{"id":"255cf220-0fd0-4106-87c1-45c4e078c753","arxiv_id":"2412.21008","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The first Steklov eigenvalue of a Riemannian manifold with boundary is, up to constant factors, equal to the isocapacitary constant Γ∂, and the same holds for the bottom of the Dirichlet-to-Neumann spectrum in the non-compact case.","lead":"Mathematicians prove two-sided bounds on the first Steklov eigenvalue, a number describing how a shape responds to boundary vibrations, using a capacitary Cheeger constant. The bounds cover compact and non-compact manifolds, and are applied to hyperbolic surfaces and a hyperbolic half-ball.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lower-bound proof of Theorem 1 is not valid as written: Proposition 10 is false as stated, and the subsequent level-set step replaces ∫f_+² by ∫f².","rationale":"The reader correctly identified Proposition 10 as a false statement, but the problem is more extensive: the same lower-bound proof of Theorem 1 contains a second independent error in the level-set calculation, and this error is not merely typographical—it changes the numerical constant in the model case. The upper-bound parts of the theorems appear coherent, and the central qualitative claim may well be salvageable by correcting Proposition 10 to the d(t²) form and adding a proper treatment of the sign-changing eigenfunction (or by weakening the lower constant). However, as submitted, the displayed derivation of 1/4Γ∂ ≤ σ1 is invalid. Because the manuscript already received a CONDITIONAL verdict and these are concrete but repairable defects, I do not move the verdict label; I strengthen the conditions under which the paper should be accepted.","tokens_in":17727,"tokens_out":30793,"duration_ms":313192,"concrete_test":"Run the lower-bound chain on M=[0,1]. First, with u=x, verify Cap({x≥t},{x≤0},[0,1])=1/t: the printed Proposition 10 would require ∫_0^1 dt/t = ∞ ≤ 4, which fails, while the corrected d(t²) form gives ∫_0^1 t·(1/t)dt = 1 ≤ 2. Second, with the first Steklov eigenfunction f=2x-1 (so Γ∂=1 and σ1=2), compute ∫_0^1 t·m({f≥t})dt = 1/2, whereas the paper's 1/2∫_{∂M}f²dσ = 1; the equality used in the proof is therefore false by a factor of 2.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The lower bound in Theorem 1 rests on Proposition 10, but the printed statement is false. For M=[0,1] and u(x)=x, Cap({x≥t},{x≤0},[0,1])=1/t, so the left-hand side ∫_0^1 Cap dt diverges, while the right-hand side is 4∫_0^1|∇u|²=4. The proof actually integrates against d(t²) in line (14); the correct statement is ∫_0^∞ Cap d(t²) ≤ 4∫|∇u|², equivalently ∫_0^∞ t·Cap dt ≤ 2∫|∇u|². This correction alone does not save the theorem's proof. In the application to the first Steklov eigenfunction f, the paper uses ∫_0^∞ t·m({f≥t})dt = 1/2∫_{∂M} f² dσ. The left-hand side equals 1/2∫_{∂M}(f_+)² dσ, and f necessarily changes sign. No argument is given to control ∫f_+² in terms of ∫f². On the unit disk with f=cosθ, the left-hand side is π/4 while the paper's right-hand side is π/2, a factor of 2 discrepancy. Thus the claimed 1/4Γ∂ lower bound does not follow from the displayed chain; a separate argument handling f_+ and f_-, or a weaker constant, is needed. Theorem 2 inherits this defect through (23).","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper defines isocapacitary constants Γ∂(M) for a Riemannian manifold with boundary and claims Cheeger-type two-sided bounds: for compact M, 1/4 Γ∂(M) ≤ σ1(M) ≤ 2Γ∂(M), and for non-compact M, 1/4 Γ∂(M) ≤ inf Spec(DM) ≤ 2Γ∂(M). The proofs proceed through a capacity lemma (Proposition 10), a variational argument for the first Steklov eigenvalue, a limit argument for the non-compact case, and applications to compact hyperbolic surfaces with geodesic boundaries and to the hyperbolic half-ball. The manuscript is self-contained and the constants are computed, not fitted; the graph paper [20] is used only as an analogue, not as an input.","tokens_in":18001,"tokens_out":9075,"duration_ms":94459,"significance":"If the main