{"id":"50d4f9ef-0e71-4517-bc79-44ffb7e4d909","arxiv_id":"2504.18494","paper_version":6,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper develops a generalized Hodge theory in which sequences of overdetermined boundary-value systems satisfying a weaker order-reduction property are lifted to genuine cochain complexes whose cohomology is explicitly described.","lead":"This paper builds a general framework, called elliptic pre-complexes, that turns solvability and uniqueness of linearized boundary-value problems into cohomology computations. It applies Hodge-theoretic machinery to equations such as Killing, Hessian, and curvature equations without assuming a flat background.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Riemann example's order-reduction hinges on unverified condition (1.4.27) for connection terms; without it Theorem 1.15 does not follow.","rationale":"The reader's verdict is CONDITIONAL and identifies the same area of concern: the geometric applications depend on ellipticity and order-reduction verifications that are partly asserted rather than fully shown. I sharpen this to a single checkable condition, namely the extra hypothesis (1.4.27) in the Riemann example. This is load-bearing because if it fails, the Riemann pre-complex does not satisfy the hypotheses of Theorem 3.14, and the advertised cohomological solvability criterion for the linearized Riemann problem collapses. The Ricci case is already explicitly excluded for dim > 3 (disrupted, deferred to a separate paper [Led25]), so the Riemann condition is the best remaining concrete test of the paper's applicability claims. The central abstract theorem may still be correct; the issue is whether a headline application meets its hypotheses. Thus the verdict stays CONDITIONAL pending this check.","tokens_in":67008,"tokens_out":16930,"duration_ms":154906,"concrete_test":"Take M = B^3 with the Euclidean metric g, and pick a non-zero compactly supported vector field X with X|∂M = 0 but normal derivative nonzero on ∂M. Set σ = L_Xg. By construction σ ∈ Ker(P_tt,DA_g). For a tame connection term Γ_g(Rm_g,σ) of the type referenced in [Ham82, pp. 92–93], compute the boundary term P_tΓ_g(Rm_g,σ) on ∂M. If this quantity is nonzero for some X, condition (1.4.27) is not 'trivially' satisfied and the order-reduction proof for the Riemann example has a gap. If it vanishes identically for all such X, the paper's assertion is corroborated.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central theorem's applicability to the Riemann curvature equations—a headline application—requires the Dirichlet order-reduction property (Definition 3.13(ii)(a)) for the segment A_1 = DΓRm_g, A_2 = d_g. The paper's verification (1.4.26) shows B_2A_1σ = P_t DΓRm_gσ = d/dt(P_t Rm_{g+tσ}) = 0 only when Γ = 0; for nonzero connection term Γ, the identity holds only under the extra condition P_tΓ = 0 on Ker(P_tt,DA_g), Eq. (1.4.27). The text asserts this is 'satisfied trivially' when Γσ = Γ_g(Rm_g,σ) arises from a tame connection on M_M×C^{2,2}_M, but no proof is given. The kernel of (P_tt,DA_g) contains all infinitesimal diffeomorphisms fixing the boundary (σ = L_Xg with X|∂M = 0), and P_tΓ is a bundle map applied to σ; it is not automatic that a general tensorial connection term vanishes on this infinite-dimensional space. If (1.4.27) fails, Ker B_1 ⊄ Ker(B_2A_1), the pre-complex (1.4.23) is not an elliptic pre-complex, and Theorem 1.15's cohomological solvability criterion for the linearized Riemann problem is unsupported. This is the most concrete load-bearing gap among the advertised applications, since it is a condition whose verification is asserted rather than demonstrated.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a Hodge-type theory for sequences of Douglas--Nirenberg systems on manifolds with boundary, introducing 'elliptic pre-complexes' and 'adapted Green systems'. The central result, Theorem 1.7 (general version) and Theorem 3.14, asserts that every elliptic pre-complex can be uniquely 'lifted' to a cochain complex of adapted Green systems, with Hodge decompositions whose harmonic spaces reduce to explicit kernels built from the original operators, e.g. H_N^{α+1}=Ker(A_{α+1},A^*_α,B^*_α) and H_D^{α+1}=Ker(A_{α+1},A^*_α,B_{α+1}). The paper then presents geometric applications to exterior covariant derivatives, Killing and Hessian equations, and linearized Riemann and Ricci curvature equations, with the latter two involving substantial deferred verifications. The abstract theory is structured around pseudodifferential boundary-value calculus, overdetermined ellipticity, and order-reduction properties replacing the rigid complex condition.","tokens_in":67354,"tokens_out":4251,"duration_ms":46028,"significance":"If the framework is correct, it provides a genuinely broad generalization of Hodge theory that applies to overdetermined geometric boundary-value problems without restrictive structural assumptions on the background geometry. The explicit formulas for the cohomology spaces are valuable because they are expressed solely in terms of the original operators and boundary conditions. The manuscript is careful with definitions and provides a staged proof of the main abstract theorem, which is a strength. The Riemann curvature example, if fully verified, would be a notable advance, and