{"id":"328eb268-49b9-4096-a47f-0cc4364899e4","arxiv_id":"2607.14358","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"State-preserving completely positive maps form a Polish space; subalgebras admitting state-preserving conditional expectations form a closed subspace whose topology agrees with the Effros-Maréchal topology, and in many type III settings such subalgebras are generic only as the complement.","lead":"This paper studies the spaces of quantum maps that preserve a fixed state, showing they form a well-behaved Polish space and that this topology matches the standard Effros-Maréchal topology on subalgebras. It uses this framework to prove that many 'good' subalgebras form closed sets, and that in many type III algebras the generic subalgebra admits no state-preserving conditional expectation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection to central claims; the Isono dependency in Thm 5.5 is a secondary black-box concern, not load-bearing for the main results.","rationale":"The reader's weakest assumption points to Theorem 5.5's reliance on Isono's relative-amenability criterion. I agree this is a black box and worth checking, but the central advertised results — Polish closedness of E_φ and density of E_φ^c — do not depend on it. The internal proofs of Theorems 4.3, 4.7, and 6.2 are coherent; I verified the key inequalities and found no gaps. The CP closedness in Proposition 3.3 is actually correct because scaling all test vectors by a common factor preserves the sign of the quadratic form, so checking on the unit ball of N is sufficient. Therefore no load-bearing objection to the central claim; the paper can be accepted as is, with a specialist check of Isono's hypotheses as a minor caveat.","tokens_in":24000,"tokens_out":34569,"duration_ms":289786,"concrete_test":"Verify the hypotheses of [Iso19, Theorem 3.2 and Appendix Theorem A.6] for an arbitrary inclusion P⊆N⊆M of type III von Neumann algebras with P,N∈E_φ: in particular, confirm that the continuous cores eN, eP are semifinite and that the weak-containment criterion does not require factoriality or finiteness. If the hypotheses fail, Theorem 5.5's closedness of A_P needs additional assumptions; if they hold, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim — E_φ closed and Polish (Thm 4.3), topology equality (Thm 4.7), and generic absence of expectations (Thm 6.2) — is supported by coherent internal proofs. I checked the cruxes: the square-map continuity in Thm 4.3, the approximation argument in Thm 4.7, and the convergence of B_n→B in Thm 6.2. No fatal gap found. The reader's flagged concern about Thm 5.5 is legitimate but secondary: it uses [Iso19, Thm 3.2 & A.6] as a black box for relative amenability of arbitrary type III inclusions, and the hypotheses are not re-verified. However, this theorem is not needed for the main E_φ topology or genericity results; Prop 5.2 (closedness of amenable subalgebras) has an independent proof. Thus the central claim stands; the Isono reliance is a caveat for a standalone generalization, not for the paper's headline assertions.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies spaces of state-preserving unital completely positive (ucp) maps between von Neumann algebras equipped with faithful normal states. It proves that this space is Polish for the topology of pointwise strong (equivalently 2-norm) convergence, and it identifies the set E_φ of subalgebras admitting a φ-preserving conditional expectation with a closed subspace of the space of idempotent state-preserving ucp maps. The main topological claims are that E_φ is Polish and that its induced 1-Lipschitz topology coincides with the restriction of the Effros–Maréchal topology on S(M), recovering a theorem of Tsukada and Haagerup–Winslow. These tools are then applied to structural classes: amenable, relatively amenable, Haagerup, and weakly amenable subalgebras, with closedness or lower-semicontinuity results. In the final section, the paper shows that for certain type III algebras the subalgebras lacking a state-preserving conditional expectation form an open dense set, and characterizes non-finite von Neumann algebras by denseness of subalgebras without any normal conditional expectation.","tokens_in":24096,"tokens_out":43914,"duration_ms":417839,"significance":"If the proofs are completed, the paper makes a solid contribution: it unifies several natural topologies on UCP spaces, gives a clean Polish topology on E_φ, recovers the Haagerup–Winslow correspondence, and provides new generic-density results for subalgebras lacking conditional expectations. The overall