{"id":"8ddb0d78-6bbd-4e5b-889d-c3f1142ac0db","arxiv_id":"2412.02941","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A Geroch-type obstruction is proven: connected sums with tori admit no positive scalar curvature metric with isolated conical singularities, and any nonnegative scalar curvature such metric must be flat and extend smoothly.","lead":"This mathematics paper proves that spaces formed by taking a connected sum with a torus cannot carry a metric with an isolated conical singularity and positive scalar curvature, and that a nonnegative scalar curvature version forces the metric to be flat and smooth. The result extends a classical theorem (the Geroch conjecture) from smooth manifolds to spaces with cone-like singular points, with applications to the positive mass theorem in general relativity.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Non-spherical-link rigidity rests on an unjustified local simple-connectivity assertion; the main Theorem 1.1 non-existence and sphere-link rigidity are not affected.","rationale":"The main non-existence result for Theorem 1.1 is convincingly derived by conformal blow-up, and the imported harmonic-function lemma cited from the authors' prior work is a standard Green's-function construction whose failure would indeed destroy the proof but is not, by itself, a demonstrated gap. The concrete internal flaw is the local simple-connectivity assertion in the rigidity proof: it is false for non-spherical cross-sections, and that is precisely the generality advertised in the abstract and in Theorem 1.5. For the central manifold-point case N = S^{n-1}, the assertion is true, so Theorem 1.1 itself is not affected. The reader's conditional verdict is therefore appropriate: the main claim is essentially sound, but the stronger arbitrary-cross-section rigidity needs either a revised fundamental-group argument or an explicit restriction of the rigidity claim. My agreement with the reader is partial because their listed weakest assumption was Lemma 2.4, whereas the issue I find most load-bearing is the rigidity step for non-spherical links.","tokens_in":18780,"tokens_out":36039,"duration_ms":393544,"concrete_test":"Work with a 4-dimensional conical model whose link is RP^3 with an Einstein metric, attach the conically singular regular part to a manifold containing a T^n factor, and compute the revised fundamental group in the sense of Mondino-Wei [30] for the resulting RCD(0,n) space. If this group contains n independent infinite-order elements, the flat-torus conclusion survives without local simple connectivity; if those generators are killed by the nontrivial pi_1 of the link, the rigidity statement for arbitrary cross-sections fails and must be restricted to spherical or simply connected links.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing concern is in the rigidity step, Section 5. After proving Ric_g = 0 on the regular part, the paper applies Mondino-Wei [30, Cor. 1.4] to conclude that the conical metric-measure space is a flat torus. To do so it asserts that the metric space 'clearly is a connected and locally simply connected' and therefore has the usual universal cover and usual fundamental group. This assertion is false when the conical link N is allowed to be non-spherical: arbitrarily close to the singularity there are regular points whose small neighborhoods are homeomorphic to N x (0, epsilon), so their fundamental group is pi_1(N). For a link with nontrivial pi_1, such as RP^3 in dimension 4, the space is not locally simply connected. The assertion is valid in the manifold-point case N = S^{n-1} with n >= 3, so it does not threaten the non-existence part of Theorem 1.1 or the rigidity statement for manifolds with spherical cross-sections. However, the abstract and Theorem 1.5 explicitly claim the arbitrary-cross-section case, and that stronger rigidity conclusion depends on this step. Mondino-Wei's revised fundamental group may still apply without local simple connectivity, but the manuscript gives no argument that the n independent infinite-order generators coming from the T^n factor survive in that revised group once non-simply-connected links are allowed.