the Killing/Hessian applications are well motivated. However, the advertised applications are not all self-contained: the Riemann example depends on an unproved condition, and the Ricci theorem is deferred to a separate paper. Thus the central theory appears defensible, but the completeness of the examples is not yet at the level claimed.","major_comments":[{"comment":"The verification that the Riemann curvature sequence (1.4.23) is an elliptic pre-complex rests on the identity B_2 A_1 = 0 on Ker B_1. The computation (1.4.26) proves this for Γ=0, but for a nonzero tensorial connection term Γ one must impose P_t Γ = 0 on Ker(P_tt, D A_g), exactly as stated in (1.4.27). The text asserts this condition is 'satisfied trivially' for tame connections on M_M × C^{2,2}_M, but no proof is supplied. Since Ker(P_tt, D A_g) contains the infinite-dimensional space of infinitesimal diffeomorphisms fixing the boundary, the vanishing of a general bundle map on this kernel is not automatic. A proof of (1.4.27) for the stated class of connections is needed; without it, the order-reduction property for A_1 is not established and Theorem 1.15 is unsupported.","section":"Section 1.4.4, Eq. (1.4.27)"},{"comment":"Theorem 1.16, the main result on the Ricci curvature equations, is stated as a theorem proved in a separate paper [Led25], and the present manuscript contains no proof of the disrupted elliptic pre-complex construction for this example. Since the abstract and introduction explicitly list the Ricci curvature equations as a demonstrated application, the manuscript should either include the necessary proof or clearly mark Theorem 1.16 as an announced result from a companion paper rather than as a theorem proved here.","section":"Section 1.4.5, Theorem 1.16"},{"comment":"The cohomological formulations in Theorems 1.8 and 1.9 are stated for the lifted operators D_α, but the solvability conditions of the original boundary-value problems would require translating the conditions D_{α+1}Θ=0 and the orthogonality to H_N^{α+1} or H_D^{α+1} into conditions expressed only through the original data. The refinement (1.2.25) gives the orthogonality spaces explicitly, but the conditions D_{α+1}Θ=0 and B_{α+1}Θ=0 are not obviously expressible purely in terms of A_•, B_• without further explanation. Please clarify how the main theorems apply to problems involving the original operators rather than only the lifted ones.","section":"Section 1.2.4, Theorems 1.8 and 1.9"}],"minor_comments":[{"comment":"The spelling of the relevant systems is inconsistent: 'Douglas–Nirenberg' appears in some places, 'Douglas-Nirenberg' in others, and 'Douglis–Nirenberg' in the abstract. Please standardize the spelling.","section":"Throughout"},{"comment":"The word 'rudienmantry' appears and should be 'rudimentary'.","section":"Section 2.2, text after Definition 2.11"},{"comment":"The notation ord(A_α) is used for the order of a differential operator in the prototypical setting, while later the paper works with sharp tuples and corresponding orders for Douglas–Nirenberg systems. A brief notational bridge between these two usages would improve readability.","section":"Section 1.2.1, Definition 1.1"},{"comment":"The comparison with elliptic quasicomplexes is informative and much needed, but the displayed formula (3.2.32) and the surrounding discussion would benefit from explicitly stating which of the L^2-orthogonal properties carry over and which do not, since the text notes that the decompositions are not necessarily L^2-orthogonal.","section":"Section 3.2.5"}],"recommendation":"major_revision","confidential_remarks":"The paper's abstract and introduction promise more than the body delivers for the Ricci equations, and the Riemann example has a specific unverified condition that is load-bearing for Theorem 1.15. These are fixable either by adding proofs or by reframing the affected claims as announcements, so I do not see a need for rejection. I would ask that the verification of (1.4.27) be either supplied or the theorem be weakened accordingly, and that the status of Theorem 1.16 be made explicit in the introduction and abstract."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear X,\n\nThis is a serious paper. The central idea—replacing the cochain condition A_{α+1}A_α = 0 with the order-reduction property, that the composition has order no larger than A_α's—is a genuine relaxation, and the lifting theorem is a real step beyond elliptic quasicomplexes. The cohomology identification H_N^{α+1} = Ker(A_{α+1}, A*_α, B*_α) is clean and is new relative to [KL25]; the Dirichlet version is genuinely new. The overall construction via adapted Green systems and Douglas–Nirenberg weights is careful, not hand-wavy.\n\nThe geometric examples give credit: exterior covariant derivatives for arbitrary connections, the Killing/Hessian complexes on arbitrary backgrounds, and the Riemann curvature linearization are the right test cases. The paper is honest about what is deferred: the Ricci theorem is outsourced to [Led25], and several Lopatinskii–Shapiro computations are announced rather than shown.