{"id":"f0b8d372-6ace-4762-a4c4-64b13403043a","arxiv_id":"1909.00529","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The equivariant higher index map is injective for bounded-geometry spaces with properly isometric, bounded-distortion group actions whose quotient and group both coarsely embed into Hilbert space.","lead":"This paper proves the equivariant coarse Novikov conjecture for discrete metric spaces when a group action has bounded distortion and both the quotient and the group embed coarsely into Hilbert space. The result gives an injectivity statement for higher index maps, which can be used to obstruct positive scalar curvature metrics on noncompact manifolds.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Bounded-distortion hypothesis in Theorem 1.2 is ill-posed for non-free actions: the defining fundamental domain Y must satisfy X = ⊔_γ γY, which forces the action to be free, while the paper's own setup and Section 4 use finite stabilizers.","rationale":"I read the paper in good faith and the central strategy is recognizable: prove injectivity of the evaluation map from the equivariant π-localization algebra, introduce twisted Roe and localization algebras using coarse embeddings of X/Γ and Γ, and apply Bott periodicity and Eilenberg swindle arguments. The reader's verdict of CONDITIONAL is appropriate. My stress-test sharpens the reader's weakest assumption: the bounded-distortion hypothesis is not merely restrictive or dependent on a choice of fundamental domain; as literally defined through the disjoint union X = ⊔γY, it requires the action to be free. The paper elsewhere engages with finite subgroups and non-free actions, so Theorem 1.2 as stated is ill-posed for a natural class of proper actions. I do not claim the proof is incorrect for free actions; rather, the theorem's hypothesis needs either an explicit freeness assumption or a generalized fundamental-domain definition with finite stabilizers, and the proof of Theorem 4.9 would need corresponding adjustment. This does not overturn the reader's CONDITIONAL verdict, so I recommend UNCHANGED.","tokens_in":23996,"tokens_out":24055,"duration_ms":501034,"concrete_test":"Take Γ = Z/2 acting on X = Z by reflection, x ↦ -x, with the usual metric. This action is proper and isometric, X/Γ is a half-line with bounded geometry, and Γ embeds into Hilbert space. Check whether any subset Y ⊆ Z satisfies Z = Y ⊔ (-Y). The fixed point 0 makes this impossible, so Theorem 1.2's bounded-distortion hypothesis cannot be satisfied despite the geometric hypotheses holding. Then attempt to re-run the key step of Theorem 4.9 using a Borel fundamental domain such as Y = {0} ∪ {positive integers}, tracking the finite stabilizer F_i = {1} (or the stabilizer of 0). If the uniform boundedness of ~F_i · Δ'_j(R) still holds, the paper needs an explicit generalized fundamental-domain definition; if it fails, the theorem must be restricted to free actions.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 1.2 assumes only a proper, isometric action with bounded distortion, where bounded distortion is defined in Section 1 using a fundamental domain Y with X = ⊔_{γ∈Γ} γY. Such a strict disjoint-union fundamental domain exists only if the Γ-action is free. The paper elsewhere explicitly allows non-free actions: Definition 2.3 concerns finite subgroups, Proposition 2.6 covers cocompact actions without freeness, and Theorem 4.9 repeatedly uses finite subgroups F_i and spaces Γ ×_{F_i} S_i. For a non-free proper action, no Y as defined exists, so the hypothesis 'with bounded distortion' is undefined and the theorem silently excludes all actions with nontrivial isotropy. This is not merely cosmetic: the proof of Theorem 4.9 chooses Y and uses representatives x_j ∈ Y to prove that ~F_i · Δ'_j(R) is uniformly bounded and that the simplices Δ_j(S) are F_i-invariant. If one tries to replace Y by a Borel fundamental domain accommodating finite stabilizers, the argument needs additional estimates that are not present. The statement also does not quantify whether bounded distortion is required for 'some' or 'all' fundamental domains; the definition depends on Y, and no independence or compatibility discussion is given. For applications such as Corollary 1.3, where the acting group may have finite isotropy, this is a genuine scope gap in the central theorem.