{"id":"602ef1a6-2844-4866-95ba-9da16d4a3ad8","arxiv_id":"1908.06584","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For prescribed mean curvature equations with a small W^{1,p} forcing term and small boundary data, a W^{2,q} solution exists near any given graphical minimal surface.","lead":"This mathematics paper proves that a prescribed mean curvature equation has a solution near a given minimal surface whenever the forcing term is small in a sharp Sobolev norm. The result gives a rigorous existence theory for such equations with unbounded, non-smooth data, motivated by a singular perturbation model.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (2.29) is not merely unproved; it is false. The expansion used to build the fixed-point operator drops a term (∇h·∇w)h_{ii}/(1+|∇h+∇w|²)^{3/2}, so a fixed point of T solves a different equation than (1.8).","rationale":"The reader identified (2.29) as an unproved identity and made the report conditional. A first-principles expansion shows the identity is not merely unproved but false: the minimal surface constraint h_{ii}=z_i z_j h_{ij}/T substituted into A_{ij}(z+w)h_{ij} produces a z_i z_j/T term, not a δ_{ij} term. The missing contribution survives at every fixed point of the Leray-Schauder operator, so the argument solves a different PDE. This is the single most load-bearing step in the paper, and the flaw is concrete and checkable. Because the main theorem's proof depends entirely on this reduction, the appropriate verdict is REJECT as written, while noting the theorem might be repairable with the corrected expansion. The paper's other aspects, including the function space setup, Lemma 2.3, and the compactness argument, are standard and were not found defective.","tokens_in":8614,"tokens_out":17664,"duration_ms":146251,"concrete_test":"Compute both sides of the claimed identity (2.29) for a Scherk minimal surface h(x,y)=log(cos y/cos x) at a point with ∇h≠0, e.g., (0.5,0), and a small test function w(x,y)=εx with ε=0.1. The left side of (2.26) and the right side of (2.29) will differ by (∇h·∇w)h_{ii}/(1+|∇h+∇w|²)^{3/2} ≈ 0.0096, whereas the paper's identity requires equality. This settles the concern.","verdict_should_be":"REJECT","load_bearing_attack":"Equation (2.29) is the algebraic hinge of the proof of Theorem 1.1. Let z=∇h, w=∇ũ, S=1+|z+w|², T=1+|z|². The minimal surface equation (1.5) gives h_{ii}=z_i z_j h_{ij}/T. Substituting this into A_{ij}(z+w)h_{ij} yields the correct expansion A_{ij}(z+w)h_{ij} = h_{ij} S^{-3/2} [ z_i z_j(|w|^2+2z·w)/T - w_iw_j - w_i z_j - w_j z_i ]. Equation (2.29) instead contains [ (|w|^2+z·w)δ_{ij} - w_iw_j - w_i z_j - w_j z_i ]. Contracting the difference against h_{ij} gives -(z·w)h_{ii}S^{-3/2}, which is not zero in general. For example, the Scherk minimal graph h(x,y)=log(cos y/cos x) has h_{ii}=Δh≠0, and at (0.5,0) with w=(0.1,0) the two sides of (2.29) differ numerically by about 0.0096. Consequently a fixed point of T in (2.31) satisfies (1.8) with an extra right-hand side (∇h·∇w)h_{ii}/(1+|∇h+∇w|²)^{3/2}, not the stated equation. The theorem may be true, but the proof as written does not establish it.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves (Theorem 1.1) that for a bounded C^{1,1} domain Ω and a W^{2,∞} minimal graph h, if G ∈ W^{1,p}(Ω×R) and φ ∈ W^{2,q}(Ω) are sufficiently small (with p ∈ ((n+1)/2, n+1) and q = np/(n+1−p)), then for any measurable H with |H| ≤ |G| there exists u ∈ W^{2,q}(Ω) with u − h − φ ∈ W^{1,q}_0(Ω) solving div(∇u/√(1+|∇u|^2)) = H(x,u,∇u) and ‖u−h‖_{W^{2,q}} < ε. The proof linearizes the mean curvature operator around the minimal surface h, builds a Leray–Schauder fixed point operator on a compact set of C^{1,α} graphs, and uses a trace lemma for the composition map G(·,v(·)).","tokens_in":8950,"tokens_out":9690,"duration_ms":88077,"significance":"If the proof were correct, the result would be a welcome perturbative existence theorem for prescribed mean curvature equations with right-hand sides only in a critical Sobolev space, extending earlier