{"id":"af2a9417-f532-4afd-93ed-f4f302de6228","arxiv_id":"1908.04021","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"The optimal thickness exponent in the nonlinear geometric rigidity inequality is at least 4/3 (hyperbolic), 1 (elliptic), and 3/2 (parabolic) shells.","lead":"Thin shells are classified by curvature, and this paper proves that the smallest possible exponent in a key rigidity estimate is at least 4/3 for saddle-shaped shells, 1 for dome-shaped shells, and 3/2 for flat-curved shells under an extra condition. The result sets limits on how much a shell can deform without stretching.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The elliptic lower bound in Theorem 1.1(b) depends entirely on inequality (2.48), imported without proof from the author's own arXiv preprint [47]; if that bound fails, the elliptic result is unsupported.","rationale":"The reader's verdict is CONDITIONAL with weak assumption (2.48). My review agrees that this is the most load-bearing concern: unlike the hyperbolic and parabolic proofs, which contain self-contained estimates and explicit ansätze, the elliptic proof is a bare citation to an unpublished preprint [47]. The central claim μ≥1 for elliptic shells stands or falls with (2.48). I also note a second, more systematic issue: all three proofs bound the ratio using Q=I, while the geometric rigidity inequality involves the best Q. This is not a fatal flaw for the hyperbolic and parabolic cases because the oscillations ensure the average skew part is negligible, but the paper does not make that argument; in the elliptic case, (2.48) provides no such control, so the Q=I issue compounds the external dependency. Because the elliptic proof can be repaired by verifying [47], and the hyperbolic/parabolic arguments are otherwise plausible, I do not change the CONDITIONAL verdict. The abstract should also mention assumption (1.7) for the parabolic case.","tokens_in":13803,"tokens_out":18315,"duration_ms":171605,"concrete_test":"Read [47, Theorem 1.4] and verify the full statement and proof of (2.48): check that the ansatz y = -tDw + w n is constructed for arbitrary C^3 elliptic middle surfaces, not only for surfaces in a single principal curvature coordinate, and that the constants in the L∞ and L2-ratio bounds are independent of h. If (2.48) is not proved or is false, Theorem 1.1(b) should be removed or reclassified. Separately, for the constructed deformation in the elliptic case, compute the optimal constant rotation Q_h minimizing ||∇u - Q||_{L2} over SO(3); if inf_Q ||∇u-Q|| / ||∇u-I|| tends to 0, the proof's use of Q=I is invalid.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 1.1(b) is a citation: it assumes (2.48), namely ||∇y||_{L∞} ≤ C/h^{1/2} and σ/h^{1/2} ≤ ||∇y||_{L2}/||sym∇y||_{L2} ≤ C/h^{1/2}, for the elliptic ansatz y = -tDw + w n 'given in the proof [47, Theorem 1.4]'. No part of this estimate is proved or reproduced here. The entire elliptic result μ≥1 follows by combining (2.48) with Lemma 2.4; if (2.48) fails, or is not applicable to C^3 elliptic surfaces without the single-principal-coordinate hypothesis used in [18], then Theorem 1.1(b) is unsupported. The dependence is delicate because the lower bound in (2.48) is a ratio of norms encoding both the optimal Korn scaling and the L∞ control needed for the geometric-rigidity ansatz. In addition, the paper uses ||∇u - I|| as the numerator in the rigidity ratio, whereas (1.3) requires the distance to the best constant rotation Q. The elliptic proof does not justify that the optimal Q is close enough to I for ||∇u - Q|| to be comparable to ||∇u - I||; this gap is also present in the hyperbolic and parabolic cases, but there the detailed oscillatory estimates make it plausible, whereas (2.48) gives no information about the skew-symmetric average that governs the optimal Q.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs explicit deformation fields (Ansätze) for thin shells whose middle surface is hyperbolic, elliptic, or parabolic and proves lower bounds for the optimal exponent μ(Ω) in the geometric rigidity inequality (1.3). The main results are Theorem 1.1 (μ ≥ 4/3 for hyperbolic, μ ≥ 1 for elliptic) and Theorem 1.2 (μ ≥ 3/2 for parabolic under the boundary-crossing condition (1.7)). The proofs aim to transfer the scalings known for the optimal constants in Korn's inequalities to the nonlinear geometric rigidity setting by taking u = z + h^τ y and estimating the ratio ‖∇u - I‖/‖dist(∇u, SO(3))‖ through Lemma 2.4.","tokens_in":14183,"tokens_out":20397,"duration_ms":187678,"significance":"If the proofs are completed, the paper would establish matching