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves an equivariant coarse Novikov conjecture: for a countable discrete group Γ acting properly, isometrically, and with bounded distortion on a bounded-geometry discrete metric space X, if both X/Γ and Γ coarsely embed into Hilbert space, then the equivariant higher index map Ind^Γ from lim_d K^Γ_*(P_d(X)) to K_*(C^*(X)^Γ) is injective (Theorem 1.2). The proof uses equivariant localization algebras, π-localization algebras defined via the quotient coarse embedding, twisted equivariant Roe algebras using both the quotient and group coarse embeddings, Bott periodicity, Mayer-Vietoris sequences, and Eilenberg swindle arguments. The main theorem is applied to show nonvanishing of equivariant higher indices of Dirac operators and hence obstructions to Γ-invariant metrics of uniformly positive scalar curvature (Corollary 1.3).","tokens_in":24256,"tokens_out":5561,"duration_ms":47203,"significance":"If the result is correct, it is a substantial equivariant extension of Yu's coarse Baum-Connes theorem, covering a broad class of non-cocompact actions under natural coarse embeddability hypotheses. The paper relies on established machinery and on the authors' earlier work [7], and it gives a clearly stated injectivity criterion that is potentially algorithmic for nonvanishing of equivariant higher indices. The architecture of the proof is coherent, and the use of twisted equivariant localization algebras is well motivated. However, the central theorem's hypothesis and several load-bearing proof steps need careful revision before the result can be accepted.","major_comments":[{"comment":"The definition of a fundamental domain as a subset Y with X = ⊔_{γ∈Γ} γY forces the Γ-action to be free: any point with nontrivial stabilizer would lie in both eY and γY for a nontrivial γ in its stabilizer. Yet the paper explicitly works with non-free actions elsewhere: Definition 2.3 concerns finite subgroups F, Proposition 2.6 covers cocompact actions without a freeness assumption, the proof of Theorem 4.9 uses spaces Γ ×_{F_i} S_i for finite subgroups F_i, and the Bott map in Section 5 averages over stabilizers Γ_x. Consequently the hypothesis \"with bounded distortion\" in Theorem 1.2 is undefined for every action with nontrivial isotropy, which is a genuine scope gap in the central theorem. In addition, the definition depends on a choice of Y, and the paper does not state whether bounded distortion is required for some or for all fundamental domains, nor prove independence of the choice. Please either restrict Theorem 1.2 (and Corollary 1.3) to free actions or reformulate bounded distortion using a Borel or measurable fundamental domain and supply the additional estimates needed in the proof of Theorem 4.9.","section":"Section 1, definition of fundamental domain and Theorem 1.2"},{"comment":"The reduction from C*_L(P_d(X), A(H,ξ)_{O_{r,k}} ⊗ (A_n)_{Γ×_{F_i}S_i})^Γ to the product over C*_L(\\tilde F_i · Δ'_j(R), ...)^{F_i} is asserted with the phrases \"it is not very difficult to prove\" and \"in a way similar to the proof of Lemma 5.16 in [30]\", after the statement that A_n is Γ-proper over Γ ×_{F_i} S_i. These identifications are load-bearing because they are what allow the Eilenberg swindle on the simplices Δ_j(S). The subsequent claim that the bounded-distortion hypothesis makes \\tilde F_i · Δ'_j(R) uniformly bounded for j ∈ J_k is also stated as \"easy to prove\"; this estimate is exactly where the new bounded-distortion condition enters, and it must be shown explicitly. Please expand this part of the proof with full details.","section":"Section 4, proof of Theorem 4.9"},{"comment":"The proof of Theorem 5.1 consists of the assertion that the Mayer-Vietoris sequence, the five lemma, induction on skeleta of P_d(X), and Bott periodicity in the 0-dimensional case suffice. This is too compressed for a step on which Theorem 1.2 directly depends: the Mayer-Vietoris/five-lemma argument requires showing that the Bott map (β_L)_* is compatible with the six-term exact sequences for the localization algebras, and the 0-dimensional equivariant case, including