work that requires bounded or L∞ data. The overall strategy is standard and the trace lemma (Lemma 2.3) is a careful and useful application of Ziemer's results. However, the validity of Theorem 1.1 hinges on identity (2.29), which is false; the proof as written does not establish the theorem.","major_comments":[{"comment":"Equation (2.29) is false. Let z = ∇h and w = ∇ũ, and set S = 1+|z+w|² and T = 1+|z|². The minimal surface equation gives h_{ii} = (z_i z_j / T) h_{ij}. Substituting this into A_{ij}(z+w) h_{ij} yields A_{ij}(z+w) h_{ij} = S^{-3/2} [ (|w|²+2z·w) z_i z_j / T − w_i w_j − w_i z_j − w_j z_i ] h_{ij}. The formula in (2.29) instead contains (|w|² + z·w) δ_{ij} in the bracket. The difference between the correct expression and (2.29) is S^{-3/2} (z·w) h_{ii}, which does not vanish in general. For example, the Scherk minimal surface h(x,y) = log(cos y / cos x) has h_{xx}+h_{yy} ≠ 0, and at (0.5,0) with w = (0.1,0) the two sides of (2.29) differ numerically. Hence (2.26) and (2.29) are not equivalent.","section":"Section 2, Eq. (2.29)"},{"comment":"The fixed-point operator T is defined by (2.31), whose left-hand side is derived from the incorrect identity (2.29). Consequently, a fixed point of T satisfies (2.26) with an additional term (∇h·∇ũ) h_{ii} / (1+|∇h+∇ũ|²)^{3/2} on the left-hand side, not the stated equation. The assertion in the proof of Theorem 1.1 that a fixed point of T satisfies (2.26)–(2.27) therefore does not follow. This is a load-bearing gap for the main theorem.","section":"Section 2, Eq. (2.31) and proof of Theorem 1.1"}],"minor_comments":[{"comment":"The statement of the Leray–Schauder fixed point theorem is incorrect as written: the bound should hold for all u satisfying u = λ T u with λ ∈ [0,1], not for all u ∈ B. The subsequent proofs use the standard theorem, so this is a presentation issue.","section":"Theorem 2.4"},{"comment":"The last norm in the chain should be W^{2,q}(Ω) rather than W^{2,p}(Ω), since the solution w constructed by Theorem 2.1 belongs to W^{2,q}(Ω).","section":"Equation (2.19)"},{"comment":"There are numerous OCR/typos in the exponent notation, for example 'C^{1, 1/2 − n/(2q)}' appears as 'C1, 12− n2q'. The paper should be carefully typeset.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The reader's report gave a conditional verdict and noted that identity (2.29) was asserted without derivation but appeared correct on a spot check. In fact the identity is false; the spot check in that report was not sufficiently detailed. The proof of Theorem 1.1 is therefore invalid as written, although the theorem may be salvageable by replacing (2.29) with the correct expansion and including the missing quadratic term in the definition of T. This will require additional estimates, so the revision is substantial."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper has a reasonable idea and one genuinely nice lemma, but the proof of the main theorem has a false algebraic step. Equation (2.29) is not a correct conversion of (2.26). Let z=∇h, w=∇ũ, S=1+|z+w|², T=1+|z|². The minimal surface equation gives Δh = (z_i z_j h_ij)/T. Expanding A_ij(z+w)h_ij and using that relation gives a term proportional to (z·w)Δh / S^{3/2} which does not appear in (2.29). The stress-test example (Scherk's surface) shows the disagreement is numerically nonzero. So a fixed point of the operator T in (2.31) satisfies a PDE with an extra right-hand side (∇h·∇ũ)Δh/(1+|∇h+∇ũ|²)^{3/2}; it does not solve (1.8). This is not a small gap, it's the hinge of the whole reduction. The theorem may be true, but this proof does not establish it.\n\nWhat is genuinely new here is the dimensionally sharp Sobolev control: H is allowed to be unbounded as long as it is dominated by G ∈ W^{1,p} with p > (n+1)/2, and the exponent q = np/(n+1-p) is the natural one. Earlier papers (Gerhardt, Miranda, Bergner, Marquardt) all assume bounded H or monotonicity. Lemma 2.3, which controls the composition via a measure-density argument, is carefully done and seems correct. The compactness and continuity arguments in the Leray–Schauder setup are standard but executed cleanly.