{"id":"439ae2c6-9ad5-494d-a609-70e6110cd1da","arxiv_id":"2411.16418","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A general Schauder theory for uniformly degenerate elliptic equations characterizes optimal boundary Holder regularity via a characteristic polynomial.","lead":"This mathematical preprint builds a general boundary regularity theory, a Schauder theory, for elliptic equations that degenerate smoothly at the boundary. It states sharp Holder continuity of solutions and some derivatives up to the boundary, controlled by roots of a characteristic polynomial.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The omitted induction in Theorem 6.2 for m≥2 is the load-bearing gap: higher-order normal regularity of ∂_t^m u requires a recursive boundary-value construction and Q(m+α)<0 control, neither of which is displayed.","rationale":"The reader's weakest_assumption identifies exactly the omitted induction in Theorems 5.2 and 6.2. My review confirms this is the single most load-bearing concern: Theorem 2.2 is the paper's central claim, and its k≥1 content requires the normal-derivative induction for m≥2. The m=1 case in Theorem 6.2 is detailed and plausible, and the structure for higher m is described verbally (shifted characteristic polynomial Q^{(m)}(μ)=Q(μ+m), boundary values determined by the equation), but the actual recursive step is not supplied. The paper itself flags the omission twice ('The proof for general ℓ is based on induction and hence omitted' and 'The proof for general m is based on induction and hence omitted'), so this is an explicit gap rather than an artifact. The tangential induction in Theorem 5.2 is easier to fill because differentiating along x' does not change the characteristic polynomial; the normal induction is the delicate part because each differentiation shifts Q and produces a new boundary condition that must be shown to be non-circular. The concrete test I propose directly checks the m=2 step; if it passes, the conditional verdict can be upgraded toward acceptance, and if it fails, the theorem as stated is unsupported. I do not see a more fundamental objection: the base cases, the supersolution Lemma 3.1, Lemma 3.2, and Theorem 4.3 are written out in detail, and the sharpness examples in Section 2 are consistent with the stated condition P(k+α)<0. The existence proof in Theorem 2.3 is also sketched, but it uses only k=0 and the omitted δ-independent estimates are less central than the induction needed for the full regularity theorem. Since the reader already assigned CONDITIONAL, my finding does not change that verdict; it sharpens the reason by pinpointing the m=2 recursion as the step to verify first.","tokens_in":24,"tokens_out":7231,"duration_ms":243468,"concrete_test":"Write out the induction step m=2 in the flat model G_1: with v=∂_t u, differentiate L^{(1)}v=f_1 with respect to t and reorganize terms so that the operator acting on ∂_t^2 u has no t^{-1} factors. Verify (i) the resulting operator is t^2a_ij∂_ij+t(b_i+4a_in)∂_i+(c+2b_n+2a_nn), so Q^{(2)}(μ)=Q(μ+2); (ii) f_2 ∈ C^α(bar G_r) with ||f_2||_{C^α} ≤ C(|u|_{L∞}+|f|_{C^{2,α}}), using only the m=1 and tangential estimates; (iii) u_2 = f_2/(c+2b_n+2a_nn) on Σ_1 contains no ∂_t^2 u. If all three hold, the omitted induction is fillable and Theorem 2.2 for k=1 is supported; if (iii) fails, the boundary recursion is circular and the higher-order normal regularity does not follow.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 2.2 for k≥1 depends on Theorem 6.2 with m≥2, yet its proof is omitted: after the m=1 case, the paper states 'The proof for general m is based on induction and hence omitted.' The m=1 argument is explicit: ∂_t u solves L^{(1)}(∂_t u)=f_1, where L^{(1)}=t^2a_ij∂_ij+t(b_i+2a_in)∂_i+(c+b_n), Q^{(1)}(μ)=Q(μ+1), and the boundary value is u_1=f_1/(b_n+c) on Σ_1. For m=2, one must differentiate (6.1) in t, isolate the ∂_t^2 u terms so that the new operator has the form t^2a_ij∂_ij+t(b_i+4a_in)∂_i+(c+2b_n+2a_nn), and verify Q^{(2)}(μ)=Q(μ+2). The new right-hand side f_2 must lie in C^α(bar G_r), using u∈C^{3,α}(G_1) and the already established estimates for ∂_t u, tD∂_t u, and tangential derivatives. The boundary value u_2=f_2/(c+2b_n+2a_nn) on Σ_1 must be expressible using only f, coefficients, and lower-order derivatives of u; if ∂_t^2 u appears on the right-hand side, the recursion is circular. The same recursive construction is needed for the 'more generally' claim after Lemma 6.1 that u is C^{m,α} at the boundary when Q(m+α)<0, and for the step 'Theorem 6.2 implies Theorem 2.2' via localization. None of these steps is written out, so the global C^{k,α} statement for k≥1 is not verifiable from the given arguments.