the averaging over the stabilizer 1/|Γ_x| in the definition of β_t, is not carried out. Please provide the details or a precise reference for this equivariant Bott isomorphism.","section":"Theorem 5.1"}],"minor_comments":[{"comment":"The notation is inconsistent: the algebra of functions is first called C*_{π,L}(X)^Γ and then the same symbol is reused for its norm closure; presumably the first object should be C*_{π,L,alg}(X)^Γ.","section":"Section 3, Definition 3.1"},{"comment":"There are repeated grammatical slips, such as \"an coarse embedding\" (Propositions 3.12, 3.13 and Theorem 4.9), \"bounded geometrical metric space\" (proofs of Proposition 3.12 and Theorem 4.9), \"distorsion\" (proof of Theorem 4.9), \"Lipchitz\" (Definition 3.3), and \"Let (C0(X), Γ,φ) is an admissible covariant system\" (Proposition 2.6). These should be corrected.","section":"Throughout"},{"comment":"The sentence \"There is no difficulty to check\" (and later \"it is no difficult to see\") is a stylistic slip; it should read \"There is no difficulty in checking\" or \"It is not difficult to see\".","section":"Section 2, after Definition 2.7"},{"comment":"The definition of A_n as a direct limit of ideals that are Γ-proper over cocompact Γ-spaces is not fully explained; in particular, the notion of \"Γ-proper over W\" should be defined or referenced precisely, since it is used in the reduction to Γ ×_{F_i} S_i.","section":"Section 4, proof of Theorem 4.9"},{"comment":"The notation Y_d is introduced as \"a fundamental domain of X_d\" and then used with x ∈ Y_d and y ∈ X_d; this inherits the free-action issue from Section 1 and should be reconciled with the stabilizer-averaging formula 1/|Γ_x| that appears in the same definition.","section":"Section 5, definition of β_t"}],"recommendation":"major_revision","confidential_remarks":"The paper is potentially publishable after substantial revision. The main concern is not the overall strategy but the ill-posedness of the bounded-distortion hypothesis for non-free actions, which conflicts with the paper's own use of finite stabilizers, and the compressed proofs of Theorems 4.9 and 5.1. I would ask the authors to reconcile the free-action requirement of the fundamental-domain definition with the non-free cases their framework treats, and to expand the omitted details before resubmission."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a real extension of the equivariant coarse Novikov conjecture, but the bounded-distortion hypothesis as written is not well-defined for non-free actions, so Theorem 1.2 is narrower than it claims. The paper deserves a serious referee, but the authors need to fix that hypothesis.\n\nWhat is actually new: previous results covered the trivial-group case (Yu) and X=Γ (Skandalis–Tu–Yu). Here they prove injectivity of the equivariant higher index map under separate coarse embeddability of X/Γ and Γ with bounded distortion. The application to invariant positive scalar curvature is a genuine payoff.\n\nI read the proof as following the standard Yu localization-algebra route: build the equivariant π-localization algebra, inject the evaluation map using the quotient embedding, then compare twisted localization algebras using the group embedding. The architecture is coherent and the citations are honest. The Bott map and the K-theory identifications follow the expected pattern.\n\nThe soft spots. The stress-test note is right: Section 1 defines a fundamental domain as X = ⊔_γ γY, which only exists if Γ acts freely. But the paper explicitly allows non-free actions—Definition 2.3 uses finite subgroups, Proposition 2.6 covers cocompact actions without freeness, and Theorem 4.9 uses spaces Γ ×_{F_i} S_i for finite subgroups. So for any proper action with nontrivial isotropy, the bounded-distortion hypothesis has no meaning. This is not cosmetic: the proof of Theorem 4.9 uses the fundamental domain Y and representatives x_j ∈ Y to show the sets ~F_i · Δ'_j(R) are uniformly bounded and that the simplices are F_i-invariant. Without a free action there is no such Y, and the argument does not go through. A Borel fundamental domain accommodating finite stabilizers would need new estimates that are not present. The statement also never says whether bounded distortion must hold for some or all fundamental domains.