theorems are correct, the paper gives a uniform control of the first Steklov eigenvalue (or the bottom of the Dirichlet-to-Neumann spectrum) by a purely geometric isocapacitary quantity, up to a universal factor of 8. This would partially answer the Buser-type question from Colbois–Girouard–Gordon–Sher and extend Maz'ya's isocapacitary framework to the Steklov setting. The applications to hyperbolic surfaces produce explicit upper bounds and eigenvalue decay statements that are concrete and checkable. The paper is readable and the overall strategy is credible, but the central lemma and one key application step need correction before the claims are established.","major_comments":[{"comment":"Proposition 10 as printed is false. For M=[0,1] and u(x)=x, the set {u≥t} is [t,1] and {u≤0} is {0}; the capacity of these two sets is 1/t, so the left side ∫_0^1 Cap dt diverges while the right side is 4∫_0^1 |u'|^2 dx = 4. The proof in Eq. (14) integrates with respect to d(t(ψ)^2), which indicates that the intended statement is ∫_0^∞ t·Cap({u≥t},{u≤0},M) dt ≤ 2∫_{M_u≥0} |∇u|^2, equivalently ∫_0^∞ Cap d(t^2) ≤ 4∫|∇u|^2. This correction is local and repairable, but the lemma as stated is load-bearing for Eq. (19) and must be fixed.","section":"§3, Proposition 10"},{"comment":"The lower-bound chain for a sign-changing Steklov eigenfunction is incomplete. With the corrected form of Proposition 10, the displayed argument yields σ1∫ f^2 ≥ 1/2 ∫_0^∞ t·Cap(...) dt ≥ 1/2 Γ∂(M) ∫_0^∞ t·m({f≥t}) dt = 1/2 Γ∂(M) · (1/2 ∫_{∂M} (f_+)^2 dσ), not 1/4 Γ∂(M)∫_{∂M} f^2 dσ. The identity used in the paper identifies ∫ t·m({f≥t}) dt with 1/2∫ f^2, but the correct identity is 1/2∫ (f_+)^2 because f changes sign. Assumption (18) controls the measures of the positive and negative sets but gives no control of ∫(f_+)^2 relative to ∫ f^2. Thus the claimed lower bound 1/4 Γ∂(M) does not follow without an additional argument handling f_+ and f_-, or a weaker constant.","section":"§4, proof of Theorem 1, Eq. (19)–(20)"},{"comment":"The non-compact part silently assumes properties that are not stated. The construction of Hf via the limit over M∩B(p,r), the monotonicity of ξ1(Mr, ∂I Mr), and the equality Γ∂(M)=lim Γ∂(Mr,M) in Lemma 12 require that the geodesic balls B(p,r) have relatively compact intersection with M, which is automatic for complete Riemannian manifolds but not for arbitrary non-compact manifolds with embedded boundary. The completeness hypothesis should be added, or the statements should be restricted to a class of manifolds for which the exhaustion is valid.","section":"§2.2 and §4, Theorem 2 and Lemma 12"},{"comment":"The definition of Cap(F,M) in the proof of Theorem 2 is misprinted: it reads 'f is taken over all smooth functions such that f ≡ 0 on F and f ∈ C_c^∞(M)', but the correct condition, as given in the introduction, is f ≡ 1 on F. With the printed condition, Cap(F,M)=0 for every F and Γ∂(M) becomes identically zero, which would make Theorem 2 vacuous. This is an obvious typo, but it affects the meaning of Lemma 12 and Theorem 2 and should be corrected.","section":"§4, definition of Cap(F,M)"}],"minor_comments":[{"comment":"There are numerous typos, e.g., 'Huasdorff' for Hausdorff, 'Togther' for Together, 'exsit' for exist, and several instances of inconsistent spacing around equations. A careful proofreading pass is needed.","section":"Throughout"},{"comment":"The line defining f^- is misprinted as 'f − = 1/2 = (f − |f|)' and should read 'f_- = 1/2(f - |f|)'.","section":"§2.2, Eq. after f_+ and f_-"},{"comment":"The notation for the DtN operator is inconsistent: D^{n+1}_+ appears as D^{n+1}_+, DD^{n+1}_+, and DD^n_+ in different places. Please unify the notation.","section":"§5.2, Theorem 14 and surrounding text"},{"comment":"The statement 'σ1(Sg,n) ≤ e + e^{-1}' for l0 < 1 < ρ0 should specify how ρ0 is chosen in terms of l0; currently the inequality depends on the choice of ρ0 via the collar lemma, and the text should make the dependence explicit.","section":"§5.1, Case II"}],"recommendation":"major_revision","confidential_remarks":"This is a promising manuscript with a credible strategy, and the capacity-constant framework is not circular. However, the printed Proposition 10 is false, the application to sign-changing Steklov eigenfunctions has a real gap, and the non-compact statements omit a completeness hypothesis. These are local and fixable rather than fatal, so I recommend major revision rather than rejection. Please ask the authors to correct Proposition 10, rework the lower-bound proof for the first Steklov eigenvalue (or state a weaker constant), and add the missing completeness assumption."