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a global Schauder-type regularity theory for uniformly degenerate elliptic operators of the form L = ρ^2 a^{ij} ∂_{ij} + ρ b^i ∂_i + c in bounded domains, where ρ is a defining function. The central result, Theorem 2.2, asserts that under C^{k+1,α} boundary and C^{k,α} coefficients, with c ≤ −c0 and P(k+α) ≤ −c_{k+α} on the boundary, every bounded solution of Lu = f with boundary value u = f/c belongs to C^{k,α}(Ω̄), with ρ∇^{k+1}u and ρ^2∇^{k+2}u in C^α(Ω̄), and with estimate (2.10). The proof is organized through a flat-boundary local theory: decay estimates (Lemmas 3.1, 3.2), Hölder regularity (Theorem 4.3), tangential regularity (Theorem 5.2), and normal regularity (Theorem 6.2), followed by a regularization argument for existence (Theorem 2.3). Sharpness is illustrated by explicit monomial and logarithmic solutions in Example 2.1.","tokens_in":22892,"tokens_out":3638,"duration_ms":33727,"significance":"If the full statement holds, Theorem 2.2 is a substantial and useful result: it gives optimal boundary regularity in standard Hölder spaces for uniformly degenerate elliptic equations without imposing growth or decay assumptions on the data, and it includes the sharp boundary condition u = f/c and the vanishing of weighted first and second derivatives. The paper has notable strengths: the characteristic polynomial P(μ) is derived directly from the operator, the sharpness examples in Section 2 are explicit constructions rather than re-statements of the theorem, and the base cases of the local theory (Lemmas 3.1, 3.2, 4.2, Theorem 4.3, Theorem 5.2 for ℓ = 1, Lemma 6.1, and Theorem 6.2 for m = 1) are proved in detail. The main weakness is that the higher-order inductions that are essential for the global C^{k,α} statement for k ≥ 1 are not displayed, so the claimed full generality is not verifiable from the given arguments.","major_comments":[{"comment":"The proof of Theorem 6.2 for m ≥ 2 is omitted after the m = 1 case, with the sentence 'The proof for general m is based on induction and hence omitted.' This is a load-bearing gap for Theorem 2.2 when k ≥ 1: the C^{k,α} regularity of u and the estimates for ρ∇^{k+1}u and ρ^2∇^{k+2}u require control of normal derivatives ∂_t^m u for m up to k. The induction is not a routine repetition of the m = 1 argument, because for m = 2 one must differentiate (6.1) in t, isolate the ∂_t^2 u terms so that the new operator has the form t^2a^{ij}∂_{ij} + t(b^i + 4a^{in})∂_i + (c + 2b^n + 2a^{nn}), verify Q^{(2)}(μ) = Q(μ+2), prove that the new right-hand side f_2 lies in C^α(Ω̄), and, crucially, express the boundary value u_2 = f_2/(c + 2b^n + 2a^{nn}) using only f, coefficients, and lower-order derivatives of u. Without this displayed recursion, the argument could be circular if ∂_t^2 u appears in f_2. The same recursive construction is needed for the 'more generally' claim after Lemma 6.1 and for the step 'Theorem 6.2 implies Theorem 2.2.' I therefore cannot verify the k ≥ 1 part of the main theorem from the submitted text.","section":"Theorem 6.2, proof for general m"},{"comment":"The proof of Theorem 5.2 for ℓ ≥ 2 is also omitted after the ℓ = 1 case, stated as 'The proof for the general ℓ is based on induction and hence omitted.' This induction is needed to obtain tangential regularity D_{x'}^τ u for all τ ≤ ℓ and hence for the full C^{k,α} conclusion in Theorem 2.2. The ℓ = 1 case uses the boundedness lemma (Lemma 5.1) and Theorem 4.3 applied to (5.1); the higher-order version requires an analogous construction in which differentiating (5.1) in a tangential direction produces a right-hand side in C^α(Ω̄) with estimates that depend only on the data and on the previously established lower-order regularity. The omitted induction is not a trivial formal exercise because