{"id":"f53f78f0-17d6-4b89-ac0c-ff62bc85f255","arxiv_id":"2607.24462","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.5,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"A twisted homology class is realised by a submanifold iff its Poincaré dual is pulled back from the twisted Thom class of M^tw O(n) via a map over BO(1); the first non-realisable examples in non-orientable manifolds are given.","lead":"The paper extends Thom's classical realisability theorem to homology with twisted integer coefficients, introducing a twisted Thom space and a parametrised Pontryagin–Thom construction. It supplies the first explicit non-realisable integral homology classes in non-orientable manifolds, answering a recent open question.","discovery_kind":"extension","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"The obstruction theorem hinges on Lemma 5.4(ii): the Borel/homotopy-orbit representative of the equivariant first k-invariant must be exactly Gitler’s primary twisted operation β_tw∘P^1_3∘ρ^tw_3, not merely an operation with the same untwisted restriction.","rationale":"This is the same soft spot named by the reader, narrowed to the exact point where the obstruction becomes computable: not the Pontryagin–Thom bijection, but the identification of the homotopy-orbit k-invariant with a named twisted primary operation. The paper gives a credible mechanism—Z2-simplicity of MSO(n), equivariant Postnikov towers à la May, and Gitler’s operations—and the sign behaviour of St^5_3 is consistent with nontrivial Z2-actions on π_n and π_{n+4}. So I would not lower the verdict on present evidence. Still, because the sufficiency theorem and Examples 6.3/6.5 use the precise operation St^5_{3,tw}, the proposed Borel-spectral-sequence check is the cleanest way to confirm that no base contribution or secondary operation has been absorbed into the notation. If the check agrees, the ACCEPT verdict is strengthened; if it finds a discrepancy, the main obstruction claim would become conditional rather than geometrically wrong.","tokens_in":32116,"tokens_out":4334,"duration_ms":663674,"concrete_test":"Compute independently H^{n+5}(L(Z,n),BZ2;L_Z^n) via the Borel spectral sequence H^s(BZ2; H^t(K(Z,n);Z_sign)) for t=n,n+4,n+5, identify the generator represented by St^5_3//Z2, and compare it with the class β_tw(P^1_3(ρ^tw_3(e_n))) obtained from the local-coefficient Bockstein and Gitler operation. If they differ by a nonzero class supported on the BZ2 base or by a secondary term, Lemma 5.4(ii) and Theorem 5.5(ii) need correction.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The geometric realisability criterion (Thm. 4.11) looks solid once M^twO(n)=MSO(n)//Z2 and the twisted Thom class are accepted. The load-bearing step for Thm. 1.2/5.5 is the Postnikov identification in §5.2: from a Z2-equivariant Postnikov tower for MSO(n), the paper forms tower (5.2) over BZ2 and asserts that St^5_3//Z2 : L(Z,n)→L(Z,n+5) represents the twisted primary operation St^5_{3,tw}=β_tw◦P^1_3◦ρ^tw_3 (Lemma 5.4(ii)). The untwisted restriction to the fibre K(Z,n) is indeed Thom’s Steenrod cube, and sign/oddness arguments make equivariance plausible. But the conclusion requires more: that Gitler’s twisted P^1_3 for the sign local system is precisely the homotopy-orbit of the ordinary equivariant P^1_3, that the twisted Bockstein is the one from the exact local-coefficient sequence 0→A--3→A→A/3→0, and that the first parametrised k-invariant has no hidden base/BZ2 component. If the homotopy-orbit class differed from β_twP^1_3ρ^tw_3 by a class pulled back from H^*(BZ2;−) or by a secondary/indeterminacy term, the necessity statement might survive but the stated sufficiency range dim X≤n+5 and the computed examples could fail.","agreement_with_reader":"agree"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The paper studies when a (co)homology class with twisted integer coefficients on a manifold is realised by a submanifold, thereby extending Thom's classical theory to non-orientable ambient manifolds. The authors introduce a twisted Thom space M^tw O(n), defined as the pushout of D(γ_n) along w_1 : S(γ_n) → BO(1), identify it with the homotopy quotient MSO(n)//Z_2, and equip it with a universal twisted Thom class. They define ξ-oriented cobordism sets L_k(X; O_X ⊗ O_ξ) and prove a twisted Pontryagin–Thom bijection [X, M^tw O(n)]_{RP∞} ≅ L_k(X; O_X ⊗ O_ξ) (Theorem 4.9) via explicit geometric constructions (parametrised transversality, tubular-neighbourhood collapse maps). This yields the realisability criterion of Theorem 4.11: α ∈ H^n(X; A) is realisable iff