\n\nThe softer issues are real but minor: the introduction claims 'well-posedness' when only existence is proven; uniqueness appears only as a remark with extra hypotheses. The self-citation [11] is for motivation, not for a proof step, and that is fine.\n\nFor the right audience—people working on prescribed mean curvature with low-regularity data—this would be a useful theorem if the proof were correct. As it stands, I would not cite it or bring it to a reading group. The author needs to redo the expansion in (2.29), either correcting the missing term or finding another reduction. If a fixed version appears, it deserves a serious referee; the current one does not.","headline":"Plausible result and a solid trace lemma, but the proof rests on a false identity in (2.29) and does not currently prove Theorem 1.1.","tokens_in":9512,"tokens_out":9912,"would_cite":false,"duration_ms":89121,"reading_group":"no","serious_thinker":"no","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35J60","35J93","53A10"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that for any minimal graph h and small Sobolev-class data, the prescribed mean curvature Dirichlet problem has a nearby solution in W^{2,q}.","keywords":["prescribed mean curvature","Dirichlet problem","minimal surface","Sobolev spaces","Leray-Schauder fixed point","graphical solutions","mean curvature operator","trace inequality"],"falsifier":"Take a concrete nontrivial minimal graph h, for instance h(x)=x_1 on a disk, and a nonzero small perturbation v, and evaluate both sides of (2.29) symbolically or numerically at a point of Ω; if the identity fails to hold as an algebraic equality in the second derivatives of u, the proof's reduction is broken and the theorem's conclusion is unsupported.","tokens_in":8382,"feed_emoji":"📐","tokens_out":5910,"duration_ms":51741,"temperature":0.7,"pith_summary":"The paper proves an existence theorem for the prescribed mean curvature Dirichlet problem: if the boundary data and the right-hand side are small in a certain Sobolev norm, then a solution exists in a neighbourhood of any given graphical minimal surface. The norm on the forcing term is $W^{{1,p}}$(Ω×R) with (n+1)/2 < p < n+1, which the paper describes as dimensionally sharp, and the solution space is $W^{{2,q}}$ for q = np/(n+1-p). This matters because the natural vector-field right-hand sides arising from singular perturbation problems have exactly this $W^{{1,p}}$ regularity and need not be bounded in L∞. The proof works by linearizing the mean curvature operator about the minimal surface, turning the problem into a fixed point of a compact operator on a Hölder ball, and applying Leray-Schauder. The paper also gives a corollary for the geometric case H(x,t,z)=ν(z)·f(x,t), with f small in $W^{{1,p}}$, and notes a uniqueness statement under monotonicity.","feed_headline":"Near a minimal graph, small Sobolev data force a solution","feed_subtitle":"The forcing term needs only a Sobolev bound, so it can be unbounded—the case arising in interface motion.","key_machinery":"The proof rests on three mechanisms. First, the linearized operator L[z](u)=A_{ij}(z)u_{x_ix_j} with A_{ij}(z)=(1+|z|^2)^{-1/2}(δ_{ij}-z_iz_j/(1+|z|^2)) is uniformly elliptic with ellipticity constant (1+‖v‖^2_{$C^{{1,α}}$})^{-3/2}, so standard $W^{{2,q}}$ regularity applies. Second, to handle an unbounded right-hand side, Lemma 2.3 proves ‖G(·,v(·))‖_{L^q(Ω)}≤c‖G‖_{$W^{{1,p}}$(Ω×R)} by viewing the graph of v as a codimension-one set and using a Radon-measure trace inequality for functions in $W^{{1,p}}$. Third, the entire argument is organized around the identity in (2.29), which expands the mean curvature operator around the minimal surface h and introduces the linear first-order term B(∇v)·∇w that makes the fixed-point map T contractively compact on a $C^{{1,1/2−n/(2q)}}$ ball.","core_discovery":"The central claim is Theorem 1.1: under the stated smallness condition on G and φ, there exists