the boundary value of D_{x'}^τ u is determined by the equation through u0 = f/c, and one must check that the recursion preserves that structure at every step.","section":"Theorem 5.2, proof for general ℓ"},{"comment":"The proof of Theorem 2.3 is only sketched. Steps 2 and 3 assert, respectively, a boundary decay estimate and a weighted C^{2,α} estimate for the regularized solutions u_δ, with constants independent of δ, but no details are given. This is not a routine application of the flat-boundary theory because the regularized operator L_δ = L + δΔ has coefficients that are only C^α(Ω̄), the boundary is only C^{1,α}, and the uniform-in-δ estimates must control the passage to the limit in a way that preserves the boundary value u = f/c and the vanishing of ρ∇u and ρ^2∇^2u. Since Theorem 2.3 is the existence half of the claimed isomorphism L : C^{k,α}_2(Ω̄) → C^{k,α}(Ω̄), the absence of a complete argument is a load-bearing gap.","section":"Section 7, proof of Theorem 2.3"},{"comment":"The transition from the flat-boundary local results to the global statement of Theorem 2.2 is compressed into the sentence 'Theorem 6.2 implies Theorem 2.2 easily.' This transition requires a boundary-coordinate change that preserves the structure of the operator, a partition of unity, and a verification that the weighted tangential and normal estimates combine into the stated C^{k,α}(Ω̄) norms and the boundary conditions ρ∇^{k+1}u = 0 and ρ^2∇^{k+2}u = 0 on ∂Ω. Given that the higher-order normal regularity is the delicate part of the paper, the global assembly should be written out rather than left as an exercise.","section":"Sections 4–6, localization to the global theorem"}],"minor_comments":[{"comment":"In the display after 'A simple computation yields Q^{(−κ)}(μ) = Q(μ − κ)', the symbol Q1 should be G1.","section":"Lemma 3.2, proof"},{"comment":"The text 'we write 4 anβ∂tβu = 2anβ∂tβu + 2anβ∂tβu' contains a typo: the first factor should likely be 2, not 4, to match the displayed operator L^{(1)}.","section":"Section 6, heuristic discussion before (6.1)"},{"comment":"The proof of Lemma 4.1 is declared 'standard and hence omitted.' This is acceptable for a calculus lemma, but since the lemma is used to convert pointwise boundary decay into global Hölder regularity, a one-sentence indication of the covering argument would improve readability.","section":"Lemma 4.1"},{"comment":"The hypotheses state Q(1 + α) ≤ −c_{1+α} in G1, while the text also uses Q(1) = b^n + c < 0; it would be clearer to state explicitly that Q(1) < 0 follows from the assumed negativity of Q(0) and Q(1+α), since that is why the denominator in (6.4) is nonzero.","section":"Section 6, statement of Lemma 6.1"},{"comment":"The paper is labeled a survey, but Theorem 2.2 and Theorem 2.3 are presented as new results with proofs; the relation to existing literature (e.g., Graham–Lee [25] and Mazzeo's theory) could be clarified regarding which parts are new and which are expository, but this does not affect the mathematical content.","section":"General presentation"}],"recommendation":"major_revision","confidential_remarks":"The paper has a credible and well-motivated framework, and the base cases of the local theory are proved carefully. The recommendation is based on the fact that the higher-order inductions in Theorems 5.2 and 6.2, and the localization step, are load-bearing for the general C^{k,α} statement and are not supplied. These gaps are fixable within the manuscript's scope by writing out the inductions, so I do not recommend rejection. I would also flag that the proof of Theorem 2.3 needs a substantially more detailed treatment before the existence claim can be considered established."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nHere's the quick take: the paper states a general Schauder theory for uniformly degenerate elliptic operators in standard Hölder spaces, and the base cases are proved carefully. The higher-order normal regularity (Theorem 6.2 for m≥2) is the load-bearing gap: the paper says 'The proof for general m is based on induction and hence omitted,' and that induction is exactly what's needed for the full C^{k,α} statement for k≥1.