{"id":"a7b891bc-6340-4ed4-9a10-606c68fd358e","arxiv_id":"2604.03845","paper_version":4,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"Localisation for cohomological theories with open–closed recollement yields a torsor of supported refinements whose secondary indeterminacy is controlled by the connecting morphism, with canonicity under Gysin and Euler injectivity.","lead":"This paper builds a categorical theory of localization for cohomology with open–closed decompositions: a class vanishing on the open part yields a torsor of supported refinements, not a unique class. Algebraic geometers and K-theorists may care because it organizes when Euler-denominator formulas are canonical and tools for singularity defects.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.5","headline":"Canonicity of the supported refinement (and recovery of Euler-denominator formulae) rests on explicit Gysin hypotheses plus injectivity of Euler multiplication; these are not checkable from the abstract and may fail for some intended recollements.","rationale":"The reader correctly isolates the Gysin/Euler-injectivity package as the weakest load-bearing premise for the uniqueness half of the strongest claim. The torsor construction itself is a plausible formal consequence of any localization triangle with open–closed recollement and does not appear internally inconsistent; the risk is confined to the canonicity upgrade and the subsequent recovery of classical formulae. Because the full text (and therefore the precise hypotheses and their verification) remains unavailable, no stronger objection can be substantiated and no change of verdict is warranted. The concrete test above would settle the issue once the paper is accessible: either the Gysin package holds in the basic geometric example and the claim stands, or it fails and the uniqueness statements must be restricted.","tokens_in":1991,"tokens_out":588,"duration_ms":12592,"concrete_test":"Once the full text is available, extract the precise Gysin hypotheses (expected near the canonicity criterion) and test them on the standard open–closed pair (A^1, G_m) with closed point {0} in ordinary cohomology or Borel–Moore homology with integer coefficients: compute the localization triangle, the set of supported refinements of the zero class, and the action of Euler multiplication; if the refinement is non-unique or the Euler map is non-injective, the pre-Euler canonicity claim fails in this basic case and the recovery of Euler-denominator formulae does not hold as advertised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that a vanishing-on-open class determines only a torsor of supported refinements (with secondary indeterminacy governed by the connecting morphism) is coherent as a formal consequence of an open–closed localization triangle. The load-bearing step that turns the torsor into a preferred class—and thereby recovers the classical Euler-denominator formulae—is the pre-Euler canonicity criterion: under “explicit Gysin hypotheses,” injectivity of multiplication by the Euler class makes the refinement unique before any coefficient localization. Because the abstract supplies neither the precise statement of those Gysin maps nor a verification that they (and the injectivity) hold for the cohomological theories and geometric situations contemplated later (purity, concentration, equivariant K-theory, Milnor torsors), it is impossible to confirm that the uniqueness and recovery claims are available in the intended range. If the Gysin maps fail to be defined or Euler multiplication fails to be injective, the torsor remains non-trivial and the usual formulae are not recovered as stated.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript develops a categorical and algebro-geometric theory of localisation for cohomological theories with open–closed recollement. Its central claim is that a class whose restriction to the open complement vanishes need not determine a preferred class on the closed stratum; the localisation triangle instead associates a torsor of supported refinements, with secondary indeterminacy governed by the connecting morphism from the open complement. Compatibility with excision, base change, proper pushforward, external products and local indices is asserted, as is the statement that compatible supported constructions factor through this torsor. Under explicit Gysin hypotheses, injectivity of Euler multiplication is claimed to give a pre-Euler canonicity criterion making the refinement unique before coefficient localisation; purity, concentration and Euler rigidification are then said to recover the usual Euler-denominator formulae. Further claims relate the secondary boundary group to link transgression, treat equivariant algebraic K-theory as a multiplicative analogue, and introduce Milnor localisation torsors for characteristic-class defects of singularities.","tokens_in":2181,"tokens_out":910,"duration_ms":16000,"significance":"If the stated results hold with the claimed scope, the paper would give a precise cohomological account of when localisation produces a canonical supported class versus a residual torsor, clarifying the origin of classical Euler-denominator formulae and extending the formalism to equivariant algebraic K-theory and singularity defects. Framing supported constructions as factoring through a localisation torsor, with secondary indeterminacy controlled by the connecting morphism, is a coherent and potentially unifying contribution for theories with open–closed recollement. The programme is ambitious; its value hinges on the precision of the Gysin hypotheses and on verification that Euler multiplication is injective in the intended geometric settings.","major_comments":[{"comment":"The load-bearing