{"id":"8ff78a94-59a3-4506-9811-1e5e6d9b36f6","arxiv_id":"2506.10529","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Pureness of C*-algebras is preserved under extensions: an algebra is pure iff every closed ideal and its quotient are pure.","lead":"The paper proves that a C*-algebra is pure exactly when both an ideal and its quotient are pure, and uses this to show certain stable multiplier algebras are pure. It introduces a 'separably determined' framework that lets nonseparable algebras be checked through their separable subalgebras.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Backward direction of Theorem 4.11 establishes only (1,1)-pureness; the final step to pureness is delegated to companion result [APTV24, Thm 5.7], with Prop 4.8 and Lemma 4.3 proofs omitted. Conditional verdict is appropriate.","rationale":"The paper's central theorem is significant and the overall architecture of the proof is coherent, but the backward direction of Theorem 4.11 does not prove purity of A on its own. It proves (1,1)-pureness of Cu(A) and then invokes the reduction phenomenon from [APTV24, Theorem 5.7]. That reduction is nontrivial and is not contained in this manuscript; since (1,1)-pureness is strictly weaker than pureness in abstract Cu-semigroups, the missing external theorem is load-bearing. The reader's weakest_assumption identified exactly this dependency, together with the omitted Lemma 4.3. I agree with that assessment. Proposition 4.8 is also explicitly omitted and is needed for the divisibility half of the extension theorem, so it compounds the concern. None of this amounts to evidence that the theorem is false; rather, it means the central claim is not fully verified within the present text and depends on a companion preprint. The conditional verdict is therefore appropriate, and no change to the reader's verdict is needed.","tokens_in":16347,"tokens_out":15558,"duration_ms":188963,"concrete_test":"Check [APTV24, Theorem 5.7] directly: (1) confirm its statement and proof cover arbitrary, not necessarily separable, nuclear, or simple, C*-algebras and all m,n in N; (2) confirm it is proved without invoking the extension permanence result of the present paper; (3) write out the omitted proof of Proposition 4.8 following the pattern of Proposition 4.5, using Lemma 4.3 and Proposition 3.6, and verify that T/(T intersect I) is isomorphic to pi_I(T) for T in the club C'. If (1) or (3) fails, Theorem 4.11 is not established; if both hold, the conditional concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing step is the backward direction of Theorem 4.11. From purity of I and A/I, Propositions 4.5 and 4.8 give only that Cu(A) is (1,1)-pure, i.e. 1-comparison and 1-almost divisibility. The conclusion that A is pure is then imported from [APTV24, Theorem 5.7], a companion preprint by the same research group whose proof is not reproduced or even sketched here. This is not a cosmetic dependency: at the level of Cu-semigroups, (1,1)-pureness is strictly weaker than pureness, so the argument relies entirely on the C*-algebraic reduction phenomenon. If [APTV24, Theorem 5.7] has an unstated hypothesis, such as separability, nuclearity, simplicity, or a restriction on m and n, that is not satisfied by arbitrary C*-algebras or by the multiplier algebras in Example 4.16, then the main biconditional and the advertised application both fail. A second unresolved point is Proposition 4.8, whose proof is omitted but which supplies the divisibility half of the extension theorem; Lemma 4.3 is likewise omitted and is used to transfer comparison and divisibility from separable sub-Cu-semigroups to arbitrary ones. These are precisely the places where an incorrect or inapplicable reduction would break the central claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper establishes a permanence theorem for pureness of C*-algebras under extensions: for a closed ideal I in a C*-algebra A, A is pure if and only if I and A/I are pure (Theorem 4.11). The proof has two main ingredients: a separability-reduction framework for properties of C*-algebras and Cuntz semigroups (Sections 2 and 3), and extension results for (m,n)-pure Cuntz semigroups (Section 4). The authors also apply the theorem to show that the stable multiplier algebras M(C*_red(F_n) ⊗ K) are pure for n = 2,3,...,∞ (Example 4.16).","tokens_in":16641,"tokens_out":4039,"duration_ms":47504,"significance":"If correct, the main theorem is a substantial contribution: it makes pureness an extension-closed property in the non-simple and non-nuclear setting, analogous to the Toms--Winter permanence of Z-stability. The separably-determined framework developed in Sections 2--3 is a useful conceptual tool, and the detailed proofs of the separable base cases (Lemma 4.1 and Lemma 4.7) are carefully written. The application to stable multiplier algebras of reduced free group C*-algebras is new and interesting. The paper is also honest about the limitations of its methods, explicitly leaving open whether strict comparison or almost divisibility individually pass to