u∈$W^{{2,q}}$(Ω) with u−h−φ∈$W^{{1,q}}$_0(Ω) satisfying div(∇u/√(1+|∇u|^2))=H(x,u,∇u) in Ω and ‖u−h‖_{$W^{{2,q}}$}<ε. Equivalently, the set of prescribed mean curvature solutions is not empty at every sufficiently small perturbation of a minimal graph, measured in Sobolev norms adapted to the dimension. The novelty relative to earlier prescribed mean curvature results is that H is controlled only through a $W^{{1,p}}$ bound on |G(x,t)|, so H need not be bounded; the proof imports a trace-type estimate that controls G(x,v(x)) in L^q(Ω) from the $W^{{1,p}}$ norm of G on Ω×R.","pith_inferences":["If identity (2.29) is verified, the same fixed-point scheme likely extends to perturbations of any stable minimal surface that is only C^{2,α}, by freezing coefficients at a smoother approximation; the W^{1,p} restriction estimate is the part that genuinely uses the graphical structure.","The dimensionally sharp range (n+1)/2n."],"supporting_citations":[{"why":"Supplies the elliptic regularity and maximum-principle estimates (Theorems 9.1, 9.13, 9.15) and the Leray-Schauder fixed point theorem (Theorem 11.3) used throughout the proof.","marker":"[4]"},{"why":"Used as Theorem 2.2, the measure-theoretic trace inequality that gives the L^q estimate for G restricted to the graph of v.","marker":"[12]"},{"why":"The singular perturbation problem that motivates the W^{1,p} right-hand side and the geometric form H=ν·f.","marker":"[11]"}],"fun_headline_variants":["Small Sobolev data guarantee solutions near minimal graphs","Unbounded prescribed curvature: existence near minimal graph","No boundedness needed: mean curvature solutions near minimal graph","Small Sobolev forcing: solutions near minimal graphs exist","Unbounded data but still a solution near minimal graphs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The algebraic identity (2.29), which rewrites the mean curvature operator around the minimal surface h, is stated without derivation; the entire fixed-point operator is built from it, so if the identity were false the constructed fixed point would solve a different equation.","fun_headline_variants_meta":{"raw":{"variants":["Small Sobolev data guarantee solutions near minimal graphs","Unbounded prescribed curvature: existence near minimal graph","No boundedness needed: mean curvature solutions near minimal graph","Small Sobolev forcing: solutions near minimal graphs exist","Unbounded data but still a solution near minimal graphs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000827,"raw_usage":{"total_tokens":3525,"prompt_tokens":766,"completion_tokens":2759,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":382,"completion_tokens_details":{"reasoning_tokens":2690}},"tokens_in":382,"tokens_out":2759,"duration_ms":20164,"temperature":1.0,"reasoning_tokens":2690,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:40:54.767149+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a concrete nontrivial minimal graph h, for instance h(x)=x_1 on a disk, and a nonzero small perturbation v, and evaluate both sides of (2.29) symbolically or numerically at a point of Ω; if the identity fails to hold as an algebraic equality in the second derivatives of u, the proof's reduction is broken and the theorem's conclusion is unsupported.","supporting_citations":[{"cited_title":"Gilbarg, N","cited_arxiv_id":null,"evidence_quote":"Supplies the elliptic regularity and maximum-principle estimates (Theorems 9.1, 9.13, 9.15) and the Leray-Schauder fixed point theorem (Theorem 11.3) used throughout the proof."},{"cited_title":"Ziemer, Weakly differentiable functions, Springer-Verlag (1989)","cited_arxiv_id":null,"evidence_quote":"Used as Theorem 2.2, the measure-theoretic trace inequality that gives the L^q estimate for G restricted to the graph of v."},{"cited_title":"A diffused interface with the advection term in a Sobolev space","cited_arxiv_id":"1904.00525","evidence_quote":"The singular perturbation problem that motivates the W^{1,p} right-hand side and the geometric form H=ν·f."}],"review_version":1}