{"id":"979ea04d-1727-4abd-8ca0-82cd5eb52a3a","arxiv_id":"2411.12956","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For every n≥4 there are infinitely many closed n-manifolds admitting negatively curved Einstein metrics but no locally symmetric metric.","lead":"For every dimension at least four, this paper constructs infinitely many closed manifolds that carry a negatively curved Einstein metric but no locally symmetric metric. The construction extends the four-dimensional work of Fine and Premoselli to all higher dimensions and includes a new rigidity theorem for branched covers of hyperbolic manifolds.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 4.1's uniform L2 spectral gap for the linearized Einstein operator is imported from [FP20] without verification for the long-neck sequence; Proposition 4.2 and the inverse-function step collapse if the gap degenerates.","rationale":"The reader's conditional verdict is appropriate. Lemma 4.1 is the single least-supported load-bearing input: it is cited to a dimension-four paper even though the present sequence has a neck whose length diverges and is required to satisfy uniform spectral and geometric bounds in every dimension n≥4. I found no internal contradiction in the main construction; the model metrics are explicitly built and the algebraic non-local-symmetry argument in Section 5 is substantial. The theorem's 'pairwise non-homeomorphic' claim is indeed not proved explicitly in the text, but it is fillable: since the branched cover p:X_k→M_k has degree d, the simplicial volume satisfies ||X_k||≥d||M_k||=d vol(M_k)/v_n, and the base manifolds M_k can be chosen with vol(M_k)→∞, so infinitely many distinct homotopy types follow. Thus the conditional verdict stands, subject to the spectral and geometric bounds being supplied.","tokens_in":25217,"tokens_out":28744,"duration_ms":331899,"concrete_test":"Verify the imported analytic input directly on the model neck: for the metric (2.9) with V=V_{a(d)} and the cutoff of §2.3, computed on a truncated tube u∈(u_{a(d)},U) with U=cosh(R), prove explicit bounds |sec|≤Λ, inj≥i0, and compute the bottom of the quadratic form h↦∫⟨Lh,h⟩ as R→∞. If the bottom stays ≥λ(n,d)>0 and the geometric bounds are independent of R and of the covers M_k, then Lemma 4.1 and Proposition 4.2 are justified and the existence argument is sound; if any bound degrades, the main theorem lacks support.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 4.3 rests on the uniform invertibility of L = (DΦ_{¯g_k})_{¯g_k} proved in Proposition 4.2. The key input is Lemma 4.1, which asserts a uniform L2 spectral gap λ∫|h|² ≤ ∫⟨Lh,h⟩ for all h, with λ(n,d)>0 independent of k. This lemma is not proved here; the text refers to [FP20, Proposition 4.3] as 'a bit more general.' That reference was written for the four-dimensional constructions of [FP20], and it is not demonstrated in the present paper that its hypotheses are satisfied by the sequence of approximate metrics constructed in §2.3–§3: the gluing parameter is U_glue=(U_max)^{1/2}, the neck length Rν_k →∞, and the branch-locus geometry and covering degree vary. Proposition 4.2 also uses Lemma 2.2 and Schauder estimates that require a two-sided sectional curvature bound |sec|≤Λ and injectivity radius inj≥i0 uniformly in k; Proposition 2.3 only records sec≤−c<0 and no injectivity bound. If any of these uniform geometric or spectral bounds degenerates, the a priori estimate ||h||_2≤C||Lh||_0 fails, and the inverse function theorem argument in §4.2 no longer produces the Einstein metrics that are the core of Theorem 1. This is an unverified import rather than a demonstrated contradiction, but it is the least secure load-bearing step in the paper.