step that turns the localisation torsor into a preferred supported class—and thereby recovers the classical Euler-denominator formulae—is the pre-Euler canonicity criterion: under “explicit Gysin hypotheses,” injectivity of Euler multiplication makes the refinement unique before coefficient localisation. The abstract neither states those Gysin maps nor verifies injectivity for the contemplated settings (purity, concentration, equivariant K-theory, Milnor torsors). If the maps are undefined or Euler multiplication fails to be injective, the torsor remains non-trivial and the recovery claims do not hold as stated. This step must be checked in the full text.","section":null},{"comment":"The abstract asserts a full compatibility package (excision, base change, proper pushforward, external products, local indices) and that compatible supported constructions factor through the torsor. These are central structural claims for the theory; their precise categorical hypotheses and proofs are not visible from the abstract alone and must be verified before the factoring-through-the-torsor statement can be accepted as established.","section":null},{"comment":"The “localisation torsor of supported refinements” and the “Milnor localisation torsors” for characteristic-class defects of singularities are introduced as new objects. Their existence, the group law on the torsor, the identification of secondary indeterminacy with the connecting morphism, and the relation of Milnor torsors to classical singularity invariants require explicit definitions and proofs that cannot be assessed from the abstract.","section":null}],"minor_comments":[{"comment":"The abstract is dense and introduces several technical notions (localisation torsor, secondary indeterminacy, pre-Euler canonicity, Milnor localisation torsors) without even schematic notation; a short notational roadmap in the introduction would help readers.","section":null},{"comment":"References to the “usual Euler-denominator formulae” and to link transgression are left unspecified; citing the classical sources being recovered would orient the reader.","section":null}],"recommendation":"uncertain","confidential_remarks":"Only the abstract was available for this review; the assessment is therefore provisional. The central formal claim (vanishing-on-open classes determine a torsor of supported refinements) is coherent as a consequence of an open–closed localisation triangle, but the canonicity and recovery claims rest on Gysin hypotheses that cannot be checked without the full manuscript. I recommend obtaining the complete text before a definitive editorial decision; if the Gysin package and injectivity are carefully verified in the intended range, the paper may well merit major or minor revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This is an abstract-only look at a pure theory paper in algebraic geometry / categorical cohomology. The punchline is simple: they package localisation for theories with open–closed recollement so that a class vanishing on the open does not pick a preferred supported class on the closed stratum; you get a torsor of refinements, with secondary indeterminacy controlled by the connecting morphism. Compatible constructions are said to factor through that torsor, and under Gysin hypotheses plus Euler-multiplication injectivity they claim a pre-Euler canonicity that recovers the usual Euler-denominator formulae. They also flag Milnor localisation torsors for singularity defects and a multiplicative analogue in equivariant K-theory.\n\nWhat looks new, on the abstract’s own terms, is the torsor language for supported refinements and the secondary-indeterminacy story, plus the Milnor torsors. That is a clean organising idea if the proofs land. Circularity risk is low for this kind of work: it is framed as a derivation from recollement and Gysin structure, not a fit to data. No free parameters, no invented constants.\n\nThe soft spot is exactly the one the stress-test names, and it is load-bearing: the canonicity criterion and recovery of classical formulae rest on “explicit Gysin hypotheses” and injectivity of Euler multiplication that we cannot check from the abstract. Compatibility claims (excision, base change, proper pushforward, external products, local indices) are asserted without visible proofs or examples. That is not a manufactured flaw; it is the usual abstract-only limitation, and it keeps soundness provisional. The central formal claim about the torsor is coherent as a consequence of a localisation triangle; whether uniqueness holds in the intended geometric range (purity, concentration, equivariant K-theory, singularities) is simply not visible yet.\n\nWho this is for: people who work with supported cohomology, characteristic classes, or equivariant K-theory and care about how much canonicity you actually get before coefficient localisation. It deserves a serious referee if the full text supplies the Gysin package and checks the injectivity hypotheses in the settings they advertise. I would not cite from the abstract alone, and I would not bring it to reading group until the proofs are readable. Send it to peer review rather than desk-reject; the organising claim is sharp enough to warrant referee time, with the expectation that the Gysin/Euler step will be scrutinised hard.","headline":"Abstract-only package on localisation torsors for open–closed recollement; coherent framing, but Gysin/Euler canonicity and all proofs are invisible.","tokens_in":2848,"tokens_out":596,"would_cite":false,"duration_ms":5199,"reading_group":"no","serious_thinker":"unclear","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"A vanishing class on an open complement determines a torsor of supported refinements on the closed