\n\nThe other soft spot is compression: Theorem 5.1's Bott isomorphism is asserted after a short paragraph, and Theorem 4.9 contains 'it is not very difficult to prove' for identifications that are load-bearing. Those are probably fillable from [39] and [7], but a referee should ask for details.\n\nBottom line: the central idea is right for free actions, and the flawed hypothesis is fixable. The paper is worth refereeing. A careful referee should push the authors to either add a freeness assumption or reformulate bounded distortion for non-free actions, and to expand the compressed proofs.","headline":"A real extension of the equivariant coarse Novikov conjecture, but the bounded-distortion hypothesis is not well-defined for non-free actions, so the main theorem is narrower than stated.","tokens_in":24816,"tokens_out":2458,"would_cite":false,"duration_ms":22551,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["19K56","58J22","46L80"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that the equivariant coarse Novikov conjecture—injectivity of the equivariant higher index map—holds for a bounded-geometry space when both its group-quotient and its symmetry group coarsely embed into Hilbert space.","keywords":["equivariant coarse Novikov conjecture","higher index","equivariant Roe algebra","localization algebra","coarse embedding into Hilbert space","bounded distortion","Rips complex","K-theory"],"falsifier":"Take a concrete non-cocompact case covered by the theorem, such as Γ=Z acting by translation on X=Z×Z with an appropriate bounded-geometry metric for which X/Γ≅Z and Γ both coarsely embed into Hilbert space and bounded distortion holds with respect to a fundamental domain, then compute the direct-limit equivariant higher index map directly: any nonzero class in lim_d K_*^Z(P_d(X)) must have nonzero image in K_*(C^*(X)^Z). Finding such a class with vanishing image would refute Theorem 1.2; verifying that no such class exists in a family of these actions would support it.","tokens_in":23770,"feed_emoji":"🎯","tokens_out":8269,"duration_ms":77559,"temperature":0.7,"pith_summary":"The paper aims to prove the equivariant coarse Novikov conjecture for a discrete metric space X with bounded geometry whenever a countable group Γ acts on X properly, isometrically, and with bounded distortion, and both the quotient X/Γ and Γ coarsely embed into a Hilbert space. The conjecture says that the equivariant higher index map from equivariant K-homology of Rips complexes to the K-theory of the equivariant Roe algebra is injective. Injectivity matters because it turns a computable K-theory class into a reliable nonvanishing obstruction: if the class of an elliptic operator is nonzero, the operator cannot be deformed away, and for spin manifolds this obstructs invariant metrics of uniformly positive scalar curvature. The proof works by reduction to localization algebras, then compares ordinary and projected localization through the two coarse embeddings, using an Eilenberg swindle step that is where bounded distortion is needed.","feed_headline":"Two coarse embeddings make the equivariant higher index map injective","feed_subtitle":"For bounded-distortion group actions, injectivity turns the higher index into a computable nonvanishing obstruction.","key_machinery":"The load-bearing machinery is the equivariant π-localization algebra C^*_{π,L}(P_d(X))^Γ, the algebra of uniformly continuous paths into the equivariant Roe algebra whose propagation, measured after quotient projection π, decays to zero, together with its twisted counterparts C^*(P_d(X), A(H,ξ)⊗A(H,η))^Γ built from the coarse embeddings of X/Γ and Γ. The quotient embedding ξ controls the decay of projected propagation and supplies the twisted Roe algebra through which the evaluation map becomes an isomorphism (Proposition 3.12); the group embedding η produces a Γ-proper C*-algebra over a continuous field of Hilbert spaces, and the bounded-distortion hypothesis guarantees that certain orbit-translated simplex families used in the Eilenberg swindle stay uniformly bounded and invariant under finite subgroups. The final step is the periodicity map from C(Z)-coefficient localization