organization is clear, and the main constructions — idempotent UCP maps, projection convergence, approximation arguments — are mostly presented in detail. The use of prior work, especially [FMMP24], is explicit. However, the proof of the key topology-equality theorem (Theorem 4.7) contains an incorrect operator identity, and one relative-amenability result relies on an external black box whose hypotheses are not verified. These issues need to be addressed before the paper can be accepted.","major_comments":[{"comment":"The proof of the second half of Theorem 4.7 uses the identity E_N^φ(x)=e_N x e_N, stated before Theorem 4.3, to justify the equality sup_{x∈M1} |⟨E_N(x)ξ,ξ⟩−⟨E_Nn(x)ξ,ξ⟩| = sup_{x∈M1} |⟨x e_N ξ,e_N ξ⟩−⟨x e_Nn ξ,e_Nn ξ⟩| for an arbitrary vector ξ. This identity is false on all of L^2(M,φ); the correct relation is e_N x e_N = E_N^φ(x)e_N, i.e. the identity holds only on L^2(N,φ). For example, take M=M_2(C), N the diagonal matrices, φ=tr, ξ=E_{12}, and x=E_{11}. Then E_N(x)ξ=E_{11}E_{12}=E_{12}, so ⟨E_N(x)ξ,ξ⟩=1, while e_N ξ=0, so ⟨x e_N ξ,e_N ξ⟩=0. Since this equality is the step that proves Maréchal convergence from pointwise 2-norm convergence, the proof of the second half of Theorem 4.7 is incomplete. The theorem itself is true (Tsukada/HW98), so this is repairable, but the proof as written needs correction.","section":"§4, Theorem 4.7 (second half)"},{"comment":"The proof of closedness of relatively amenable subalgebras A_P passes through the continuous core and invokes [Iso19, Theorem 3.2 and Appendix Theorem A.6] as black boxes. The manuscript does not state or verify the hypotheses of these results in the stated generality of arbitrary type III inclusions with a faithful normal state φ. If Isono's criterion requires additional assumptions (e.g. factoriality or specific core conditions), the closedness assertion may fail as stated. This theorem is not needed for the main E_φ topology results or for Theorem 6.2, so the issue is secondary, but it should be addressed by either verifying the hypotheses or restricting the statement.","section":"§5.1, Theorem 5.5"}],"minor_comments":[{"comment":"There are several typos: 'Tsukuda' should be 'Tsukada' (Theorem 4.7 heading), 'Hany' should be 'any' (Remark 5.11), 'idempodent' should be 'idempotent' (Lemma 4.2), and 'posses' should be 'possess' (Section 4).","section":"Throughout"},{"comment":"Definition 5.6 says 'We say that M has the Haagerup property relative to N', but Theorem 5.7 concerns 'N has the Haagerup property relative to P'. Please clarify which algebra is being tested and make the notation consistent.","section":"§5.2, Definition 5.6 and Theorem 5.7"},{"comment":"The notation E^ω(fM) appears in the proof of Theorem 5.5, but the paper elsewhere writes E_φ. Please define E^ω (or E_ω) explicitly when it is first used.","section":"§4, Theorem 4.7 and §5.5"},{"comment":"The proof assumes the existence of an increasing sequence q_n∈M_φ with q_n→1 strongly and p_n=1−q_n nonzero for all n. This is true under the stated hypothesis that M_φ is infinite-dimensional, but a one-sentence justification (e.g. existence of a countable partition of unity in the σ-finite algebra M_φ) would help.","section":"§6, Theorem 6.2"},{"comment":"Remark 4.6 says that separate continuity of composition implies Φ↦Φ^2 is Baire class 1 for the pointwise ultraweak topology. This is not immediate and, if used, needs a reference or a proof. Since the remark is non-essential, this is a minor point.","section":"§3, Remark 4.6"}],"recommendation":"major_revision","confidential_remarks":"The main obstacle is the flawed proof of the second half of Theorem 4.7. Because the theorem is known, I expect the authors can fix it by citing Tsukada/HW98 directly or by supplying a correct argument. The rest of the paper appears sound, but I would like to see the Isono dependency in Theorem 5.5 clarified before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper delivers what the abstract promises: a clean Polish-space framework for state-preserving ucp maps, a closed subspace E_phi of subalgebras admitting a phi-preserving conditional expectation, and a demonstration that the induced topology matches Effros–Maréchal. The authors are explicit that the topology result recovers Haagerup–Winslow; the genuinely new content is the framework itself and the applications: closedness of amenable and Haagerup subalgebras in E_phi, lower semicontinuity and failure of continuity of the Cowling–Haagerup constant, and the open-density results for subalgebras without state-preserving expectations. I checked the main proofs (Prop 3.3, Thm 4.3, Thm 4.7, Thm 6.2, Thm 6.4) and they hold together. The square-map continuity argument, the projection convergence via Lemma 2.3, and the approximation in Thm 4.7 are all sound. The self-citation to [FMMP24] is legitimate; the Polishness of Maréchal topology is an independent prior result and the paper says so.