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs, for every n≥4 and ε>0, infinitely many pairwise non-homeomorphic closed n-manifolds X that admit a Riemannian metric with sectional curvature in [−1−ε,−1+ε], an Einstein metric with negative sectional curvature, and no locally symmetric metric. The examples are cyclic branched covers of a sequence of arithmetic hyperbolic manifolds, following Gromov–Thurston. The proof combines a Fine–Premoselli-type approximate Einstein metric, an inverse function theorem for the Einstein operator with uniform a priori estimates, and a rigidity theorem (Theorem 5.5) showing that at most one branched cover in a relevant family can be hyperbolic. The paper also contains a result on conical Einstein metrics (Remark 4.4).","tokens_in":25531,"tokens_out":12817,"duration_ms":116919,"significance":"If the analytic gaps are closed, this is a major advance: it extends the four-dimensional examples of Fine–Premoselli to all dimensions n≥4 and gives the first negatively curved Einstein metrics on non-locally-symmetric manifolds in dimensions n≥5. The algebraic construction via subgroup separability and virtual retractions is novel and carefully executed, and Theorem 5.5 is a result of independent interest. The paper is honest about its reliance on prior work, explicit about the L2 estimate, and the inverse function theorem framework is standard. However, the central analytic input, the uniform spectral gap, is imported without verification for this specific family.","major_comments":[{"comment":"The uniform L2 spectral gap for L on the approximate metrics ¯g_k is the central analytic input, but it is not proved. The text refers to [FP20, Proposition 4.3] as 'a bit more general,' yet no verification is given that the hypotheses of that result hold for the sequence here, where the gluing parameter is U_glue=(U_max)^{1/2}, the neck length R_k=R^ν_k/2 tends to infinity, and the branch locus may be disconnected. If the spectral gap degenerates as the neck grows, Proposition 4.2 and the inverse function theorem in Section 4.2 collapse, and Theorem 4.3 (hence Theorem 1) is unsupported. The authors should provide a proof of Lemma 4.1 in this setting, or at minimum a detailed verification that the constants in [FP20, Proposition 4.3] depend only on n and d and not on the neck length.","section":"§4.1, Lemma 4.1"},{"comment":"The a priori estimate uses Lemma 2.2, which requires a two-sided sectional curvature bound |sec|≤Λ and a uniform lower injectivity radius bound inj≥i0 for all k. Proposition 2.3 records only sec≤−c<0 and no injectivity bound for the approximate metrics. Since the sequence develops a long neck, the uniform lower injectivity radius and uniform upper curvature bound are not automatic, and without them the C0-estimate (4.5) is not justified. The uniform invertibility of L therefore rests on a second unverified geometric input. The authors should prove these uniform geometric bounds for the approximate metrics or modify the argument to avoid them.","section":"§4.1, Proposition 4.2"}],"minor_comments":[{"comment":"'mututally' should be 'mutually'.","section":"Abstract"},{"comment":"'Perelmann' should be 'Perelman'.","section":"Introduction"},{"comment":"The section title 'C0-.' appears truncated; it should read 'C0-estimates'.","section":"§2.2"},{"comment":"There is an inconsistency in the Hölder exponent: (2.8) states ||g^φ_{ij}||_{C^{1,α}} ≤ C, but the following sentence refers to the C^{2,α} norm.","section":"§2.2, (2.8)"},{"comment":"The definitions of the hybrid norms leave implicit the harmonic charts used to define the Hölder norms; uniformity in k of the chart size is needed and should be stated explicitly.","section":"§4.1, (4.1)-(4.2)"},{"comment":"The statement 'at most one d ∈ 4Z' should read 'at most one d ∈ 4N', since d is a positive covering degree.","section":"§5, Theorem 5.5"},{"comment":"The notation M^{2π/d_i}_cut is introduced without a precise definition; the cone structure and the path isometric boundary should be described more carefully.","section":"§5.2"},{"comment":"The infiniteness of the family X_k is not explicitly argued; it follows from the volumes of M_k (and hence X_k) tending to infinity, but this should be stated.","section":"Theorem 1 proof"}],"recommendation":"major_revision","confidential_remarks":"The paper depends heavily on [FP20] for both the approximate metric construction and the spectral gap. For a journal with high standards, the import of Lemma 4.1 from [FP20] is not sufficient without verifying its hypotheses; the authors should