algebras to A(H)-twisted ones.","core_discovery":"In the paper's own terms, the central discovery is Theorem 1.2: under the stated hypotheses, Ind^Γ: lim_{d→∞} K_*^Γ(P_d(X)) → K_*(C^*(X)^Γ) is injective. The proof first replaces the equivariant higher index map with the evaluation map from an equivariant localization algebra to the equivariant Roe algebra, whose injectivity is equivalent (Theorem 2.11). It then introduces an equivariant π-localization algebra where propagation is measured after projection to X/Γ; the coarse embedding of X/Γ makes the evaluation map from this algebra injective (Theorem 3.13). The coarse embedding of Γ enters through a proper affine action on a continuous field of Hilbert spaces, giving a twisted algebra A(H,η); Theorem 4.9 shows that with bounded distortion the localization and π-localization versions of the twice-twisted algebra have isomorphic K-theory, Theorem 4.10 yields injectivity of the comparison map, and the periodicity isomorphism (Theorem 5.1) completes the diagram. Bounded distortion appears in Theorem 4.9, where it ensures the sets Γ·Δ'_j(R) are uniformly bounded and the simplices Δ_j(S) are F_i-invariant, which makes the Eilenberg swindle work.","pith_inferences":["The paper's reliance on bounded distortion is tied to a chosen fundamental domain, and the condition is not shown to be invariant under that choice; one could test whether a weaker, choice-independent condition—such as uniform displacement on orbits after passing to a metric quotient—would suffice for the same Eilenberg swindle.","The structure suggests a template: if a quotient's coarse geometry controls propagation and the symmetry group's coarse embedding supplies enough Bott-type twisting, injectivity of the equivariant index may persist in other targets than Hilbert space, such as uniformly convex Banach spaces, wherever an analogue of the periodicity isomorphism holds.","One might try replacing bounded distortion by a bound on displacement only along a generating set of Γ; if the Eilenberg swindle still closes, the theorem would cover a larger class of non-cocompact actions."],"forward_implications":["For a complete spin manifold M with bounded geometry and a proper, isometric, bounded-distortion action of Γ on an invariant ε-net X, a nonzero class φ_*([D]) in lim_d K_*^Γ(P_d(X)) rules out any Γ-invariant Riemannian metric of uniformly positive scalar curvature (Corollary 1.3).","When Γ is trivial, Theorem 1.2 reduces to the classical coarse Novikov conjecture for spaces coarsely embeddable into Hilbert space; when X=Γ, it recovers the known result for groups that coarsely embed into Hilbert space.","The injectivity result makes the equivariant higher index an algorithmically usable invariant, because the domain lim_d K_*^Γ(P_d(X)) and the push-forward class are computable in the stated setting.","The theorem extends the reach of the equivariant coarse Novikov conjecture from cocompact actions, where the equivariant Roe algebra is Morita equivalent to the reduced group C*-algebra, to many non-cocompact actions as long as the quotient is coarsely embeddable."],"supporting_citations":[{"why":"Supplies the twisted Roe algebra and localization algebra framework, the asymptotic morphisms, and the Eilenberg swindle argument that the proof adapts.","marker":"[39]"},{"why":"Establishes the equivalence between coarse embeddability of Γ and existence of a proper affine action on a continuous field of Hilbert spaces, used to build A(H,η).","marker":"[32]"},{"why":"Provides the periodicity-style isomorphism for the C*-algebras attached to finite-dimensional affine subspaces of Hilbert space.","marker":"[19]"},{"why":"Supplies the descent technique used to relate equivariant index theory to group C*-algebras.","marker":"[17]"},{"why":"Gives the equivariant coarse Baum-Connes machinery and asymptotic morphisms used in the localization-algebra comparison.","marker":"[7]"},{"why":"Constructs twisted Roe algebras for coarse embeddings and proves the non-equivariant evaluation-map isomorphism that the proof generalizes.","marker":"[34]"},{"why":"Provides the equivariant Roe algebra and the Morita equivalence to the reduced group C*-algebra in the cocompact case, framing the