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nQuick take: the headline results are plausible and the isocapacitary idea transfers to Steklov naturally, but the lower-bound proof has two concrete errors, one typo-level and one load-bearing. The theorem may be true, but the printed proof does not establish the stated constant.\n\nWhat's genuinely new: the two-sided isocapacitary bounds for compact and non-compact manifolds, the explicit DtN operator on the hyperbolic half-ball, and the bottom-of-spectrum computation (2/pi in dimension 2). The upper bounds via Gamma_delta are clean. The applications to hyperbolic surfaces with short geodesic boundaries are sensible and give concrete improvements over existing estimates. This is real work.\n\nSoft spots. Proposition 10 is stated with integral of Cap dt, but the proof integrates against d(t^2); the printed statement is false (u=x on [0,1]). That is an easy fix.\n\nMore seriously, even with the d(t^2) version, the application in Theorem 1's lower bound is wrong. For the Steklov eigenfunction f, the chain gives sigma_1 * int f^2 >= (1/4) Gamma_delta * int_0^∞ t m({f>=t}) dt = (1/4) Gamma_delta * (1/2) int (f_+)^2. The paper replaces int (f_+)^2 by int f^2. Since f has zero mean and changes sign, these are not equal; on the unit disk with f=cos theta the missing factor is 1/2, so the proof at best yields sigma_1 >= (1/8) Gamma_delta without additional argument. No control of int (f_+)^2 relative to int f^2 is given. That is a genuine gap in the central theorem, and Theorem 2 inherits it through (23).\n\nThe non-compact part also silently assumes completeness when using geodesic balls; minor, but should be stated.\n\nNet: the framework is promising, and the upper bounds, the DtN formula, and the examples look solid. The lower bound needs repair. I would send this to a serious referee, but the authors should expect a major revision, and the headline constant may change.","headline":"A plausible and well-motivated transfer of Maz'ya's isocapacitary method to Steklov eigenvalues, but the lower-bound proof has a sign-level gap that likely forces a weaker constant; deserves serious refereeing with expected major revision.","tokens_in":18524,"tokens_out":6406,"would_cite":false,"duration_ms":62920,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["58J50","35P15","53C21"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves two-sided Cheeger-type bounds for Steklov eigenvalues: on a compact Riemannian manifold with boundary, the first nonzero Steklov eigenvalue σ1(M) is at least one quarter and at most twice the isocapacitary constant…","keywords":["Steklov eigenvalue","Dirichlet-to-Neumann operator","isocapacitary constant","Cheeger inequality","capacity","coarea formula","hyperbolic surface","bottom of spectrum"],"falsifier":"Compute the two sides of the printed Proposition 10 for u(x)=x on the interval [0,1]: Cap({x≥t},{x≤0}) = 1/t, so ∫$_0^{1}$ Cap dt diverges while the right side is 4, disproving that printed form. For the main theorem itself, compute σ1 and Γ∂ explicitly on a flat strip or thin cylinder; if the ratio σ1/Γ∂ ever falls outside [1/4,2], the two-sided bound fails.","tokens_in":17519,"feed_emoji":"📐","tokens_out":7093,"duration_ms":71985,"temperature":0.7,"pith_summary":"This paper proves a Cheeger-type sandwich for Steklov eigenvalues: on a compact Riemannian manifold with boundary, the first nonzero Steklov eigenvalue σ1(M) is bounded below by a