\n\nThe soft spots are real but not load-bearing. Theorem 5.5, closedness of relatively amenable subalgebras, rests on Isono’s relative-amenability/core criterion ([Iso19, Thm 3.2, A.6]) as a black box, with hypotheses not re-verified for arbitrary type III inclusions. A specialist should check that step before building on it. But the reader and stress-test are right: this theorem is not needed for the paper’s headline results, and Proposition 5.2 (amenable subalgebras) has an independent proof. The examples in 5.15/5.17 use external results (Ozawa, Cowling–Haagerup) but those citations are standard and the logic is transparent. There are also quite a few typos and minor notational slips; they don’t obscure the arguments.\n\nThis is a careful, honest paper. The central claims are proven in detail, the literature is engaged with fairly, and the new applications are nontrivial. It deserves a serious referee, and I would send it out as-is rather than desk reject.","headline":"Solid, well-written operator algebra paper: the Polish-space framework for state-preserving ucp maps is real, the recovery of Haagerup–Winslow is honest, and the new applications (closedness of amenable/Haagerup subalgebras, semicontinuity of Lambda_cb, genericity of no expectation) are genuine, with the main caveat being a black-box use of Isono in Theorem 5.5.","tokens_in":24773,"tokens_out":914,"would_cite":true,"duration_ms":11330,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["46L10","54H05"],"pacs":[],"model":"deepseek-v4-flash","headline":"Subalgebras admitting state-preserving conditional expectations form a closed Polish space, and in many factors their absence is generic.","keywords":["ucp maps","conditional expectations","Effros–Maréchal topology","Polish space","von Neumann algebras","amenability","Haagerup property","Cowling-Haagerup constant"],"falsifier":"A single counterexample to Theorem 5.5: a sequence N_n of subalgebras in E_φ converging to N with each N_n amenable relative to P but N not amenable relative to P. Such a sequence would directly falsify the closedness claim. A second falsifier for Theorem 6.2 would be finding a type III_λ factor with λ<1 and a faithful normal state for which subalgebras lacking the expectation are not dense.","tokens_in":23736,"feed_emoji":"🧮","tokens_out":9165,"duration_ms":85801,"temperature":0.7,"pith_summary":"This paper establishes that for a von Neumann algebra equipped with a faithful normal state, the collection of subalgebras that are the ranges of state-preserving conditional expectations forms a closed subspace of the space of all state-preserving unital completely positive maps. Being closed in a Polish space, it inherits a Polish topology, which the paper shows coincides with the Effros–Maréchal topology on subalgebras. On this space, amenability, the Haagerup property, and bounded Cowling–Haagerup constant are closed properties, and the Cowling–Haagerup constant is lower semicontinuous. In the opposite direction, the paper proves that for many purely infinite factors, the subalgebras lacking a state-preserving conditional expectation form an open and dense set, so a generic subalgebra is not the range of such an expectation. The upshot is a clear topological picture of which subalgebras can be \"expected onto\" and which structural properties are stable under limits.","feed_headline":"No state-preserving expectation for the generic subalgebra in type III factors","feed_subtitle":"The paper proves the expected subalgebras form a closed Polish space—and that for many factors, generic subalgebras avoid it.","key_machinery":"The key machinery is the bijection between subalgebras in E_φ and their unique φ-preserving conditional expectations, which are exactly the idempotent state-preserving ucp maps. The map sending a ucp map to its square is continuous in the topology of pointwise strong convergence, so the idempotent ones form a closed set; then each idempotent is identified with the conditional expectation onto its fixed-point algebra. For the density results, the construction replaces a corner of a given subalgebra with a subalgebra of a corner algebra that admits no conditional expectation, using the classical characterization of finite von Neumann algebras by the existence of a masa with no normal