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves Geroch-type theorems for Riemannian metrics with isolated conical singularities on manifolds of the form M#T^n. The main non-existence claim (Theorem 1.1) is that for n>=3, with either n<=7 or M spin, there is no metric on M#T^n with finitely many isolated conical singularities whose scalar curvature on the regular part is nonnegative and strictly positive somewhere. The rigidity claim is that if such a metric has nonnegative scalar curvature and is scalar flat, then it is flat and extends smoothly across the singularities. The authors give two proofs of non-existence, both by conformally blowing up the singular point(s) to obtain a complete metric and then invoking known smooth theorems (Chodosh-Li, Wang-Zhang). The rigidity proof first shows Ricci flatness via a deformation argument using the conformal Laplacian, then applies the RCD rigidity theorem of Mondino-Wei. Theorem 1.5 extends the statement to compact manifolds with boundary whose boundary components are pinched to conical points, allowing non-spherical links.","tokens_in":19049,"tokens_out":34798,"duration_ms":359241,"significance":"If the arguments are completed, the paper would be a valuable extension of the Geroch conjecture and of Li-Mantoulidis-type rigidity to isolated conical singularities, including topological singularities with non-spherical links. The two independent conformal blow-up proofs are clearly structured, and the reduction of the rigidity problem to RCD rigidity is elegant. The paper also draws on a substantial analytic framework from the authors' prior work on conifolds, which is published separately. The main value lies in the combination of conformal analysis on conifolds with known smooth scalar-curvature obstructions and synthetic Ricci curvature rigidity. However, two load-bearing gaps currently prevent the stated theorems from following from the proofs as written.","major_comments":[{"comment":"The proof of the non-existence part of Theorem 1.1 is written for a single conical singularity: Lemma 2.4 and Lemma 4.1 are stated for one cone point, and the proofs begin with a manifold (M#T^n, g, o) with a single singularity. Theorem 1.1 and Corollary 1.3 claim the result for finitely many isolated conical singularities, but no reduction to the single-singularity case is supplied. If there are k>1 singularities and the conformal factor has a pole at only one of them, the conformally changed metric is still genuinely singular at the other k-1 points, so the resulting object is not a smooth complete manifold and the cited theorems in [5] and [39] do not apply. The authors should either prove a multi-pole version of the harmonic function or Green's function with prescribed leading asymptotics at every cone tip, or restrict the theorem to a single isolated singularity.","section":"§2 and §4 (proofs of Theorem 1.1)"},{"comment":"The paper defines 'positive scalar curvature' as Sc>=0 and strictly positive somewhere, and the proofs of non-existence produce a complete metric with nonnegative scalar curvature that is strictly positive at some point. The contradiction is then invoked with Theorem 3 of [5] and Theorem 1.1 of [39], which in the literature are standardly stated for pointwise positive scalar curvature (Sc>0 everywhere). Since the constructed scalar curvature can vanish on large sets, the cited theorems do not directly rule out the constructed metric. The authors need either to cite versions of [5,39] that apply to the weaker condition, or to add a conformal deformation argument that promotes the constructed metric to one with pointwise positive scalar curvature while preserving completeness. This is a load-bearing step in both non-existence proofs and hence also in the rigidity conclusion.","section":"§2, Eq. (2.9) and §4, Eq. (4.10)"}],"minor_comments":[{"comment":"The assertion that the metric space is locally simply connected is correct even for non-spherical links: small balls about the cone tip are truncated cones, which are contractible, and small balls about regular points are Euclidean coordinate balls. The authors may wish to add one sentence making this explicit, since the point is currently asserted rather than justified.","section":"§5"},{"comment":"The abstract's first sentence says 'an isolated conical singularity', while Theorem 1.1 says 'finitely many isolated conical singularities'; please align the terminology, especially in light of the first major comment.","section":"Abstract and Theorem 1.1"},{"comment":"Lemma 2.4 quotes Lemma 4.1 of [10] without restating its hypotheses. Since the entire first proof depends on this lemma, the authors should spell out the geometric and analytic conditions on the asymptotically flat manifold (e.g., nonnegative scalar curvature, dimension, asymptotic order) so that the reader can verify that the surgically constructed manifold satisfies them.","section":"§2, Lemma 2.4"},{"comment":"Definition 2.1 defines only a single conical singularity; the finitely many case is merely said to be analogous. For a rigorous statement of Theorem 1.1, the definition should be extended explicitly to finitely many points, including the decay condition for the perturbation h near each point.","section":"§2, Definition 2.1"},{"comment":"There is a typo 'Green's funcction' in Proposition A.1, and some