include a proof or a detailed verification in the appendix. The references to [HJ22] and [HJ24] are for auxiliary technical lemmas, which is acceptable if those are accepted as standard; however, the referee could not fully assess them from the manuscript alone. The topic is well within the scope of a differential geometry journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Ursula and Frieder have extended Fine-Premoselli from n=4 to all n≥4, constructing negatively curved Einstein metrics on Gromov-Thurston manifolds. The main analytic setup is a direct adaptation: build approximate Einstein metrics on branched covers, get an L2 bound on the error from a clever volume estimate, then invert the linearized Einstein operator. The genuinely new part is Section 5: Theorem 5.5, which says that for a branched cover of a hyperbolic manifold with a totally geodesic hypersurface fixed by an orientation-reversing involution, at most one degree d divisible by 4 can yield a hyperbolic manifold. That is a real result, and the proof using Mostow rigidity and fixed-point sets of isometries is intricate and, as far as I can tell, coherent.\n\nThe paper is well organized and honest about what is imported. The two soft spots are real but not fatal. First, Lemma 4.1, the uniform L2 spectral gap for the linearized Einstein operator, is cited to [FP20, Prop 4.3] rather than proved. The stress-test note worries that FP20 was written for dimension four and that the long neck in the approximate metrics could degenerate the gap. I can't rule that out from the text; the authors don't state the hypotheses of FP20's proposition or verify them for their sequence. That should be supplied. Second, the abstract promises 'mutually not homotopic' and Theorem 1 states 'pairwise non-homeomorphic', but the proof of Theorem 1 is only sketched as an immediate consequence of Theorem 4.3, Proposition 5.1, and Theorem 5.5. It is presumably true because the volumes of the covers tend to infinity, but the text never says so explicitly. A referee should ask for a sentence.\n\nThe curvature bounds needed for the Schauder estimates are also a bit loose—Proposition 2.3 gives an upper curvature bound but no lower bound, and the cited Schauder estimates require two-sided control. I suspect the lower bound follows from the explicit model, but it should be stated.\n\nOverall, this is a significant paper. The construction is not routine, and the rigidity theorem is new and interesting. The analytic argument is standard modulo the imported spectral gap. I would send it to a good referee, and I expect the gaps to be fillable.","headline":"Solid extension of Fine-Premoselli to all n≥4 with a new rigidity theorem; two fillable gaps keep me from endorsing it unconditionally.","tokens_in":26056,"tokens_out":4664,"would_cite":true,"duration_ms":46562,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C25","53C21","22E40"],"pacs":[],"model":"deepseek-v4-flash","headline":"For every n≥4 there exist infinitely many closed manifolds with negatively curved Einstein metrics that are not locally symmetric.","keywords":["Einstein metrics","negative curvature","Gromov–Thurston manifolds","branched covers","hyperbolic manifolds","locally symmetric spaces","linearized Einstein operator","arithmetic lattices"],"falsifier":"Compute the principal eigenvalue of the operator L = (1/2)Δ_L + (n−1)id on a family of explicit Gromov–Thurston manifolds with growing neck length; if it approaches zero, the uniform spectral gap in Lemma 4.1 fails and the inverse-function-theorem step collapses. Separately, exhibiting two distinct degrees d1,d2 ∈ 4Z for which the cyclic branched covers of the same hyperbolic manifold both admit hyperbolic metrics would disprove Theorem 5.5.","tokens_in":25022,"feed_emoji":"📐","tokens_out":6754,"duration_ms":59074,"temperature":0.7,"pith_summary":"The paper proves that in every dimension n≥4 there are infinitely many pairwise non-homeomorphic closed smooth manifolds that admit a negatively curved Einstein metric. The same manifolds also carry metrics with sectional curvature arbitrarily close to the constant −1, so they are