stratum, not a preferred class.","keywords":["localisation","open-closed recollement","supported refinements","torsor","Gysin maps","Euler multiplication","algebraic geometry","cohomological theories"],"falsifier":"Exhibit a concrete open-closed pair satisfying the paper's recollement and Gysin hypotheses in which Euler multiplication fails to be injective, and check whether genuinely distinct supported refinements remain after all compatible constructions are imposed.","tokens_in":2817,"feed_emoji":"📐","tokens_out":842,"duration_ms":14174,"temperature":0.7,"pith_summary":"This paper builds a categorical and algebro-geometric theory of localisation for cohomological theories that admit open-closed recollement. Its central claim is that a class whose restriction to the open complement vanishes does not automatically select a preferred class on the closed stratum. Instead the localisation triangle produces a torsor of supported refinements, with secondary indeterminacy controlled by the connecting morphism from the open complement. Compatible supported constructions are shown to factor through this torsor, and the construction is compatible with excision, base change, proper pushforward, external products and local indices. Under explicit Gysin hypotheses, injectivity of Euler multiplication supplies a pre-Euler canonicity criterion that makes the refinement unique before any coefficient localisation; purity, concentration and Euler rigidification then recover the familiar Euler-denominator formulae. The same circle of ideas yields a relation between the secondary boundary group and link transgression, treats equivariant algebraic K-theory as a multiplicative analogue, and introduces Milnor localisation torsors that measure characteristic-class defects of singularities.","feed_headline":"Vanishing classes yield torsors of refinements, not unique classes","feed_subtitle":"Localisation triangles organise supported cohomology without preferred choices on closed strata","key_machinery":"The localisation triangle of an open-closed recollement, which replaces a single preferred supported class by a torsor of refinements whose secondary indeterminacy is controlled by the connecting morphism; under Gysin hypotheses, injectivity of Euler multiplication yields pre-Euler canonicity.","core_discovery":"A class whose restriction to the open complement vanishes need not determine a preferred class on the closed stratum; the localisation triangle associates with it instead a torsor of supported refinements, whose secondary indeterminacy is governed by the connecting morphism from the open complement, and compatible supported constructions factor through this torsor.","pith_inferences":["The torsor picture suggests that many classical Euler-denominator formulae arise from uniqueness of a refinement rather than from an a-priori choice of coefficients.","The same secondary-indeterminacy mechanism may organise support conditions in other six-functor formalisms that possess open-closed recollements.","Testing the pre-Euler canonicity criterion on concrete Gysin maps in Chow groups or motivic cohomology would delineate the geometric range of uniqueness.","Milnor localisation torsors offer a possible dictionary between singularity defects and secondary boundary data already visible in link geometry."],"forward_implications":["Compatible supported constructions factor through the torsor of refinements rather than through a single preferred class.","Under the stated Gysin hypotheses, injectivity of Euler multiplication makes the supported refinement unique before coefficient localisation.","Purity, concentration and Euler rigidification recover the classical Euler-denominator formulae as special cases.","The secondary boundary group is identified with link transgression, and equivariant algebraic K-theory appears as a multiplicative analogue.","Milnor localisation torsors organise characteristic-class defects of singularities."],"fun_headline_variants":["Vanishing open classes yield torsors of closed refinements","Localisation triangles give refinement torsors not unique classes","Open-vanishing yields supported refinement torsors via connecting maps","Supported constructions factor through localisation torsors","No preferred closed class: open vanishing produces refinement torsors"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The explicit Gysin hypotheses under which injectivity of Euler multiplication forces the supported refinement to be unique before any coefficient localisation.","fun_headline_variants_meta":{"raw":{"variants":["Vanishing open classes yield torsors of closed refinements","Localisation triangles give refinement torsors not unique classes","Open-vanishing yields supported refinement torsors via connecting maps","Supported constructions factor through localisation torsors","No preferred closed class: open vanishing produces refinement torsors"]},"model":"grok-4.5","effort":"low","cost_usd":0.005218,"raw_usage":{"total_tokens":1395,"prompt_tokens":692,"num_sources_used":0,"completion_tokens":79,"cost_in_usd_ticks":52180000,"prompt_tokens_details":{"text_tokens":692,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":624,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":692,"tokens_out":79,"duration_ms":4785,"temperature":1.0,"reasoning_tokens":624,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-13T12:13:19.382830+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Exhibit a concrete open-closed pair satisfying the paper's recollement and Gysin hypotheses in which Euler multiplication fails to be injective, and check whether genuinely distinct supported refinements remain after all compatible constructions are imposed.","supporting_citations":[],"review_version":2}