\n\nWhat's genuinely new is the formulation: standard Hölder spaces, no growth or decay assumptions on f or u, and a pure PDE treatment of normal derivatives that does not require a constant indicial root. The local estimates for α-Hölder regularity (Lemma 3.2, Theorem 4.3), tangential derivatives (Theorem 5.2, ℓ=1), and the first normal derivative (Lemma 6.1, Theorem 6.2 m=1) are written out in detail. The sharpness examples in Section 2 are explicit and illustrate why the condition P(k+α)<0 is necessary. That is real content.\n\nThe soft spots are where the paper stops being self-contained. The induction for tangential derivatives in Theorem 5.2 for ℓ≥2 is omitted. More seriously, the induction for normal derivatives in Theorem 6.2 for m≥2 is omitted. Even the m=2 case requires differentiating (6.1) in t, verifying the new operator has the claimed form with Q^{(2)}(μ)=Q(μ+2), and constructing a boundary value u_2 that is not circular. None of that is displayed. The line 'Theorem 6.2 implies Theorem 2.2 easily' also hides the localization from flat boundary to a curved boundary, which involves the ρ∇^{k+2}ρ conditions and is not trivial. The proof of Theorem 2.3 sketches the δ-regularization estimates, but those are at least plausible from the local theory.\n\nI don't think the gaps are structural in the sense that the approach is wrong; the base arguments are sound and the missing steps have a clear shape. But as the paper stands, Theorem 2.2 for k≥1 is not verifiable from the given arguments. If this is a survey of results proved elsewhere, the authors should say where. If it is meant to be self-contained, the inductions need to be written out, or at least one nontrivial case (m=2, ℓ=2) given.\n\nWho is the audience? Geometric analysts and PDE people who want the statement of the general result and a feel for the method. The paper is readable and the examples are illuminating. I would send it to review, but the referee should insist on filling or referencing the missing inductions before accepting the main theorem.","headline":"A promising general Schauder theory for uniformly degenerate elliptic operators, with careful base proofs, but the higher-order induction steps the main theorem depends on are omitted.","tokens_in":23497,"tokens_out":4168,"would_cite":false,"duration_ms":33083,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35J70","35B65","35J25"],"pacs":[],"model":"deepseek-v4-flash","headline":"For uniformly degenerate elliptic operators, solutions inherit exactly the Hölder regularity the boundary permits—no more, no less—and the paper proves this up to the boundary.","keywords":["uniformly degenerate elliptic equations","boundary regularity","Schauder theory","characteristic polynomial","indicial roots","Hölder spaces","optimal regularity","degenerate elliptic equations"],"falsifier":"In the flat model $L=t^2a^{nn}\\partial_t^2+tb^n\\partial_t+c$ with constant coefficients, take $a^{nn}=1$, $b^n=0$, $c=-1$, so $Q(\\mu)=\\mu(\\mu-1)-1$ and the positive root is $(1+\\sqrt5)/2$. Solve $Lu=f$ explicitly for smooth $f$ and compute the boundary Hölder exponent of the solution and of $\\partial_t^2u$; the theorem predicts the solution is $C^{k,\\alpha}$ exactly when $Q(k+\\alpha)<0$ and no better. The same computation with $f$ chosen so the second normal derivative develops a logarithmic term would test the omitted $m=2$ induction directly.","tokens_in":22259,"feed_emoji":"📐","tokens_out":13814,"duration_ms":120857,"temperature":0.7,"pith_summary":"This survey paper proves a complete boundary Schauder theory for uniformly degenerate elliptic operators $L=\\rho^2 a^{ij}\\partial_{ij}+\\rho b^i\\partial_i+c$, where the second-order part vanishes at the boundary like the square of the defining function $\\rho$. Under the condition that the characteristic polynomial $P(\\mu)$ satisfies $P(k+\\alpha)<0$ on the boundary, any bounded solution of $Lu=f$ with boundary value $u=f/c$ is $C^{k,\\alpha}$ up to the boundary, and the weighted derivatives $\\rho\\nabla^{k+1}u$ and $\\rho^2\\nabla^{k+2}u$ are $C^\\alpha$ and vanish on the boundary, with a full estimate in terms of $|u|_{L^\\infty}$ and $|f|_{C^{k,\\alpha}}$. The condition is sharp: explicit solutions of the form $\\psi\\rho^{k+\\alpha}$ (with logarithmic factors when $k+\\alpha$ is an integer) have arbitrarily smooth data