Gromov–Thurston-style examples. The proof builds approximate Einstein metrics on cyclic branched covers of hyperbolic manifolds, shows their Einstein-tensor error tends to zero in L2, and then perturbs them to exact Einstein metrics using a uniform invertibility estimate for the linearized Einstein operator. A separate rigidity argument shows that at most one of the branched covers in the constructed family can be hyperbolic, which rules out locally symmetric metrics. This extends, from dimension four to all dimensions, the earlier result of Fine and Premoselli.","feed_headline":"Negative-curvature Einstein metrics in every dimension ≥4","feed_subtitle":"Infinitely many new closed manifolds in every dimension n≥4, with Einstein metrics and no locally symmetric structure.","key_machinery":"The central object is the Gromov–Thurston manifold X_k, a cyclic d-fold branched cover of a hyperbolic manifold M_k along a null-homologous totally geodesic codimension-two submanifold Σ_k. The argument is carried by three tools: (1) the Fine–Premoselli approximate Einstein metric \\bar g_k obtained by interpolating the model metric g = $du^{2}$/V(u) + V(u)$dθ^{2}$ + $u^{2}$ g_S (with V(u)=$u^{2}$−1+a(d)/$u^{{n−3}}$ chosen so the cone angle is 2π/d) with the hyperbolic metric; (2) the uniform invertibility of the linearized Einstein operator L = (1/2)Δ_L + (n−1)id acting on symmetric two-tensors, which follows from a uniform L2 spectral gap and De Giorgi–Nash–Moser estimates; and (3) a volume bound on the gluing region, obtained from subgroup separability and a geometric retraction, which makes the L2 norm of Ric(\\bar g_k)+(n−1)\\bar g_k tend to zero. The exclusion of locally symmetric metrics then uses Mostow rigidity applied to fixed point sets of deck-group isometries.","core_discovery":"Theorem 1 asserts that for any n≥4 and any ε>0 there are infinitely many pairwise non-homeomorphic closed n-manifolds X that admit a Riemannian metric with sectional curvature in [−1−ε,−1+ε], an Einstein metric with negative sectional curvature, and are not homeomorphic to any closed locally symmetric space. The examples are Gromov–Thurston manifolds: cyclic d-fold covers of closed arithmetic hyperbolic manifolds branched along null-homologous totally geodesic codimension-two submanifolds. The Einstein metric is found by showing that a sequence of approximate Einstein metrics, obtained by gluing a model Einstein metric to the hyperbolic metric, has Einstein-tensor error tending to zero in L2, and then applying a quantitative inverse function theorem to the Einstein operator in Bianchi gauge. This yields exact Einstein metrics close to the approximate ones, with negative sectional curvature. For n≥5 these are the first known negatively curved Einstein metrics on manifolds that are not locally symmetric.","pith_inferences":["A likely extension is that the same gluing-plus-perturbation scheme applies to branched covers along null-homologous codimension-two submanifolds in other non-positively curved Einstein spaces, provided analogous model metrics with the correct cone angle exist and the linearized operator has a uniform spectral gap.","If the imported spectral gap could be proved directly for the long-neck family rather than cited, the construction would become self-contained and might yield quantitative control on the Einstein metric in terms of the geometry of Σ_k.","Theorem 5.5 suggests a stronger statement, which the authors mention as forthcoming: no nontrivial branched cover of a closed hyperbolic n-manifold admits a hyperbolic metric; if true, the restriction to degree d∈4Z in the theorem is an artifact of the current proof.","One could test the construction numerically in dimension four on explicit arithmetic examples to estimate the size of the L2 error and the spectral gap; such data would indicate how large d may be for fixed M."],"forward_implications":["In every dimension n≥4 and for every ε>0, one obtains closed n-manifolds that simultaneously have (1+ε)-pinched negatively curved metrics and genuine Einstein metrics, a combination not previously available outside dimension four.","For n≥5, the Einstein