condition","core_discovery":"The central discovery is that E_φ—the set of von Neumann subalgebras of M that admit a φ-preserving conditional expectation—is closed in the Polish space of state-preserving ucp maps, making it Polish, and that its topology is exactly the restriction of the Effros–Maréchal topology. This is obtained by identifying E_φ with the set of idempotent state-preserving ucp maps, observing that Φ ↦ Φ² is continuous, so idempotency is closed. The paper also shows the complement direction: when M is purely infinite and the centraliser M_φ is infinite-dimensional, the subalgebras without a φ-preserving expectation are open and dense, so the existence of such an expectation is a rare property.","pith_inferences":["If the generic-density result extends to all states on type III_1 factors (the generic state has trivial centraliser), then the phenomenon would be even more sweeping: expectation-free subalgebras would be residual for a generic state, not just for states with infinite-dimensional centraliser.","The closedness of relatively amenable subalgebras suggests a tool for detecting relative amenability by approximation: a limit of relatively amenable subalgebras remains relatively amenable, which may be useful in rigidity questions for subfactor lattices.","The dichotomy—closed Polish space of expected subalgebras versus open dense complement—could be read as an obstruction to any continuous parametrization of conditional expectations over the whole space of subalgebras.","A testable extension: determine whether the set Exp(M) in Theorem 6.4 is Borel for non-finite M; the paper leaves this as an open question."],"forward_implications":["Amenable subalgebras with φ-preserving expectations form a closed set; a non-amenable factor admitting such an expectation cannot be approximated from inside by amenable subalgebras with expectations.","The Cowling–Haagerup constant is lower semicontinuous on E_φ, and the set of weakly amenable subalgebras is F_σ; moreover the paper constructs examples where the constant is not continuous, so it is genuinely only semicontinuous.","Subalgebras with the relative Haagerup property form a closed set within E_φ, giving a unified proof of closedness for amenable and Haagerup classes.","For purely infinite algebras whose state has infinite-dimensional centraliser—including all type III_λ factors with λ<1 and many type III_1 factors—the generic subalgebra does not admit a φ-preserving conditional expectation.","The Effros–Maréchal topology restricted to E_φ is Polish and coincides with pointwise convergence of the conditional expectations, extending the classical convergence criterion to this full space."],"fun_headline_variants":["Generic subalgebras have no state-preserving expectation","State-preserving expectations are rare for generic subalgebras","For generic subalgebras, no state-preserving conditional expectation","In many type III factors, generic subalgebras avoid state-preserving expectations"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The paper's proof that relatively amenable subalgebras are closed inherits a criterion from the literature that equates relative amenability with a weak-containment condition at the level of continuous cores; the paper does not verify the hypotheses of that criterion, so if it fails for arbitrary type III algebras, that closedness conclusion is not established.","fun_headline_variants_meta":{"raw":{"variants":["Generic subalgebras have no state-preserving expectation","State-preserving expectations are rare for generic subalgebras","For generic subalgebras, no state-preserving conditional expectation","In many type III factors, generic subalgebras avoid state-preserving expectations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000825,"raw_usage":{"total_tokens":3473,"prompt_tokens":799,"completion_tokens":2674,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":543,"completion_tokens_details":{"reasoning_tokens":2615}},"tokens_in":543,"tokens_out":2674,"duration_ms":19313,"temperature":1.0,"reasoning_tokens":2615,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T02:19:50.207017+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A single counterexample to Theorem 5.5: a sequence N_n of subalgebras in E_φ converging to N with each N_n amenable relative to P but N not amenable relative to P. Such a sequence would directly falsify the closedness claim. A second falsifier for Theorem 6.2 would be finding a type III_λ factor with λ<1 and a faithful normal state for which subalgebras lacking the expectation are not dense.","supporting_citations":[],"review_version":1}