extensions (Questions 4.13 and 4.14). However, the full force of the main theorem rests on unproved or externally imported ingredients, which is why the present assessment is conditional.","major_comments":[{"comment":"The backward direction of Theorem 4.11 proves only that Cu(A) is (1,1)-pure and then invokes [APTV24, Theorem 5.7] to conclude that A is pure. Since at the level of Cu-semigroups (1,1)-pureness is strictly weaker than pureness, this final step is genuinely load-bearing and the main biconditional is conditional on a companion preprint whose hypotheses are not stated in the present paper. The authors should either state [APTV24, Theorem 5.7] explicitly with all hypotheses, confirm that it applies to arbitrary C*-algebras (including the nonseparable multiplier algebras M(A⊗K) used in Example 4.16), or include a proof or at least a detailed sketch of the reduction. This is not a cosmetic dependency.","section":"Section 4, Theorem 4.11"},{"comment":"The proof of Lemma 4.3 is omitted, with only 'We omit the details.' This lemma is used in Proposition 4.5 to transfer comparison from clubs of separable sub-Cu-semigroups of S/I and I to the ambient Cu-semigroup S. If the club property or the pullback construction fails, Proposition 4.5 does not follow, and hence the comparison half of Theorem 4.11 collapses. The authors should provide a full proof of Lemma 4.3, or give a precise reference that covers exactly the Cu-semigroup statement, including the verification that the pullback of a club is σ-complete and cofinal.","section":"Section 4, Lemma 4.3"},{"comment":"Lemma 4.4 is similarly stated with 'We omit the details.' It is cited as the Cu-semigroup analog of [Thi23, Lemma 3.2(1)], but the transfer from the C*-algebra setting to Cuntz semigroups is not automatic, especially for the σ-completeness of the collection {T ∈ Sep(S) : I ∩ T ∈ D}. Since Lemma 4.4 is also used in the proof of Proposition 4.5, a complete proof or an exact reference is needed.","section":"Section 4, Lemma 4.4"},{"comment":"Proposition 4.8 supplies the divisibility half of the main theorem, yet its proof is omitted ('This is proved analogous to Proposition 4.5. We omit the details.'). The separable case Lemma 4.7 is proved in detail, but the club-transfer argument for n-almost divisibility is not a purely mechanical repetition of Proposition 4.5: the Löwenheim–Skolem step for divisibility (Proposition 3.6) is more delicate than for comparison. The authors should write out the proof or provide a complete reference.","section":"Section 4, Proposition 4.8"}],"minor_comments":[{"comment":"The sentence 'remains open of this property is axiomatizable' should read 'remains open whether this property is axiomatizable'.","section":"Remark 2.5"},{"comment":"There is a typo: 'I ∩ T has m2-comparsion' should be 'm2-comparison'.","section":"Proposition 4.5"},{"comment":"The phrase 'the reduced C*-algebra of the free groupFn' is missing a space; it should be 'the free group F_n'.","section":"Introduction"},{"comment":"The reference to [Thi23, Lemma 3.2(1)] is helpful, but the authors should clarify whether the proof in the Cu-semigroup context requires modifications beyond 'similar methods', since the definition of club in Sep(S) differs from that in the C*-algebra setting.","section":"Section 4, Lemma 4.4"}],"recommendation":"major_revision","confidential_remarks":"The manuscript depends heavily on the companion preprint [APTV24] and on several other preprints by essentially the same group. This is not by itself a problem, but the editorial process should verify that [APTV24] is in a stable form and that Theorem 5.7 indeed applies to arbitrary C*-algebras, including the multiplier algebras in Example 4.16. The omitted proofs of Lemmas 4.3, 4.4, and Proposition 4.8 are load-bearing and must be supplied or replaced with precise references before the paper can be accepted. If those gaps are closed, the paper would be a solid contribution; as it stands, the main theorem is conditional on unpublished material and on unstated hypotheses."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Main take: the paper proves a clean permanence theorem—pureness passes to extensions and quotients—and the proof is mostly solid. The one thing to know is that the final step borrows the reduction from (1,1)-pure to pure from a companion preprint, and three technical lemmas are explicitly omitted. I think the conditional verdict is fair; I'd send it to referees.\n\nWhat's new: Theorem 4.11 is not in the literature. Prior work did Z-stability in extensions (Toms–Winter), and pureness for reduced group C*-algebras, but not the extension permanence. The separably determined framework is a useful tool in its own right. The application to stable multiplier algebras of free groups is a nice consequence and uses known results about coronas.