yet cannot be $C^{k,\\beta}$ for any $\\beta>\\alpha$. This matters because uniformly degenerate operators arise from complete conformal metrics, complex Monge-Ampère equations, minimal hypersurfaces in hyperbolic space, and conformally compact Einstein metrics, where earlier regularity results lived in weighted Hölder spaces; the paper removes growth and decay assumptions and obtains the classical Hölder spaces directly.","feed_headline":"Degenerate elliptic PDEs: boundary smoothness is exactly optimal","feed_subtitle":"The characteristic polynomial fixes the sharp Hölder exponent; weighted derivatives stay continuous to the boundary.","key_machinery":"The load-bearing object is the characteristic polynomial $P(\\mu)=\\mu(\\mu-1)a^{ij}\\nu_i\\nu_j+\\mu b^i\\nu_i+c$ on the boundary (with the flat-model version $Q(\\mu)=\\mu(\\mu-1)a^{nn}+\\mu b^n+c$). Its positive root is the sharp Hölder threshold: the condition $P(k+\\alpha)<0$ says the desired exponent lies below that root. The argument is carried by a family of homogeneous supersolutions $\\psi=t^\\sigma(\\varepsilon|x'-x_0'|^2+t^2)^{(\\mu-\\sigma)/2}+K t^\\mu$, whose construction in Lemma 3.1 turns the inequalities $Q(\\sigma)\\le -c_\\sigma$ and $Q(\\mu)\\le -c_\\mu$ into pointwise decay of solutions near the boundary. A calculus lemma then upgrades decay to $C^\\alpha$ bounds on $u$, $tDu$, and $t^2D^2u$. Tangential regularity comes from differentiating the equation along the boundary; normal regularity comes from differentiating in $t$, which produces a shifted operator $L^{(m)}$ with characteristic polynomial $Q(\\mu+m)$ and with boundary values of $\\partial_t^m u$ fixed by the equation.","core_discovery":"The central claim is that the degeneracy does not cause any loss of differentiability beyond the amount already encoded in the characteristic polynomial. For $L=\\rho^2 a^{ij}\\partial_{ij}+\\rho b^i\\partial_i+c$ with $c\\le -c_0$ and $P(k+\\alpha)\\le -c_{k+\\alpha}$ on $\\partial\\Omega$, the Dirichlet problem forces $u=f/c$ on the boundary, and the paper proves that $u\\in C^{k,\\alpha}(\\bar\\Omega)$, $\\rho\\nabla^{k+1}u\\in C^\\alpha(\\bar\\Omega)$, and $\\rho^2\\nabla^{k+2}u\\in C^\\alpha(\\bar\\Omega)$, with these weighted derivatives equal to zero on the boundary and the estimate (2.10). The proof works locally near the boundary and treats tangential and normal derivatives by the same elliptic arguments; normal differentiations shift the characteristic polynomial to $Q^{(m)}(\\mu)=Q(\\mu+m)$, and the boundary values of $\\partial_t^m u$ are determined by the equation. Example 2.1 shows optimality: when the positive characteristic exponent is $k+\\alpha$, one can make $f$ as smooth as desired while $u$ remains only $C^{k,\\alpha}$.","pith_inferences":["A direct test of the omitted inductions is to compute the $m=2$ normal-derivative equation in the flat constant-coefficient model; if the shifted polynomial $Q(\\mu+2)$ and the boundary values $\\partial_t^2u_0$ do not control the right-hand side, the full $C^{k,\\alpha}$ statement for $k\\ge2$ would require a different argument.","The same scheme should extend to fully nonlinear uniformly degenerate equations whose linearization has this structure, since the proof of the linear estimates is the only additional input beyond standard interior Schauder theory; nonlinear boundary blow-up problems linearize exactly to the model equation (1.4).","Because the proof never assumes the positive characteristic root is constant along the boundary, the local estimates should carry over to geometric settings where the indicial root varies from point to point on the conformal boundary, a case the paper explicitly notes its PDE method is designed to handle."],"forward_implications":["For $c\\le 0$ in $\\Omega$ and $c<0$, $P(k+\\alpha)<0$ on $\\partial\\Omega$, the operator $L$ is an isomorphism from the weighted Hölder space $C^{k,\\alpha}_2(\\bar\\Omega)$ onto the usual space $C^{k,\\alpha}(\\bar\\Omega)$, so the degenerate Schauder theory matches the nondegenerate one with weighted control at the boundary.","The regularity is optimal: even with arbitrarily smooth coefficients and data, a solution can be $C^{k,\\alpha}$ but not $C^{k,\\beta}$ for any $\\beta>\\alpha$ when the positive characteristic