quarter and above by twice the isocapacitary constant Γ∂(M). The same two-sided bound governs the bottom of the spectrum of the Dirichlet-to-Neumann operator when the manifold is non-compact with embedded boundary. The isocapacitary constant is a purely geometric quantity formed by minimizing the capacity of boundary sets per unit boundary measure. This gives a partial affirmative answer to the question of whether a Buser-type upper bound exists for Steklov spectra, using capacity rather than the classical Cheeger constant.","feed_headline":"Boundary capacities pin Steklov eigenvalues to factor 8","feed_subtitle":"A new two-sided inequality bounds the first Steklov eigenvalue by the isocapacitary constant on compact and noncompact manifolds.","key_machinery":"The central object is the capacity Cap(A,B,M), the minimum Dirichlet energy of a function that is at least 1 on A and at most 0 on B. The isocapacitary constant Γ∂ is this capacity minimized over boundary sets and normalized by the smaller boundary measure. The proof machinery is a capacitary layer inequality: for a smooth function u, the integral over t of the capacity of the superlevel set {u≥t} relative to {u≤0}, weighted by t, is controlled by the Dirichlet energy of u on {u≥0}. This inequality comes from the coarea formula and Hardy's inequality applied to the distribution function ψ(t)=∫_0^t dτ / ∫_{M^τ_u}|∇u|ds. Applying this inequality to a Steklov eigenfunction converts level-set boundary masses into capacity estimates, giving the lower bound; the upper bound uses a direct test function built from a near-minimizer of the capacity ratio.","core_discovery":"The central claim is that the spectral quantity and the geometric quantity are comparable up to a universal factor. For a compact n-dimensional Riemannian manifold with smooth boundary, define Γ∂(M) = min Cap(A,B,M)/min{m(A),m(B)} over disjoint compact boundary sets A,B. The paper establishes (1/4)Γ∂(M) ≤ σ1(M) ≤ 2Γ∂(M). For a non-compact manifold with embedded smooth boundary, define Γ∂(M) = inf Cap(F,M)/m(F) over compact F⊂∂M; then (1/4)Γ∂(M) ≤ inf Spec(D_M) ≤ 2Γ∂(M). In both settings the bottom of the boundary spectral theory is thus controlled, up to the universal factor 8, by boundary capacities.","pith_inferences":["Beyond the paper: once Γ∂(M)>0 on a non-compact manifold, Theorem 2 gives a positive spectral gap above zero for the Dirichlet-to-Neumann operator, implying a form of boundary confinement that the authors do not single out.","Beyond the paper: the same capacity-versus-energy sandwich should transfer to weighted manifolds, where capacity is defined with a density, and the constants can be tested numerically on flat strips or warped cylinders.","Beyond the paper: the paper treats only the first eigenvalue, but the min-max principle suggests isocapacitary upper and lower bounds for higher Steklov eigenvalues σ_k with constants depending on k, which is a natural extension the authors do not state.","Beyond the paper: the hyperbolic-surface estimate σ1(S_{g,1})ℓ(γ) ≤ 4/3 + o(1) raises the question of whether 4/3 is asymptotically sharp for short geodesic boundaries, which the paper does not address."],"forward_implications":["On a compact manifold, the first nonzero Steklov eigenvalue is determined up to a universal factor of 8 by the isocapacitary constant Γ∂, so spectral gaps can be certified by boundary capacity computations.","The proof extends to Steklov–Dirichlet problems: for a submanifold N with mixed boundary data, the first eigenvalue ξ1(N,∂_I N) satisfies the analogous two-sided estimate (1/4)Γ_Y(N) ≤ ξ1 ≤ 2Γ_Y(N).","For compact hyperbolic surfaces, a boundary geodesic of length tending to zero forces the normalized first Steklov eigenvalue to satisfy limsup σ1(S_{g,1})·ℓ(γ) ≤ 4/3; if a surface with at least two boundaries has a boundary curve of length <1, then σ1(S_{g,n}) ≤ e + e^{-1} ≈ 3.086.","For the non-compact hyperbolic upper half-ball, the Dirichlet-to-Neumann operator has an explicit integral representation, and the bottom of its spectrum is exactly 2/π