\n\nWhat's good: the writing is clear, the structure is sensible: separability reduction in Section 3, then extension arguments in Section 4. The main proof at the Cu-semigroup level is spelled out in Lemma 4.1 and Lemma 4.7; these are the substantive parts and they look correct to me. The authors also honestly flag that strict comparison and almost divisibility individually may not pass to extensions (Questions 4.13, 4.14), which is the right amount of epistemic modesty.\n\nSoft spots: the backward direction of Theorem 4.11 only yields (1,1)-pureness of Cu(A); the final step to purity is [APTV24, Thm 5.7], a broad reduction result in a companion paper. That's a real dependency, not cosmetic, since (1,1)-pureness is strictly weaker than pureness in general Cu-semigroups. If that companion theorem had hidden hypotheses (e.g., nuclearity or separability), the main biconditional would fail. I can't check [APTV24] here, so the paper is conditional. Also, Lemma 4.3, Lemma 4.4, and Proposition 4.8 are stated with proofs omitted. They look like standard club arguments and the separable-to-general transfer, and I'd bet they fill in without trouble. But they are load-bearing, so a referee should demand them. This is a fixable gap, not a fatal flaw.\n\nCitation pattern: heavy self-citation, but to directly relevant previous work, including the very reduction theorem used. Not a problem.\n\nBottom line: this is a good paper with a real result and a transparent proof strategy. It deserves a serious referee; the referees should push for the omitted details and for confirmation that [APTV24] applies to arbitrary C*-algebras. I'd take it.","headline":"A clean permanence theorem for pureness, well argued but conditional on a companion reduction result and three omitted technical proofs.","tokens_in":17220,"tokens_out":1897,"would_cite":true,"duration_ms":20130,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["46L05","19K14","46L80","46L85"],"pacs":[],"model":"deepseek-v4-flash","headline":"Pureness of a C*-algebra is an extension property: the algebra is pure exactly when each closed ideal and its quotient are pure.","keywords":["C*-algebras","pureness","Cuntz semigroup","strict comparison","almost divisibility","separably determined properties","extensions","multiplier algebras"],"falsifier":"Produce, or find in the literature, a closed ideal $I$ in a C*-algebra $A$ such that $I$ and $A/I$ are both pure but $A$ is not pure; the theorem predicts there is no such extension. Since the paper's own argument would force such an $A$ to be $(1,1)$-pure, the search reduces to deciding whether there exists a $(1,1)$-pure C*-algebra that is not pure, which the companion reduction theorem rules out.","tokens_in":16121,"feed_emoji":"🔗","tokens_out":12181,"duration_ms":114878,"temperature":0.7,"pith_summary":"This paper establishes a permanence property for pure C*-algebras: if $I$ is a closed ideal in a C*-algebra $A$, then $A$ is pure if and only if both $I$ and the quotient $A/I$ are pure. Pureness is the regularity condition on the Cuntz semigroup that combines strict comparison of positive elements with almost divisibility, and it is the relevant notion of regularity for C*-algebras that need not absorb the Jiang-Su algebra. The authors prove that pureness is separably determined, meaning it can be certified on separable subalgebras, and they show that the quantified version of pureness, $(m,n)$-pureness, passes to extensions with an explicit but finite loss in the parameters. As an application, the stable multiplier algebras of reduced free group C*-algebras are shown to be pure even though the underlying algebras fail to be Z-stable.","feed_headline":"A C*-algebra is pure exactly when its ideal and quotient are","feed_subtitle":"The proof uses quantified regularity in Cuntz semigroups, and yields pure stable multiplier algebras of free groups.","key_machinery":"The machinery has three pieces. First, the Cuntz semigroup $\\mathrm{Cu}(A)$ -- the positive elements of $A\\otimes K$ up to Cuntz equivalence, with addition and order from subequivalence -- is the object on which pureness is defined; for an ideal $I$, $\\mathrm{Cu}(I)$ sits inside $\\mathrm{Cu}(A)$ as an ideal and $\\mathrm{Cu}(A/I)$ is the quotient semigroup, by the identification of [CRS10]. Second, the notion of a separably determined property: a property that reflects to a cofinal, $\\sigma$-complete family of separable subobjects and is preserved under inductive limits; the paper shows that pureness, $m$-comparison, and $n$-almost divisibility are separably determined, which reduces extension arguments to separable Cuntz semigroups. Third, the quantified regularity notion of $(m,n)$-pureness, which carries the comparison and divisibility information with explicit parameters and is what actually moves across extensions, with the companion reduction theorem converting any finite-degree pureness into pureness.","core_discovery":"Concretely, the central discovery is that pureness of $A$ is equivalent to pureness of $I$ and $A/I$ (Theorem 4.11). The forward implication follows because comparison and divisibility in $\\mathrm{Cu}(A)$ restrict to ideals and pass to quotients. The backward implication is the substantial part: the authors prove at the level of abstract Cuntz semigroups that an