it is pulled back from the twisted Thom class by a parametrised map over RP^∞. Using an equivariant Postnikov tower for MSO(n) and homotopy quotients, they identify the first k-invariant with the twisted Steenrod cube St^5_{3,tw} = β_tw ∘ P^1_3 ∘ ρ^tw_3 (Lemma 5.4), giving necessity of St^5_{3,tw}(α) = 0 and sufficiency when dim X ≤ n+5 (Theorem 5.5). Section 6 exhibits the first known non-realisable integral homology classes in non-orientable manifolds: 7-classes in an 11-manifold (Example 6.3) and a 10-manifold (Example 6.5).","tokens_in":32485,"tokens_out":11319,"duration_ms":383552,"significance":"If correct, this is a solid and useful contribution: it answers, with a computable obstruction, a classical question (raised by Liu) that Thom's methods cannot reach, and it provides the first concrete examples of non-realisable integral homology classes in non-orientable manifolds. Strengths deserve explicit mention: the Pontryagin–Thom bijection (Theorem 4.9) is proved by complete, geometric, parameter-free constructions (twisted Umkehr maps, push-pull formula 3.9, explicit collapse maps with continuity checks); the realisability criterion (Theorem 4.11) is a clean if-and-only-if statement; the obstruction St^5_{3,tw} is a defined, previously studied twisted cohomology operation, so the criterion is falsifiable and is actually verified nonzero in worked examples (6.3, 6.5) with careful transfer and Mayer–Vietoris computations. The framework (twisted Thom space, parametrised Postnikov tower over BZ_2) is developed with enough generality to be reusable. Nothing in the argument defines a quantity in terms of the result it predicts; the examples rely on independent input from Bohr–Hanke–Kotschick.","major_comments":[{"comment":"§5.2, Lemma 5.4(ii): the identification of the homotopy-orbit k-invariant St^5_3//Z_2 with Gitler's twisted operation β_tw ∘ P^1_3 ∘ ρ^tw_3 is the load-bearing step for Theorem 1.2/5.5, and the current proof is terse. Two points should be made explicit. (a) Existence/class of the equivariant k-invariant: state that the Z_2-equivariant Postnikov tower of MSO(n) exists by [21, Thm II.1.2] (Z_2-simplicity is checked), and that its first k-invariant must restrict to the ordinary k-invariant of the underlying tower, i.e. ±St^5_3 in H^{n+5}(K(Z,n);Z) ≅ Z/3, both signs being odd and hence equivariantly admissible — the phrase 'by the uniqueness of the first κ-invariant' should be expanded along these lines. (b) Possible base component: as a retractive map over BZ_2 the k-invariant is a class in H^{n+5}(L(Z,n), BZ_2; L^n_Z) by Proposition 2.14, and one should rule out a summand pulled back from","section":"§5.2, Lemma 5.4(ii)"}],"minor_comments":[{"comment":"Question 6.7 as stated ('a non-realisable integral homology class of dimension 3 in a non-orientable manifold of dimension 9', i.e. (n,k) = (6,3)) is already answered in the negative by the authors' own Corollary 6.1(ii): for α ∈ H^6(X^9; O_X) the obstruction St^5_{3,tw}(α) lies in H^{11}(X^9; O_X) = 0 and dim X = 9 ≤ n+5, so every such class is realisable. Presumably 'dimension 6' is intended, matching the (n,k) = (3,6) case of Remark 6.6. Please correct.","section":"§6, Question 6.7"},{"comment":"Proof of Theorem 5.5: 'P[k+4]//Z_2' should read P[n+4]//Z_2, and 'H^n(M, BZ_2; L^Th_Z)' should read H^n(M^tw O(n), BZ_2; L^Th_Z).","section":"§5.2, Theorem 5.5"},{"comment":"Lemma 2.11(iv): the fibrancy preservation of (−)//G is cited to a MathOverflow answer [13]; since this supports the fibrancy of M^tw O(n) used throughout, a published reference or a short direct proof would be preferable.","section":"§2.4, Lemma 2.11"},{"comment":"§1.1: in the definition of realisability, clarify whether the isomorphism ι^*(A ⊗ O_X) ≅ O_M is part of the data or merely required to exist; later the realisation map µ (§4.5) uses the specific isomorphism induced by the ξ-orientation φ, so a word of reconciliation would help.","section":"§1.1"},{"comment":"Remark 5.6 states a twisted analogue of Thom's sharper range (n ≥ 8, dim X ≤ n+8, realisable iff St^5_{3,tw}(α) = 0); since this is invoked implicitly in the minimality discussion of Remark 6.6, consider promoting it to a numbered corollary with a one-line proof.","section":"§5.2, Remark 5.6"},{"comment":"Figure 1 is genuinely helpful for Construction 4.7; consider also marking the section σ : T → D(E) explicitly, since it is the non-obvious ingredient in the collapse map.","section":"§4.4, Figure 1"}],"recommendation":"minor_revision","confidential_remarks":"The one place a skeptical reader will push is Lemma 5.4(ii) (identification