reference titles contain typos (e.g., 'Compacitification' in [8]). Please proofread the bibliography.","section":"Appendix A, Proposition A.1"}],"recommendation":"major_revision","confidential_remarks":"The two gaps identified in the major comments are fixable but are not purely cosmetic: the finite-singularity gap concerns the match between the statement and the proof, and the scalar-curvature-condition gap concerns the logical bridge to the cited smooth theorems. If the authors can close both gaps, the paper is likely a solid contribution. The heavy reliance on [10] for the key harmonic function lemma is acceptable since [10] is published, but a full restatement of its hypotheses would be helpful. The local-simple-connectivity concern raised in the review does not appear to be valid, for the reasons given in the minor comments."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Three things you should know. First, the main non-existence theorem is real: for n ≤ 7 or spin, M#T^n admits no metric with isolated conical singularities and positive scalar curvature on the regular part. The proof via conformal blow-up is standard in outline but the analytic setup on conifolds — harmonic functions with the right asymptotics near the singularity and at infinity — is handled carefully. The alternative proof through the conformal Laplacian on the compact conifold is a nice addition, and the Schoen-Yau-Lohkamp compactification appendix is a useful contribution. The rigidity statement for spherical cross-sections (scalar-flat implies flat and smoothly extendable) is also new, and I see no gap there.\n\nSecond, the advertised generalization to non-spherical links has a genuine gap. In Section 5, after proving Ric = 0, the paper applies Mondino-Wei's RCD rigidity and needs the metric space to be locally simply connected. The text says this is \"clear.\" It is not clear; it is false for non-spherical links. For a link like RP^3 in dimension 4, small punctured neighborhoods have the homotopy type of RP^3, so the space is not locally simply connected. The usual fundamental group does not control Mondino-Wei's revised fundamental group, and the paper gives no argument that the T^n generators survive in that revised group. This does not affect the main Theorem 1.1 in the manifold-point case (where the link is S^{n-1}, simply connected for n ≥ 3), nor the non-existence part for arbitrary links. But Theorem 1.5 and the abstract's claim about arbitrary cross-sections rely on this step. The authors need to either supply the missing argument or weaken the rigidity claim to links that are simply connected.\n\nThird, the citation pattern looks fine. The paper relies on the authors' earlier analytic framework, but those results are independent and published separately; no circularity.\n\nWho is this for? Geometric analysts working on scalar curvature and singular spaces. The main theorem is a substantial extension of Geroch's conjecture and deserves a serious referee. I would send it out, with a clear request to fix or trim the arbitrary-link rigidity claim.","headline":"Main non-existence theorem is solid and new; the arbitrary-cross-section rigidity claim rests on a false local simple-connectivity assumption and needs repair.","tokens_in":19548,"tokens_out":3472,"would_cite":true,"duration_ms":36429,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C21","53C24","53C23","58J50"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that no conically singular metric on $M\\#\\mathbb{T}^n$ can have positive scalar curvature, and that nonnegative scalar curvature forces flatness plus smooth extension.","keywords":["scalar curvature","isolated conical singularity","Geroch conjecture","torus","conformal blow-up","rigidity","RCD spaces","asymptotically flat manifolds"],"falsifier":"A concrete counterexample would settle the matter: a metric on $\\mathbb{T}^3$ with a single isolated conical singularity and positive scalar curvature on the regular part, or on $M\\#\\mathbb{T}^n$ in any allowed dimension. A more targeted check is to solve for the harmonic function used in the blow-up on a cone with a non-spherical cross section and test numerically whether the required asymptotics hold.","tokens_in":18600,"feed_emoji":"📐","tokens_out":9312,"duration_ms":95122,"temperature":0.7,"pith_summary":"The paper proves that the classical obstruction to positive scalar curvature on tori survives when the metric is allowed to have isolated conical singularities. On any connected sum $M\\#\\mathbb{T}^n$ with $n\\ge 3$, under a mild topological hypothesis ($n\\le 7$ or $M$ spin), there is no Riemannian metric whose regular part has nonnegative scalar curvature that is positive at some point. It also proves rigidity: if such a metric has merely