extension of an $m_1$-comparing ideal and an $m_2$-comparing quotient is $(m_1+m_2+1)$-comparing, and an extension of an $n_1$-almost divisible ideal and $n_2$-almost divisible quotient is $\\max\\{2n_1+1,2n_2+1\\}$-almost divisible. Pure algebras are $(0,0)$-pure, so an extension of pure algebras is at least $(1,1)$-pure; the reduction phenomenon stated in a companion paper upgrades this back to pureness. This yields purity for stable multiplier algebras of reduced free group C*-algebras, such as $M(C^*_{\\mathrm{red}}(F_n)\\otimes K)$ for $n=2,3,\\ldots,\\infty$.","pith_inferences":["We infer that the separably determined framework should apply to other Cuntz-semigroup regularity properties phrased with explicit quantifiers, while unquantified variants like controlled comparison can fail to pass to limits (as the paper notes).","We infer that the mechanism behind the free-group application is general: any multiplier algebra sitting over a pure ideal with a purely infinite quotient will be pure, so the result likely extends to other groups satisfying the same combination of stable rank one, unique trace, and strict comparison hypotheses.","We infer that the boundary marked by Questions 4.13 and 4.14 is real: the paper proves only the combined $(1,1)$-pure bound for extensions, so either strict comparison or almost divisibility alone may fail to pass to extensions even though their conjunction does."],"forward_implications":["If a C*-algebra sits as an extension of two pure C*-algebras, then it is pure, and conversely every pure C*-algebra has pure ideals and pure quotients.","For any $n=2,\\ldots,\\infty$, the stable multiplier algebra $M(C^*_{\\mathrm{red}}(F_n)\\otimes K)$ is pure, so pureness can hold for multiplier algebras whose underlying stabilized group C*-algebra is not Z-stable.","Purity of a possibly nonseparable C*-algebra can be verified separably: it is pure exactly when every separable subalgebra is contained in a separable pure subalgebra.","The quantified extension theorem gives explicit bounds: an extension of pure algebras is always $(1,1)$-pure, before the reduction theorem upgrades it to pureness."],"supporting_citations":[{"why":"Supplies the reduction phenomenon that any (m,n)-pure C*-algebra is pure, which is the load-bearing upgrade in the backward direction of Theorem 4.11.","marker":"[APTV24, Theorem 5.7]"},{"why":"Identifies Cu(I) as an ideal in Cu(A) with Cu(A/I) as the corresponding quotient, connecting C*-algebra extensions to Cuntz semigroup extensions.","marker":"[CRS10]"},{"why":"Provides the club of separable sub-Cu-semigroups with order-embedding Cuntz semigroups, underpinning the separably determined machinery used throughout the proof.","marker":"[TV21, Proposition 6.1]"},{"why":"Shows the corona algebra is purely infinite under the exactness and trace hypotheses, giving the pure quotient needed in Proposition 4.15(1).","marker":"[KNP10, Theorem 4.5]"},{"why":"Gives pure infiniteness of the stable corona algebra in the stable-rank-one unique-trace setting, used in Proposition 4.15(2).","marker":"[KNZ19, Corollary 6.12]"},{"why":"Establishes strict comparison with respect to the unique tracial state for the reduced group C*-algebras appearing in Example 4.16.","marker":"[AGKEP24, Theorem B]"}],"fun_headline_variants":["C*-algebra purity equivalent to ideal and quotient purity","Purity survives extensions of C*-algebras","Ideal and quotient purity force C*-algebra purity","Extension theorem for pure C*-algebras","Stable multipliers of free group C*-algebras are pure"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the companion reduction theorem that any C*-algebra that is $(m,n)$-pure for some finite $m,n$ is automatically pure, with the proof also relying on Lemma 4.3, whose proof is omitted but which is needed to pass from separable to arbitrary Cuntz semigroups.","fun_headline_variants_meta":{"raw":{"variants":["C*-algebra purity equivalent to ideal and quotient purity","Purity survives extensions of C*-algebras","Ideal and quotient purity force C*-algebra purity","Extension theorem for pure C*-algebras","Stable multipliers of free group C*-algebras are pure"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000407,"raw_usage":{"total_tokens":2061,"prompt_tokens":839,"completion_tokens":1222,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":455,"completion_tokens_details":{"reasoning_tokens":1157}},"tokens_in":455,"tokens_out":1222,"duration_ms":13821,"temperature":1.0,"reasoning_tokens":1157,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T04:23:47.699231+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Produce, or find in the literature, a closed ideal $I$ in a C*-algebra $A$ such that $I$ and $A/I$ are both pure but $A$ is not pure; the theorem predicts there is no such extension. Since the paper's own argument would force such an $A$ to be $(1,1)$-pure, the search reduces to deciding whether there exists a $(1,1)$-pure C*-algebra that is not pure, which the companion reduction theorem rules out.","supporting_citations":[],"review_version":1}