lower bounds for the nonlinear rigidity exponents, aligning the geometric rigidity inequality with the known Korn-inequality scalings for shells. The hyperbolic and parabolic constructions are original and involve careful oscillatory and scaling arguments; the elliptic case is a reduction to an estimate from the author's prior work. The connection between Korn-type Ansätze and geometric rigidity is a useful conceptual contribution. However, the current manuscript has gaps in the transition from the constructed test functions to the optimal rotation Q and in the H¹-admissibility of the parabolic ansatz, and the elliptic case is conditional on an unproven cited estimate.","major_comments":[{"comment":"The proofs bound the ratio ‖∇u − I‖² / ‖dist(∇u, SO(3))‖², but inequality (1.3) requires the existence of a rotation Q that minimizes ‖∇u − Q‖. The identity I is only one admissible rotation, and the optimal Q may be significantly closer to ∇u than I is. To conclude μ(Ω) ≥ 4/3 (resp. 3/2, 1), the author must prove that the optimal Q is close to I and that ‖∇u − Q‖ is comparable to ‖∇u − I‖, for example by showing that the L² mean of ∇y, in particular its skew-symmetric part, is negligible compared to ‖∇y‖_{L²}. This control is not provided in the manuscript; in the elliptic case (2.48) gives no information about the mean of ∇y.","section":"Section 2, proof of Theorems 1.1 and 1.2, Eqs. (2.44)–(2.45) and (2.68)–(2.69)"},{"comment":"The elliptic lower bound μ ≥ 1 depends entirely on the estimate (2.48), quoted from the author's unpublished preprint [47, Theorem 1.4] without proof or even a description of the ansatz used to derive it. Since this estimate is load-bearing and is the author's own work, Theorem 1.1(b) is at present conditional on an unverifiable citation. The author should either reproduce the proof of (2.48) in a self-contained way or present the elliptic ansatz and the derivation of the three inequalities in (2.48).","section":"Section 2(b), Eq. (2.48)"},{"comment":"The parabolic ansatz y is defined to be zero outside S0, but the functions v, b, w in (2.62) do not generally vanish on the lateral boundary curves γ(t±(x2), β(x2,p0)), so y has a nonzero trace on ∂S0 and is not an H¹(Ω) function. Since all estimates are only meaningful for admissible deformations u ∈ H¹(Ω), the construction must be multiplied by a smooth cutoff in the x1 direction that vanishes near the boundary; this will add terms to (2.65)–(2.69) that must be incorporated into the estimates. Without such a cutoff, the parabolic lower bound is not established.","section":"Section 2, proof of Theorem 1.2, Eq. (2.64)"},{"comment":"These results rely on [47, Lemma 2.7], a local existence result for a unit vector field X with ∇_X n = 0 on a parabolic surface, quoted without proof. This is another load-bearing import from an unpublished preprint. The author should either prove the lemma or explicitly declare it as an assumption; alternatively, the theorem statement should include this existence condition as part of the hypotheses.","section":"Section 2, proof of Proposition 1.1 and Lemma 2.5"}],"minor_comments":[{"comment":"The displayed equations should read ∇_{e1}n = λ1 e1 and ∇_{e2}n = λ2 e2; as written, the first equation ∇_{e1}n = λ1 e2 and the second ∇_{e2}n = λ2 e2 are inconsistent with the symmetry of the shape operator and with the subsequent use of these relations in (2.13) and in the final computation of Z ⊗ Df.","section":"Lemma 2.2, Eq. (2.11)"},{"comment":"The notation ‖·‖_{L²(S)} is used for surface integrals, while the rigidity inequality (1.3) is over Ω. The conversion between surface and volume norms, which introduces a factor of h, is implicit. Please clarify this convention.","section":"Section 2, proof of Theorem 1.1(a)"},{"comment":"The asymptotic expansions for v, b, w use o(h^{-1/4}), o(h^{-1/2}), etc. without specifying uniformity in x ∈ S0. It would be helpful to state that the remainders are uniform.","section":"Section 2, proof of Theorem 1.2, Eq. (2.62)"},{"comment":"Reference [44] lists the year 2012 with arXiv:1310.5384; the arXiv submission date appears inconsistent. Please check and correct.","section":"References"},{"comment":"There are typographical and formatting issues, e.g., \"Ans¨atze\" for \"Ansätze\" and inconsistent use of \"geometry rigidity inequality\" vs. \"geometric rigidity inequality\". A careful proofreading is recommended.