nonnegative scalar curvature, it must be flat and smoothly extend across each singularity. This matters because it transfers a foundational smooth-manifold theorem to singular and incomplete geometries, where positive scalar curvature questions arise in the study of weak notions of curvature and in positive mass theorems for singular spaces.","feed_headline":"Torus obstruction survives conical singularities","feed_subtitle":"For M#T^n, positive scalar curvature is impossible with conical points; nonnegative curvature forces flatness and smooth extension.","key_machinery":"The mechanism is conformal blow-up. The paper finds a positive solution $u$ of the conformal Laplace equation $-\\Delta_g u + \\frac{n-2}{4(n-1)}\\mathrm{Sc}_g u=0$ that grows like $r^{2-n}$ near a conical tip and has the standard Green's-function falloff $1+A\\rho^{2-n}$ at an asymptotically flat end. The conformally changed metric $\\tilde g = u^{4/(n-2)}g$ then satisfies $\\mathrm{Sc}_{\\tilde g}=u^{-4/(n-2)}\\mathrm{Sc}_g$, so nonnegative scalar curvature is preserved, while the change of variable $s=1/r$ turns the singular tip into a complete end modeled on $ds^2+s^2g_N$. The resulting complete metric with positive scalar curvature contradicts the known torus obstruction for complete metrics, and this contradiction is the engine of both non-existence proofs. For the rigidity half, the engine is different: deform a scalar-flat metric in the direction of its Ricci tensor, use the positivity of the first eigenvalue of the conformal Laplacian on conical manifolds to produce a positive-scalar-curvature deformation, and then invoke RCD-space fundamental-group rigidity to conclude the metric is a flat torus.","core_discovery":"The central claim is Theorem 1.1: for a closed smooth $n$-manifold $M$, $n\\ge 3$, with either $n\\le 7$ or $M$ spin, the manifold $M\\#\\mathbb{T}^n$ admits no Riemannian metric $g$ with finitely many isolated conical singularities such that $\\mathrm{Sc}_g\\ge 0$ on the regular part and $\\mathrm{Sc}_g>0$ at some point. Here 'positive scalar curvature' is used in that nonnegative-and-somewhere-positive sense. The same theorem asserts rigidity: a conically singular metric on $M\\#\\mathbb{T}^n$ with nonnegative scalar curvature must be scalar flat, then Ricci flat, then flat, and must extend smoothly across the singular points; in particular the cross sections of the singularities are forced to be round spheres even when they were not assumed spherical. A parallel statement treats compact manifolds with boundary whose boundary components are pinched to conical points, and the torus can be replaced by any $\\Lambda^2$-enlargeable closed manifold in the spin case.","pith_inferences":["The conformal blow-up strategy suggests an extension to singular sets of codimension at least 3: if a Green's-type function with the right asymptotics exists, the same contradiction should rule out positive scalar curvature for those singularities.","The rigidity argument could plausibly generalize from isolated points to conical strata of codimension at least 3; the missing ingredient would be the appropriate eigenfunction asymptotics and RCD structure for stratified spaces.","One can test sharpness by placing conical singularities on manifolds with positive Yamabe invariant rather than on $M\\#\\mathbb{T}^n$; the theorem says nothing there, and explicit examples would show that topology, not just singularity type, drives the obstruction."],"forward_implications":["For every closed manifold $M$ with $3\\le n\\le 7$, or spin $M$ of any dimension, $M\\#\\mathbb{T}^n$ carries no conically singular metric with positive scalar curvature on the regular part.","Tori and, more generally, $\\Lambda^2$-enlargeable closed manifolds admit no metric with finitely many isolated conical singularities and positive scalar curvature.","Any conically singular metric on $M\\#\\mathbb{T}^n$ with nonnegative scalar curvature is flat and extends smoothly through every singular point, with spherical cross sections forced by the rigidity conclusion.","For compact manifolds with boundary, pinching each boundary component to a conical point still forbids positive scalar curvature, and nonnegative scalar curvature forces flatness and spherical boundary components.","The compactification argument used to derive positive mass theorems from torus obstructions carries over to asymptotically flat manifolds with isolated conical singularities."],"supporting_citations":[{"why":"Supplies the existence of the harmonic function with the cone-tip and infinity asymptotics that the conformal blow-up requires.","marker":"[10]"},{"why":"Supplies the surgical connected-sum construction used to turn the compact conically singular manifold into an asymptotically flat one