of the parametrised k-invariant with Gitler's twisted Steenrod cube, including the absence of a base/BZ_2 component). I checked this carefully and do not believe it is a genuine gap: the relative representability of twisted cohomology (Prop. 2.14) plus the spectral sequence for K(Z,n) → L(Z,n) → BZ_2 forces the relevant group to be Z/3 detected on the fibre, so the factorwise Gitler identification suffices. The fix is expository — one paragraph — not new mathematics, hence minor rather than major revision. The manuscript is unusually transparent otherwise (explicit homotopies in Theorem 4.9, a candid generative-AI disclosure limited to discussion of one lemma, and honest open questions). The MathOverflow citations ([13] for a fibrancy fact, [19] as the problem's source) are unusual but acceptable; I have asked the authors to firm up [13]. Good fit for the journal."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The punchline is that they close the gap Thom left for non-orientable manifolds and twisted integer coefficients. They build a twisted Thom space M^tw O(n) as the push-out of the disc bundle along w1 (equivalently MSO(n)//Z2), run a geometric parametrised Pontryagin–Thom construction, and get a clean realisability criterion: α is realised iff it is the pull-back of the twisted Thom class by a map over BO(1). The first obstruction is the twisted Steenrod cube, necessary and sufficient in dim ≤ n+5, and they produce the first concrete non-realisable 7-classes in non-orientable 10- and 11-manifolds by transferring Bohr–Hanke–Kotschick examples across free involutions and a connected-sum construction.\n\nWhat is new is the classifying object itself (not the usual fibrewise Thom space), the geometric twisted cobordism sets and collapse maps, the full PT bijection (Thm 4.9), and the examples. The geometric sections (tubular neighbourhoods, Umkehr maps, transversality over the base) are careful and self-contained; the homotopy-quotient model and the low-dimensional cases n=1,2 are clean. Citations to Thom, Gitler, May, and the equivariant literature look appropriate; no circularity.\n\nThe soft spot the stress-test flags is real but proportionate: Lemma 5.4(ii) identifies the first parametrised k-invariant with Gitler’s twisted operation β_tw P^{1}_{3} ρ^{3}_tw by taking homotopy orbits of the equivariant Steenrod cube. The untwisted restriction and sign arguments check out, and this is the standard way these towers are handled, but it is not a fully expanded calculation of possible base classes or indeterminacy. Necessity of the obstruction survives either way; the sharp sufficiency range and the explicit vanishing checks in the examples rest on the identification being exact. That is a referee-level verification, not a structural hole.\n\nThis is for algebraic and geometric topologists who care about bordism, realisability, or twisted coefficients. It deserves a serious specialist referee. I would accept it for peer review and expect it to clear with at most minor expansion of the k-invariant step.","headline":"Solid extension of Thom to twisted coefficients with a clean geometric PT construction and the first explicit non-realisable integral classes in non-orientable manifolds; the k-invariant identification is the only load-bearing soft spot and looks standard rather than broken.","tokens_in":33424,"tokens_out":593,"would_cite":true,"duration_ms":13568,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["55N22","55N25","55P91","57R19"],"pacs":[],"model":"grok-4.5","headline":"A twisted homology class is realised by a submanifold exactly when its dual pulls back the twisted Thom class from a new classifying space over BO(1).","keywords":["twisted homology","local coefficients","parametrised homotopy theory","Pontryagin–Thom construction","twisted Thom space","Steenrod cube","realisability","non-orientable manifolds"],"falsifier":"Compute the twisted Steenrod cube of the dual classes constructed in the 10- and 11-dimensional examples; if either vanishes, or if an independent geometric construction realises one of those classes by an embedded submanifold, the claimed obstruction is false.","tokens_in":33182,"feed_emoji":"∫","tokens_out":1033,"duration_ms":18175,"temperature":0.7,"pith_summary":"Classical Thom theory tells when an ordinary homology class in an orientable manifold is carried by a submanifold, but the same question for twisted integer coefficients—or for ordinary integer classes inside non-orientable manifolds—has remained open. This paper answers it by building a twisted Thom space M^tw