined by reflection of formulas to $\\Sigma_n$-correct levels $H(\\theta)$; Theorem 3.5 equates essential faintness with the existence of a proper class of such cardinals at every level $n$. On the model-theoretic side, Lemma 6.1 proves that every sufficiently large $C(n)$-weakly shrewd cardinal is a weak compactness cardinal for every abstract logic definable at level $n$, and Lemma 6.4 constructs inverse abstract logics that force any weak compactness cardinal to be $C(n)$-weakly shrewd. HOD enters through a lemma in which a $C(n)$-weakly shrewd cardinal that fails to be $C(n)$-strongly unfoldable produces a subtle cardinal in HOD.","core_discovery":"The central claim is Theorem 1.6: over ZFC, the scheme \"Ord is essentially faint\" is equivalent to the statement that every abstract logic has a weak compactness cardinal. Here a weak compactness cardinal for an abstract logic $L$ is a cardinal $\\kappa$ such that every $L$-theory of size $\\kappa$ all of whose subtheories of smaller size are satisfiable is itself satisfiable. The proof proceeds by a two-sided large-cardinal analysis: Section 3 shows that \"Ord is essentially faint\" is equivalent to the existence of a proper class of $C(n)$-weakly shrewd cardinals for every natural number $n$, and Section 6 shows that such cardinals are exactly the cardinals that can serve as weak compactness cardinals for definable abstract logics. A forcing argument then separates the scheme from strong inaccessibility: starting from a proper class of subtle cardinals, the paper produces a model with no strongly inaccessible cardinals in which a proper class of weakly shrewd cardinals survives, so that Ord is essentially faint (Theorem 1.14). Complementing this, Theorem 1.15 and Corollary 4.5 show that weak compactness for all logics forces unboundedly many ordinals to be strongly inaccessible in HOD, and in fact to be strongly inaccessible $C(n)$-weakly shrewd there.","pith_inferences":["A testable extension is to restrict the theorem to logics with a fixed occurrence number; the same bridge should yield bounded versions of essential faintness, producing a hierarchy between first-order compactness and full faintness for Ord.","The forcing construction behind Theorem 1.14 suggests that \"Ord is essentially faint\" is compatible not only with the absence of inaccessible cardinals but also with a proper class of singular cardinals violating the Singular Cardinal Hypothesis, connecting the paper's Question 7.1 to cardinal arithmetic.","Corollary 4.5 singles out an internal absoluteness question: whether \"Ord is essentially subtle\" passes from $V$ to HOD. A model where it fails would show Theorem 1.15 needs a different proof in that case, while a model where it holds would complete the first case of the corollary."],"forward_implications":["Over ZFC, \"Ord is essentially faint\" and \"every abstract logic has a weak compactness cardinal\" become two names for the same fact, settling the questions posed in the introduction without needing a definable well-ordering of the universe.","Consistency of a proper class of subtle cardinals is equivalent to consistency of \"Ord is essentially faint\" together with the statement that there are no strongly inaccessible cardinals.","Weak compactness for all abstract logics forces unboundedly many ordinals to be strongly inaccessible in HOD, and in fact to be strongly inaccessible $C(n)$-weakly shrewd cardinals there.","Ordinal subtlety gives the model-theoretic companions: \"Ord is essentially subtle\" implies stationary classes of weak and strict Löwenheim–Skolem–Tarski numbers for every abstract logic, while \"Ord is essentially faint\" implies a strict Löwenheim–Skolem–Tarski number for every abstract logic.","In any model of \"Ord is essentially faint\" in which only boundedly many cardinals are subtle in HOD, the stronger scheme \"Ord is essentially subtle\" must hold."],"supporting_citations":[{"why":"It supplies the original connection between subtlety of the ordinals and weak compactness cardinals under a definable well-ordering, which this paper removes and strengthens.","marker":"[5]"},{"why":"It introduces the \"essentially subtle\" scheme and the $C(n)$-strongly unfoldable characterization that marks the contrasting principle.","marker":"[2]"},{"why":"It defines faintness for cardinals, the property whose class version is the paper's main scheme.","marker":"[17]"},{"why":"It introduces weakly shrewd cardinals, the base large-cardinal notion whose $C(n)$-versions carry the main equivalence.","marker":"[13]"},{"why":"It proves the analogous strong-compactness characterization that the paper contrasts with the weak case.","marker":"[16]"},{"why":"It provides the canonical well-ordering of HOD and the forcing facts used in the separation and HOD arguments.","marker":"[10]"},{"why":"It introduces subtle cardinals, the consistency hypothesis behind Theorem 1.14.","marker":"[11]"}],"fun_headline_variants":["Ord's faintness is weak compactness for all logics","Weak compactness for every logic iff Ord is faint","No inaccessible cardinals required for weak compactness","Subtlety of Ord determines weak compactness","Weak compactness all logics: subtlety of Ord suffices"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the scheme \"Ord is essentially subtle\" passes from the universe to the inner model HOD: the proof of Corollary 4.5 assumes that if the scheme holds in $V$, it also holds inside HOD, even though the club and witness sequences involved need not be ordinal-definable.","fun_headline_variants_meta":{"raw":{"variants":["Ord's faintness is weak compactness for all logics","Weak compactness for every logic iff Ord is faint","No inaccessible cardinals required for weak compactness","Subtlety of Ord determines weak compactness","Weak compactness all logics: subtlety of Ord suffices"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000244,"raw_usage":{"total_tokens":1556,"prompt_tokens":990,"completion_tokens":566,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":606,"completion_tokens_details":{"reasoning_tokens":487}},"tokens_in":606,"tokens_out":566,"duration_ms":5821,"temperature":1.0,"reasoning_tokens":487,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T11:59:18.815828+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A model of ZFC in which \"Ord is essentially faint\" holds but some explicitly defined abstract logic has no weak compactness cardinal would falsify Theorem 1.6; the theorem says no such model exists. For the HOD conclusion, a model in which \"Ord is essentially subtle\" holds in $V$ but not in HOD would falsify the transfer assumption used in the first case of the proof of Corollary 4.5.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It supplies the original connection between subtlety of the ordinals and weak compactness cardinals under a definable well-ordering, which this paper removes and strengthens."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It introduces the \"essentially subtle\" scheme and the $C(n)$-strongly unfoldable characterization that marks the contrasting principle."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It defines faintness for cardinals, the property whose class version is the paper's main scheme."},{"cited_title":"Makowsky, Vopˇ enka’s principle and compact logics, J","cited_arxiv_id":null,"evidence_quote":"It proves the analogous strong-compactness characterization that the paper contrasts with the weak case."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It provides the canonical well-ordering of HOD and the forcing facts used in the separation and HOD arguments."},{"cited_title":"Jensen and Kenneth Kunen, Some combinatorial properties of L and V , handwritten notes, 1969","cited_arxiv_id":null,"evidence_quote":"It introduces subtle cardinals, the consistency hypothesis behind Theorem 1.14."}],"review_version":1}