in dimension 2 and at least (∫_0^∞ cosh^{-(n-1)}s ds)^{-1} in dimension n≥3."],"supporting_citations":[{"why":"Introduces the Cheeger-type constant h_j and proves σ1(M) ≥ (1/4)h_c(M)h_j(M), the prior lower bound that this paper extends and complements.","marker":"[21]"},{"why":"The survey that poses Open Question 4.6, asking whether a Buser-type upper bound for σ1 exists; this paper answers it partially via isocapacitary constants.","marker":"[5]"},{"why":"Buser's theorem provides the model upper bound for Laplace eigenvalues in terms of the Cheeger constant, the pattern adapted here to Steklov eigenvalues.","marker":"[2]"},{"why":"Maz'ya's isocapacitary inequality (1/4)Γ(Ω) ≤ λ1(Ω) ≤ Γ(Ω) supplies the exact template for the main theorem.","marker":"[25]"},{"why":"Maz'ya's Sobolev-space book provides the capacity theory, including the distribution-function and Hardy-inequality lemmas used in Proposition 10.","marker":"[29]"},{"why":"Reed–Simon's theorem on positive symmetric operators and closable quadratic forms is applied to define the Dirichlet-to-Neumann operator and its spectrum on non-compact manifolds.","marker":"[33]"}],"fun_headline_variants":["Steklov eigenvalues within factor 8 of boundary capacities","Isocapacitary constants bound Steklov spectrum","Cheeger-type bound for Steklov eigenvalues on manifolds","Boundary shape controls Steklov eigenvalues","Capacities pin the Steklov spectrum to 8"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The lower bound rests on the layer-capacity inequality, Proposition 10, which must compare an integral of t times capacity with the Dirichlet energy; as printed, the proposition lacks the factor t and is false for a linear function on an interval.","fun_headline_variants_meta":{"raw":{"variants":["Steklov eigenvalues within factor 8 of boundary capacities","Isocapacitary constants bound Steklov spectrum","Cheeger-type bound for Steklov eigenvalues on manifolds","Boundary shape controls Steklov eigenvalues","Capacities pin the Steklov spectrum to 8"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000213,"raw_usage":{"total_tokens":1334,"prompt_tokens":773,"completion_tokens":561,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":389,"completion_tokens_details":{"reasoning_tokens":482}},"tokens_in":389,"tokens_out":561,"duration_ms":5187,"temperature":1.0,"reasoning_tokens":482,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T23:07:47.396500+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the two sides of the printed Proposition 10 for u(x)=x on the interval [0,1]: Cap({x≥t},{x≤0}) = 1/t, so ∫$_0^{1}$ Cap dt diverges while the right side is 4, disproving that printed form. For the main theorem itself, compute σ1 and Γ∂ explicitly on a flat strip or thin cylinder; if the ratio σ1/Γ∂ ever falls outside [1/4,2], the two-sided bound fails.","supporting_citations":[{"cited_title":"Une in´ egalit´ e de Cheeger pour le spectre de Steklov","cited_arxiv_id":null,"evidence_quote":"Introduces the Cheeger-type constant h_j and proves σ1(M) ≥ (1/4)h_c(M)h_j(M), the prior lower bound that this paper extends and complements."},{"cited_title":"Some recent develop- ments on the Steklov eigenvalue problem","cited_arxiv_id":null,"evidence_quote":"The survey that poses Open Question 4.6, asking whether a Buser-type upper bound for σ1 exists; this paper answers it partially via isocapacitary constants."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Maz'ya's isocapacitary inequality (1/4)Γ(Ω) ≤ λ1(Ω) ≤ Γ(Ω) supplies the exact template for the main theorem."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Maz'ya's Sobolev-space book provides the capacity theory, including the distribution-function and Hardy-inequality lemmas used in Proposition 10."},{"cited_title":"Methods of modern mathematical physics","cited_arxiv_id":null,"evidence_quote":"Reed–Simon's theorem on positive symmetric operators and closable quadratic forms is applied to define the Dirichlet-to-Neumann operator and its spectrum on non-compact manifolds."}],"review_version":1}