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper contains promising ideas and the hyperbolic proof is quite detailed, but the elliptic theorem is essentially a citation to the author's own unpublished preprint and the parabolic construction has an H¹-admissibility gap. The transition from ‖∇u − I‖ to the optimal Q in the rigidity inequality is a logical gap in all three cases. I would encourage the editor to ask for the preprints [47] and [48] to be made available or formally published before final acceptance. The scope of the paper is appropriate for the journal, but substantial revision is needed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear [Colleague],\n\nRead Yao's paper on lower bounds for the optimal exponent in the nonlinear geometric rigidity inequality for shells. The headline: the hyperbolic lower bound (4/3) and the parabolic lower bound (3/2, under a boundary condition) are new and the proofs are largely self-contained. The elliptic case (μ≥1) is real only if the bound (2.48) imported from the author's own preprint [47] holds; the paper does not prove or reproduce it. That makes the elliptic result conditional, not a settled theorem.\n\nWhat is good: the paper constructs explicit oscillatory Ansätze and works out the scaling estimates carefully. For hyperbolic shells, the argument from Lemma 2.1 through (2.44)-(2.45) is coherent and does not rely on the single-principal-coordinate assumption (1.9) used in earlier Korn work. The parabolic construction (Section 2.6, Theorem 1.2) is also original and the geodesic coordinate lemma is a nice piece of geometry. The lower bounds match the known Korn scalings, which is expected but still worth having for the nonlinear problem.\n\nSoft spots, in order of seriousness:\n\n1. The elliptic case (Theorem 1.1(b)) is a citation. Inequality (2.48) carries the whole proof: it gives the L∞ control and the norm ratio for the ansatz. If that bound fails, the elliptic result falls. Since the cited work is an arXiv preprint, a referee needs to see the proof. This should be a condition for publication.\n\n2. The abstract says the parabolic lower bound is 3/2; it omits assumption (1.7) (the geodesic crossing condition). That overstates the theorem.\n\n3. Lemma 2.2, equation (2.11): the shape operator action is misprinted (∇_{e1} n = λ1 e2 should be λ1 e1, presumably). The subsequent formulas (2.12)-(2.14) are consistent with the corrected version, so it's a typo, but it should be fixed.\n\n4. Both the reader and the stress-test note flag that the ratio is computed with ||∇u-I|| in the numerator, whereas (1.3) infimizes over Q∈SO(3). In the hyperbolic and parabolic cases the oscillatory ansätze have skew parts with small mean, so the optimal Q is plausibly close to I and the argument can be repaired; but the paper does not say this. In the elliptic case, (2.48) gives no information about the skew-symmetric average, so this is a real gap, not just a missing sentence. If I were refereeing, I would ask for a lemma that controls ||∇u - Q_opt|| from below, or at least an explicit statement of why the difference from ||∇u-I|| is negligible.\n\nWho should read it: anyone working on rigorous derivation of shell theories or optimal constants in geometric rigidity. It is a serious paper worth referee time, but I would not accept it in current form. I'd send it back for major revision, mainly to make the elliptic case self-contained or clearly labeled as conditional, and to fix the abstract and the typo. With those changes, the hyperbolic and parabolic results would stand on their own.\n\nHope this helps.","headline":"New lower bounds for the nonlinear rigidity exponent in shells; the hyperbolic and parabolic proofs are detailed and largely self-contained, but the elliptic result is conditional on an unproved bound imported from the author's own preprint.","tokens_in":14618,"tokens_out":6191,"would_cite":true,"duration_ms":61409,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["74K20","74B20"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves lower bounds of 4/3, 1, and 3/2 on the optimal thickness exponent in the geometric rigidity inequality for hyperbolic, elliptic, and parabolic shells, respectively.","keywords":["geometry rigidity inequality","shell","nonlinear elasticity","Riemannian geometry","optimal thickness exponent","Gaussian curvature","Korn inequality","Ansatz construction"],"falsifier":"Check the imported elliptic estimate directly: compute $\\|\\nabla y\\|_{L^\\infty}$ and the ratio $\\|\\nabla y\\|_{L^2}/\\|\\operatorname{sym}\\nabla y\\|_{L^2}$ for the ansatz $y=-tDw+w\\vec n$ on a spherical shell patch. If the ratio does not stay in the interval $[\\sigma h^{-1/2}, Ch^{-1/2}]$, the elliptic theorem loses its main support; a numerical search for a deformation making (1.3) hold with some $\\mu<1$ on a hemisphere would then