conjecture.","marker":"[27]"},{"why":"Supplies admissibility and stabilization results for equivariant K-homology cycles, needed to define the equivariant index map.","marker":"[23]"}],"fun_headline_variants":["Coarse embeddings force injective equivariant higher index","Bounded-distortion actions with coarse embeddings give index injectivity","Equivariant Novikov conjecture holds under two coarse embeddings","Two coarse embeddings make higher index injective for group actions","Coarse embeddability yields injective equivariant higher index"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof collapses if the action does not have bounded distortion: for some group element γ, the displacement sup_{x∈Y} d(γx,x) over a chosen fundamental domain Y must be finite, and this finiteness is used to keep the orbit-translated simplices bounded enough for the Eilenberg swindle to run.","fun_headline_variants_meta":{"raw":{"variants":["Coarse embeddings force injective equivariant higher index","Bounded-distortion actions with coarse embeddings give index injectivity","Equivariant Novikov conjecture holds under two coarse embeddings","Two coarse embeddings make higher index injective for group actions","Coarse embeddability yields injective equivariant higher index"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000165,"raw_usage":{"total_tokens":1250,"prompt_tokens":942,"completion_tokens":308,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":558,"completion_tokens_details":{"reasoning_tokens":226}},"tokens_in":558,"tokens_out":308,"duration_ms":3254,"temperature":1.0,"reasoning_tokens":226,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:45:41.941590+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a concrete non-cocompact case covered by the theorem, such as Γ=Z acting by translation on X=Z×Z with an appropriate bounded-geometry metric for which X/Γ≅Z and Γ both coarsely embed into Hilbert space and bounded distortion holds with respect to a fundamental domain, then compute the direct-limit equivariant higher index map directly: any nonzero class in lim_d K_*^Z(P_d(X)) must have nonzero image in K_*(C^*(X)^Z). Finding such a class with vanishing image would refute Theorem 1.2; verifying that no such class exists in a family of these actions would support it.","supporting_citations":[{"cited_title":"Yu, The coarse Baum-Connes conjecture for spaces which admit a u niform embedding into Hilbert space , Invent","cited_arxiv_id":null,"evidence_quote":"Supplies the twisted Roe algebra and localization algebra framework, the asymptotic morphisms, and the Eilenberg swindle argument that the proof adapts."},{"cited_title":"Skandalis, J","cited_arxiv_id":null,"evidence_quote":"Establishes the equivalence between coarse embeddability of Γ and existence of a proper affine action on a continuous field of Hilbert spaces, used to build A(H,η)."},{"cited_title":"Higson, G","cited_arxiv_id":null,"evidence_quote":"Provides the periodicity-style isomorphism for the C*-algebras attached to finite-dimensional affine subspaces of Hilbert space."},{"cited_title":"Higson, Bivariant K-theory and the Novikov conjecture , Geom","cited_arxiv_id":null,"evidence_quote":"Supplies the descent technique used to relate equivariant index theory to group C*-algebras."},{"cited_title":"Fu and X","cited_arxiv_id":null,"evidence_quote":"Gives the equivariant coarse Baum-Connes machinery and asymptotic morphisms used in the localization-algebra comparison."},{"cited_title":"Shan and Q","cited_arxiv_id":null,"evidence_quote":"Constructs twisted Roe algebras for coarse embeddings and proves the non-equivariant evaluation-map isomorphism that the proof generalizes."},{"cited_title":"Roe, Index theory, coarse geometry and topology of manifolds , CBMS Regional Conference Series in Mathematics, American Mathematica l Society, no","cited_arxiv_id":null,"evidence_quote":"Provides the equivariant Roe algebra and the Morita equivalence to the reduced group C*-algebra in the cocompact case, framing the conjecture."},{"cited_title":"Kasparov and G","cited_arxiv_id":null,"evidence_quote":"Supplies admissibility and stabilization results for equivariant K-homology cycles, needed to define the equivariant index map."}],"review_version":1}