O(n) that lives over BO(1) and classifying twisted cobordism by parametrised maps into that space. The dual of a twisted class is realised if and only if it is the pull-back of the universal twisted Thom class. The first obstruction is the vanishing of a twisted Steenrod cube; when the ambient dimension is at most five more than the codimension the obstruction is also sufficient. The authors then produce the first concrete non-realisable integer homology classes in non-orientable 10- and 11-manifolds, showing that the obstruction is sharp.","feed_headline":"First non-realisable integer classes found in non-orientable manifolds","feed_subtitle":"A twisted Thom space and its Steenrod-cube obstruction decide which twisted classes are carried by submanifolds","key_machinery":"The twisted Thom space M^tw O(n) := MSO(n) ⎯ Z_2, a fibrant retractive space over BO(1) whose fibre is the ordinary oriented Thom space. Its twisted Pontryagin–Thom construction bijects parametrised homotopy classes with twisted cobordism classes, and its parametrised Postnikov tower supplies the twisted Steenrod-cube obstruction.","core_discovery":"A cohomology class α in H^n(X; A) with local integer coefficients A is realised by a submanifold if and only if there is a map f : X \to M^tw O(n) over BO(1) carrying the universal twisted Thom class to α. The first non-trivial obstruction is the vanishing of the twisted Steenrod cube St^5_{3,tw}(α); the condition is necessary in all dimensions and sufficient when dim X 一 n ≤ 5. Concrete 7-dimensional integer classes in non-orientable 10- and 11-manifolds are shown to be non-realisable by this criterion.","pith_inferences":["The same method should decide whether every twisted class admits a non-zero multiple that is realised, the twisted analogue of Thom’s classical multiple theorem.","Minimal-dimensional counter-examples are expected in dimension 9; a systematic search for free orientation-reversing involutions on known non-realisable oriented examples would settle the question.","Higher twisted k-invariants, once computed, will give complete realisability criteria in a larger stable range."],"forward_implications":["Every twisted class of codimension 1 or 2 is realised by a submanifold.","In codimension ≥ 3 a twisted class of dimension ≤ 5 is realised precisely when its twisted Steenrod cube vanishes.","There exist non-realisable integral homology classes in non-orientable manifolds of dimensions 10 and 11.","The same obstruction theory applies verbatim to ordinary integer classes inside non-orientable manifolds via the orientation local system."],"fun_headline_variants":["Twisted Thom space yields first non-realisable integer classes","Steenrod-cube obstruction decides twisted homology realisability","Non-realisable integer classes appear in non-orientable manifolds","Parametrised maps to M^twO(n) classify realisable twisted classes","Postnikov tower of twisted Thom space blocks some integer classes"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The first obstruction in the Postnikov tower of the twisted Thom space is identified with the twisted Steenrod cube by taking the homotopy quotient of an equivariant Postnikov tower of the ordinary oriented Thom space and invoking the existing theory of twisted cohomology operations; if that identification fails the obstruction calculus collapses.","fun_headline_variants_meta":{"raw":{"variants":["Twisted Thom space yields first non-realisable integer classes","Steenrod-cube obstruction decides twisted homology realisability","Non-realisable integer classes appear in non-orientable manifolds","Parametrised maps to M^twO(n) classify realisable twisted classes","Postnikov tower of twisted Thom space blocks some integer classes"]},"model":"grok-4.5","effort":"low","cost_usd":0.004893,"raw_usage":{"total_tokens":1450,"prompt_tokens":843,"num_sources_used":0,"completion_tokens":78,"cost_in_usd_ticks":48928000,"prompt_tokens_details":{"text_tokens":843,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":529,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":843,"tokens_out":78,"duration_ms":9142,"temperature":1.0,"reasoning_tokens":529,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T14:15:51.092323+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Compute the twisted Steenrod cube of the dual classes constructed in the 10- and 11-dimensional examples; if either vanishes, or if an independent geometric construction realises one of those classes by an embedded submanifold, the claimed obstruction is false.","supporting_citations":[],"review_version":1}