decide the matter.","tokens_in":13532,"feed_emoji":"🐚","tokens_out":10424,"duration_ms":100427,"temperature":0.7,"pith_summary":"The paper establishes curvature-dependent lower bounds on the optimal thickness exponent in the geometric rigidity inequality for thin shells. For a shell whose middle surface is hyperbolic, elliptic, or parabolic, the exponent $\\mu(\\Omega)$ in (1.3) cannot be smaller than $4/3$, $1$, or $3/2$ (the last under a boundary condition), respectively. The proof works by constructing explicit trial deformations whose ratio of full strain to distance from the rotation group forces these powers of $1/h$. These bounds matter because any shell theory derived from three-dimensional elasticity by $\\Gamma$-convergence must respect the corresponding decay rate of the rigidity constant.","feed_headline":"Shell curvature sets minimum rigidity exponents: 4/3, 1, 3/2","feed_subtitle":"Thin-shell rigidity cannot decay slower than these powers of thickness; explicit near-buckling deformations force the bounds.","key_machinery":"The carrying mechanism is the ansatz: an explicit family of test deformations built from the geometry of the middle surface. For hyperbolic shells, the ansatz uses an asymptotic coordinate system and the 90-degree rotation operator $Q$ in the tangent plane; the key identity $\\operatorname{sym} Z\\otimes Df = v\\Pi$ converts the shape operator into a symmetric gradient. For elliptic shells, the deformation is the same ansatz already used for Korn's inequality, $y=-tDw+w\\vec n$, and the argument imports sharp two-sided bounds from the linear theory. For parabolic shells, the construction uses the straight-line geodesics of Proposition 1.1 and a principal coordinate system along them; the ansatz $V+tW+b\\vec n$ produces oscillations of frequency $h^{-1/4}$ that force the $h^{-3/2}$ scaling. In all cases the key comparison is Lemma 2.4, which relates the distance to $SO(3)$ to the symmetrized strain $\\Phi(B)$, so that lower bounds on the strain ratio translate directly into lower bounds on $\\mu(\\Omega)$.","core_discovery":"For a $C^3$ shell $\\Omega=\\{p+t\\vec n(p): p\\in S,\\ |t|<h/2\\}$, the paper proves $\\mu(\\Omega)\\ge 4/3$ when the Gaussian curvature of $S$ is negative, $\\mu(\\Omega)\\ge 1$ when it is positive, and $\\mu(\\Omega)\\ge 3/2$ in the parabolic case provided the straight-line geodesic through a point of $S$ exits the boundary transversally as in (1.7). These are lower bounds on the infimum of exponents for which inequality (1.3) holds with a constant independent of $h$. The proof exhibits deformations $u(z)=z+h^\\tau y(z)$ with $\\tau$ larger than the claimed bound and shows, via Lemma 2.4, that the ratio $\\|\\nabla u-I\\|_{L^2}/\\|\\mathrm{dist}(\\nabla u,SO(3))\\|_{L^2}$ has the stated power-law size in $1/h$; no constant independent of $h$ can therefore make (1.3) valid with a smaller exponent.","pith_inferences":["If the equality conjecture in Remark 1.1 holds, the rigidity exponent becomes a curvature classifier: elliptic $1$, parabolic $3/2$, hyperbolic $4/3$; proving matching upper bounds is the natural next step.","The oscillatory ansaetze suggest the same exponents should appear in linear buckling and eigenvalue problems for shells; testing whether the first nontrivial eigenvalue scales like $h^{2\\mu}$ would connect rigidity to stability.","The hyperbolic ansatz oscillates at frequency $h^{-1/3}$ and the parabolic at $h^{-1/4}$, so the construction predicts distinct wavelength selections in near-buckling patterns, and numerical experiments could look for those wavelengths."],"forward_implications":["Any $\\Gamma$-convergence derivation of a shell model from (1.3) must use energy scalings consistent with rigidity constants that grow at least as $h^{-4/3}$, $h^{-1}$, or $h^{-3/2}$ according to the sign of the Gaussian curvature.","The explicit deformations constructed in the proofs can serve as near-optimal test fields for numerical benchmarking of shell elements, since they saturate the ratio controlled by the inequality.","For hyperbolic shells the lower bound is achieved without any boundary assumption, so the $4/3$ exponent is an intrinsic feature of negative curvature rather than of a particular boundary geometry.","For parabolic shells, the $3/2$ bound currently depends on the geodesic-crossing condition (1.7), so extending the bound to closed parabolic shells is a natural next test of the method."],"supporting_citations":[{"why":"Supplies the elliptic-shell ansatz and the two-sided ratio estimate (2.48) on which the elliptic lower bound rests.","marker":"[47]"},{"why":"Proves the geometric rigidity