metrics produced are the first known examples of negatively curved Einstein metrics on manifolds that are not locally symmetric.","In dimension four, the family constructed here is different from the Fine–Premoselli examples, and the rigidity theorem of Besson–Courtois–Gallot shows these manifolds are not homotopy equivalent to hyperbolic manifolds.","As a by-product (Remark 4.4), the hyperbolic manifolds M_k themselves carry negatively curved Einstein metrics with conical singularities of cone angle 2π/d along Σ_k.","Theorem 5.5 implies that among the cyclic branched covers with degree d∈4Z of a fixed hyperbolic manifold, at most one can be hyperbolic, so the construction yields infinitely many distinct non-homeomorphic manifolds in any dimension."],"supporting_citations":[{"why":"Introduced Gromov–Thurston manifolds as branched covers with arbitrarily pinched negative curvature and provided the indirect non-hyperbolicity argument.","marker":"[GT87]"},{"why":"Supplied the approximate Einstein metric construction for dimension four and, crucially, the uniform L2 spectral gap for the linearized Einstein operator used in Lemma 4.1.","marker":"[FP20]"},{"why":"Gives the virtual retraction and embedding results for arithmetic hyperbolic submanifolds that are used to construct the sequence of covers with good diameter-to-injectivity-radius control.","marker":"[BHW11]"},{"why":"Provides subgroup separability in arithmetic hyperbolic lattices, used to enlarge the normal injectivity radius of Σ_k while keeping it fixed.","marker":"[Ber00]"},{"why":"The volume-entropy rigidity theorem for negatively curved Einstein metrics, used to rule out locally symmetric metrics in dimension four.","marker":"[BCG95]"},{"why":"Classical spectral gap for the linearized Einstein operator on closed Einstein manifolds with negative curvature, the model for the uniform gap needed here.","marker":"[Koi78]"},{"why":"Provides the framework for the Einstein operator in Bianchi gauge and the perturbation strategy for finding Einstein metrics from approximate ones.","marker":"[And06]"},{"why":"Supplies the Schauder estimates on manifolds with a priori geometric bounds that are used in proving uniform invertibility of L.","marker":"[HJ22]"}],"fun_headline_variants":["Beyond local symmetry: Einstein metrics for n≥4","Negative curvature Einstein metrics, no local symmetry","Infinitely many Einstein manifolds, not locally symmetric","Gromov-Thurston yields Einstein metrics in every dimension","Non-symmetric Einstein metrics in all dimensions ≥4"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire perturbation step relies on a uniform L2 spectral gap for the linearized Einstein operator that is imported from Fine–Premoselli; if that gap degenerates as the approximate metrics develop longer and longer necks, the construction of exact Einstein metrics fails.","fun_headline_variants_meta":{"raw":{"variants":["Beyond local symmetry: Einstein metrics for n≥4","Negative curvature Einstein metrics, no local symmetry","Infinitely many Einstein manifolds, not locally symmetric","Gromov-Thurston yields Einstein metrics in every dimension","Non-symmetric Einstein metrics in all dimensions ≥4"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001891,"raw_usage":{"total_tokens":7318,"prompt_tokens":755,"completion_tokens":6563,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":371,"completion_tokens_details":{"reasoning_tokens":6487}},"tokens_in":371,"tokens_out":6563,"duration_ms":50709,"temperature":1.0,"reasoning_tokens":6487,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T17:02:46.542912+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the principal eigenvalue of the operator L = (1/2)Δ_L + (n−1)id on a family of explicit Gromov–Thurston manifolds with growing neck length; if it approaches zero, the uniform spectral gap in Lemma 4.1 fails and the inverse-function-theorem step collapses. Separately, exhibiting two distinct degrees d1,d2 ∈ 4Z for which the cyclic branched covers of the same hyperbolic manifold both admit hyperbolic metrics would disprove Theorem 5.5.","supporting_citations":[],"review_version":1}