exponent is $k+\\alpha$, and integer exponents introduce logarithmic factors instead of extra differentiability.","The boundary data are not free: both $u=f/c$ and the normal derivatives $\\partial_t^i u$ up to order $k$ on the boundary are determined by the equation, so the well-posed Dirichlet problem for this class has a single boundary function, namely the ratio of the right-hand side to $c$.","For $k=0$ and any $\\alpha\\in(0,1)$, the regularization argument produces a unique solution for every $f\\in C^\\alpha(\\bar\\Omega)$, with $\\rho\\nabla u$ and $\\rho^2\\nabla^2u$ continuous up to the boundary and vanishing there."],"supporting_citations":[{"why":"Introduces the boundary blow-up problem whose linearization is the model uniformly degenerate operator, motivating the degeneracy structure and the boundary condition $u=f/c$.","marker":"[52]"},{"why":"Establishes existence and regularity in weighted Hölder spaces for the same class of operators, the result the paper improves by working in standard Hölder spaces without growth or decay assumptions.","marker":"[25]"},{"why":"Poses the question of how much boundary regularity is lost for these degenerate equations and treats normal derivatives by ODE arguments that the paper replaces with pure elliptic estimates.","marker":"[50]"},{"why":"Provides the maximum-principle alternative cited for one of the decay estimates in the proof of the boundary Hölder regularity.","marker":"[10]"},{"why":"Supplies the standard calculus lemma used to convert pointwise decay near the boundary into global Hölder semi-norm estimates.","marker":"[8]"}],"fun_headline_variants":["Degenerate elliptic equations hit optimal boundary smoothness","Characteristic polynomial pins down boundary Hölder regularity","Weighted derivatives vanish on boundary; regularity is sharp","No loss beyond intrinsic degeneracy: optimal boundary estimates","Exact boundary smoothness for uniformly degenerate PDEs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument rests on the claimed but omitted induction that repeated differentiation in the normal and tangential directions keeps the equation in the same degenerate class with shifted characteristic polynomials, together with the explicit boundary conditions $c\\le -c_0$ and $P(k+\\alpha)<0$.","fun_headline_variants_meta":{"raw":{"variants":["Degenerate elliptic equations hit optimal boundary smoothness","Characteristic polynomial pins down boundary Hölder regularity","Weighted derivatives vanish on boundary; regularity is sharp","No loss beyond intrinsic degeneracy: optimal boundary estimates","Exact boundary smoothness for uniformly degenerate PDEs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00021,"raw_usage":{"total_tokens":1344,"prompt_tokens":813,"completion_tokens":531,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":429,"completion_tokens_details":{"reasoning_tokens":457}},"tokens_in":429,"tokens_out":531,"duration_ms":7663,"temperature":1.0,"reasoning_tokens":457,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:09:03.107203+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In the flat model $L=t^2a^{nn}\\partial_t^2+tb^n\\partial_t+c$ with constant coefficients, take $a^{nn}=1$, $b^n=0$, $c=-1$, so $Q(\\mu)=\\mu(\\mu-1)-1$ and the positive root is $(1+\\sqrt5)/2$. Solve $Lu=f$ explicitly for smooth $f$ and compute the boundary Hölder exponent of the solution and of $\\partial_t^2u$; the theorem predicts the solution is $C^{k,\\alpha}$ exactly when $Q(k+\\alpha)<0$ and no better. The same computation with $f$ chosen so the second normal derivative develops a logarithmic term would test the omitted $m=2$ induction directly.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes existence and regularity in weighted Hölder spaces for the same class of operators, the result the paper improves by working in standard Hölder spaces without growth or decay assumptions."},{"cited_title":"Lin, On the Dirichlet problem for minimal graphs in hyperbolic sp ace, Invent","cited_arxiv_id":null,"evidence_quote":"Poses the question of how much boundary regularity is lost for these degenerate equations and treats normal derivatives by ODE arguments that the paper replaces with pure elliptic estimates."}],"review_version":1}