inequality and the L2 framework that the shell estimate extends.","marker":"[10]"},{"why":"Establishes the plate case exponent and the general hierarchy of rigidity estimates.","marker":"[11]"},{"why":"Provides the Korn-inequality ansatz for zero Gaussian curvature used in the parabolic construction.","marker":"[13]"},{"why":"Gives exact scaling exponents for cylindrical shells, the linear precursor of the parabolic result.","marker":"[15]"},{"why":"Identifies Gaussian curvature as the classifier of shell rigidity and supplies elliptic and hyperbolic Korn exponents.","marker":"[18]"},{"why":"Gives the general upper bound $\\mu(\\Omega)\\le 2$ for shells, the benchmark the lower bounds sharpen.","marker":"[27]"},{"why":"Shows principal coordinates exist only locally for hyperbolic and parabolic surfaces, motivating the coordinate-free ansatz.","marker":"[48]"},{"why":"Provides the flow theorem used to construct principal coordinates along parabolic geodesics.","marker":"[26]"}],"fun_headline_variants":["Shell rigidity floors by geometry: 4/3, 3/2, 1","Curvature dictates minimum shell rigidity exponents","Geometry pins shell rigidity exponents: 4/3, 3/2, 1","Thin shells: rigidity can't beat 4/3, 3/2, 1","Hyperbolic, parabolic, elliptic: rigidity minima 4/3, 3/2, 1"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"For the elliptic case, the proof imports a two-sided estimate from a companion preprint: for the trial deformation $y=-tDw+w\\vec n$, the ratio of the full gradient to the symmetric gradient must grow exactly like $h^{-1/2}$. If that imported estimate fails, the claimed bound $\\mu(\\Omega)\\ge 1$ is not established.","fun_headline_variants_meta":{"raw":{"variants":["Shell rigidity floors by geometry: 4/3, 3/2, 1","Curvature dictates minimum shell rigidity exponents","Geometry pins shell rigidity exponents: 4/3, 3/2, 1","Thin shells: rigidity can't beat 4/3, 3/2, 1","Hyperbolic, parabolic, elliptic: rigidity minima 4/3, 3/2, 1"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000857,"raw_usage":{"total_tokens":3670,"prompt_tokens":839,"completion_tokens":2831,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":455,"completion_tokens_details":{"reasoning_tokens":2720}},"tokens_in":455,"tokens_out":2831,"duration_ms":22022,"temperature":1.0,"reasoning_tokens":2720,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:54:38.900791+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check the imported elliptic estimate directly: compute $\\|\\nabla y\\|_{L^\\infty}$ and the ratio $\\|\\nabla y\\|_{L^2}/\\|\\operatorname{sym}\\nabla y\\|_{L^2}$ for the ansatz $y=-tDw+w\\vec n$ on a spherical shell patch. If the ratio does not stay in the interval $[\\sigma h^{-1/2}, Ch^{-1/2}]$, the elliptic theorem loses its main support; a numerical search for a deformation making (1.3) hold with some $\\mu<1$ on a hemisphere would then decide the matter.","supporting_citations":[{"cited_title":"Optimal exponentials of thickness in Korn's inequalities for parabolic and elliptic shells","cited_arxiv_id":"1807.11114","evidence_quote":"Supplies the elliptic-shell ansatz and the two-sided ratio estimate (2.48) on which the elliptic lower bound rests."},{"cited_title":"Friesecke, R","cited_arxiv_id":null,"evidence_quote":"Proves the geometric rigidity inequality and the L2 framework that the shell estimate extends."},{"cited_title":"Friesecke, R","cited_arxiv_id":null,"evidence_quote":"Establishes the plate case exponent and the general hierarchy of rigidity estimates."},{"cited_title":"Grabovsky and D","cited_arxiv_id":null,"evidence_quote":"Provides the Korn-inequality ansatz for zero Gaussian curvature used in the parabolic construction."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives exact scaling exponents for cylindrical shells, the linear precursor of the parabolic result."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Identifies Gaussian curvature as the classifier of shell rigidity and supplies elliptic and hyperbolic Korn exponents."},{"cited_title":"Lewicka, M","cited_arxiv_id":null,"evidence_quote":"Gives the general upper bound $\\mu(\\Omega)\\le 2$ for shells, the benchmark the lower bounds sharpen."},{"cited_title":"Linear Strain Tensors and Optimal Exponential of thickness in Korn's Inequalities for Hyperbolic Shells","cited_arxiv_id":"1807.11115","evidence_quote":"Shows principal coordinates exist only locally for hyperbolic and parabolic surfaces, motivating the coordinate-free ansatz."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the flow theorem used to construct principal coordinates along parabolic geodesics."}],"review_version":1}