with nonnegative scalar curvature.","marker":"[14]"},{"why":"Gives the complete-metric non-existence theorem in dimensions at most 7 that the blown-up metric is designed to contradict.","marker":"[5]"},{"why":"Gives the same complete-metric obstruction in the spin and enlargeable case, covering arbitrary dimensions.","marker":"[39]"},{"why":"Provides the spectral properties of the conformal Laplacian on conical manifolds used in the rigidity deformation.","marker":"[11]"},{"why":"Provides the asymptotic expansion of positive eigenfunctions near conical singularities needed to keep the deformed metric conical.","marker":"[13]"},{"why":"Establishes the RCD condition for conical spaces with Ricci curvature bounded below, connecting the singular metric to synthetic Ricci bounds.","marker":"[1]"},{"why":"Supplies the fundamental-group rigidity for RCD spaces that forces the flat torus after Ricci flatness is reached.","marker":"[30]"}],"fun_headline_variants":["Conical points can't rescue torus scalar curvature","No positive scalar curvature with conical singularities","Scalar flat with conical singularities forces flatness","Torus still resists curvature with isolated cones"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that on every asymptotically flat manifold with a conical singularity of the type considered, there exists a harmonic function with exactly the growth $r^{2-n}$ near the cone tip and $1+A\\rho^{2-n}$ at infinity; if that existence statement fails for some allowed cross section, the first non-existence proof collapses.","fun_headline_variants_meta":{"raw":{"variants":["Conical points can't rescue torus scalar curvature","No positive scalar curvature with conical singularities","Scalar flat with conical singularities forces flatness","Torus still resists curvature with isolated cones"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000217,"raw_usage":{"total_tokens":1413,"prompt_tokens":899,"completion_tokens":514,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":515,"completion_tokens_details":{"reasoning_tokens":466}},"tokens_in":515,"tokens_out":514,"duration_ms":5449,"temperature":1.0,"reasoning_tokens":466,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T22:57:55.262098+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete counterexample would settle the matter: a metric on $\\mathbb{T}^3$ with a single isolated conical singularity and positive scalar curvature on the regular part, or on $M\\#\\mathbb{T}^n$ in any allowed dimension. A more targeted check is to solve for the harmonic function used in the blow-up on a cone with a non-spherical cross section and test numerically whether the required asymptotics hold.","supporting_citations":[{"cited_title":"Positive mass theo rem for asymptotically flat man- ifolds with isolated conical singularities","cited_arxiv_id":null,"evidence_quote":"Supplies the existence of the harmonic function with the cone-tip and infinity asymptotics that the conformal blow-up requires."},{"cited_title":"Blaine Lawson, Jr","cited_arxiv_id":null,"evidence_quote":"Supplies the surgical connected-sum construction used to turn the compact conically singular manifold into an asymptotically flat one with nonnegative scalar curvature."},{"cited_title":"Generalized soap bubbles and the topolo gy of manifolds with positive scalar curvature","cited_arxiv_id":null,"evidence_quote":"Gives the complete-metric non-existence theorem in dimensions at most 7 that the blown-up metric is designed to contradict."},{"cited_title":"On the generalized Geroch c onjecture for complete spin manifolds","cited_arxiv_id":null,"evidence_quote":"Gives the same complete-metric obstruction in the spin and enlargeable case, covering arbitrary dimensions."},{"cited_title":"Perelman’s λ-functional on manifolds with conical singularities","cited_arxiv_id":null,"evidence_quote":"Provides the spectral properties of the conformal Laplacian on conical manifolds used in the rigidity deformation."},{"cited_title":"Perelman’s functionals on manif olds with non-isolated conical singularities","cited_arxiv_id":null,"evidence_quote":"Provides the asymptotic expansion of positive eigenfunctions near conical singularities needed to keep the deformed metric conical."},{"cited_title":"Stratified spaces and synthetic Ricci curvature bounds","cited_arxiv_id":null,"evidence_quote":"Establishes the RCD condition for conical spaces with Ricci curvature bounded below, connecting the singular metric to synthetic Ricci bounds."},{"cited_title":"On the universal cover and th e fundamental group of an RCD∗ (K, N )-space","cited_arxiv_id":null,"evidence_quote":"Supplies the fundamental-group rigidity for RCD spaces that forces the flat torus after Ricci flatness is reached."}],"review_version":1}