\n\nThe soft spot is the Riemann example, and it is load-bearing. For the Dirichlet pre-complex (1.4.23), the order-reduction step at α=1 needs B_2A_1 = P_t DΓRm_g to vanish on Ker B_1. Equation (1.4.26) gives that only when Γ=0. For a nonzero connection term, the text imposes (1.4.27): P_tΓ = 0 on Ker(P_tt, DA_g), and says this holds trivially for tame connections. That is not trivial. The kernel of (P_tt, DA_g) contains all infinitesimal diffeomorphisms fixing the boundary, and P_tΓ is a bundle map; there is no evident reason it annihilates that infinite-dimensional space. No proof is supplied. Until this is verified, Theorem 1.15's solvability criterion for the linearized Riemann problem is unsupported, and that is one of the two headline applications. The gap doesn't kill the abstract framework, but it is a concrete hole in the advertised scope.\n\nThe math that is actually carried out looks coherent, and the citation pattern is fine—[KL25] is acknowledged and relevant. No code or formal verification, but that's not the norm in this field.\n\nMy recommendation: send to a serious referee. The framework is worth a careful check, and the referee should be asked to focus on (1.4.27) and on the deferred ellipticity claims. I would not desk-reject; I would also ask the author to move the Ricci material into this paper or clearly state which parts are established.","headline":"A serious framework that relaxes the cochain condition, but the Riemann example carries an unproved condition that needs to be supplied before the headline application is convincing.","tokens_in":67833,"tokens_out":2069,"would_cite":true,"duration_ms":20732,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["58J10","35N25","53C20","58A14","35S15","53C21","31C12"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every elliptic pre-complex can be lifted to a genuine cochain complex, and the cohomology of the lift is computed from the original operators alone, so solvability of many linearized geometric boundary-value problems reduces to a…","keywords":["elliptic pre-complex","Hodge theory","overdetermined boundary-value problems","Douglas-Nirenberg systems","order-reduction property","lifted complex","Killing equation","Riemann curvature equations"],"falsifier":"Take a compact Riemannian manifold with boundary and any connection $\\nabla$ whose curvature $R^\\nabla$ is nonzero. If $d^\\nabla\\oplus\\delta^\\nabla\\oplus P_t$ is overdetermined elliptic, the paper's Theorem 1.11 predicts that the problem $d^\\nabla\\psi=\\omega$, $\\delta^\\nabla\\psi=0$, $P_t\\psi=0$ is solvable iff $d^\\nabla\\omega=0$, $P_t\\omega=0$, and $\\omega$ is $L^2$-orthogonal to $H_D^{\\alpha+1}=\\mathrm{Ker}(d^\\nabla,\\delta^\\nabla,P_t)$. Producing one smooth $\\omega$ satisfying those three conditions for which no solution $\\psi$ exists would refute the central claim; equally, checking the claimed equality $H_D^{\\alpha+1}=\\mathrm{Ker}(A_{\\alpha+1},A^*_\\alpha,B_{\\alpha+1})$ on a small finite-element example would settle it.","tokens_in":66796,"feed_emoji":"📐","tokens_out":7115,"duration_ms":68257,"temperature":0.7,"pith_summary":"The paper develops a generalized Hodge theory for linearized boundary-value problems that do not form a cochain complex. Its central object is an elliptic pre-complex: a sequence of Douglas–Nirenberg systems tied together by Green's formulas, where the composition of consecutive operators is allowed to be nonzero as long as its order and class drop. The main theorem shows that every such pre-complex can be lifted to a genuine complex of adapted Green systems, and that the cohomology of the lifted complex is exactly the kernel of an overdetermined boundary-value problem written only in terms of the original operators. If the theorem is right, the solvability and uniqueness conditions for the original linearized problem—Killing, Hessian, or linearized Riemann curvature equations on a general compact manifold with boundary—reduce to a compatibility condition and an explicit orthogonality condition to a finite-dimensional space.","feed_headline":"Every elliptic pre-complex lifts to a true cochain complex","feed_subtitle":"Solvability of Killing, Hessian, and curvature boundary problems reduces to compatibility plus an explicit orthogonality condition.","key_machinery":"The load-bearing objects are elliptic pre-complexes and their lifted complexes. An elliptic pre-complex is a diagram of adapted Green systems—Douglas–Nirenberg pseudodifferential boundary systems—interacting through a generalized Green formula, with two extra conditions: the combined systems $A_\\alpha\\oplus A^*_{\\alpha-1}\\oplus B^*_{\\alpha-1}$ (Neumann) or $A_\\alpha\\oplus A^*_{\\alpha-1}\\oplus B_\\alpha$ (Dirichlet) are overdetermined elliptic, and the order-reduction property holds, meaning consecutive compositions have lower order and class than nominally expected. The order-reduction property replaces the strict cochain condition $A_{\\alpha+1}A_\\alpha=0$, and typically arises from linearizing geometric symmetries and constraints. The lifting construction uses auxiliary decompositions, built inductively, whose projections lie in the Boutet de Monvel calculus and are of order and class zero; these corrections preserve overdetermined ellipticity and give the explicit cohomology formulas.","core_discovery":"The paper claims, in Theorems 1.7 and 3.14, that every elliptic pre-complex $(A_\\bullet)$ induces a uniquely characterized lifted complex $(D_\\bullet)$ of adapted Green systems with $R(D_\\alpha)\\subseteq N(D_{\\alpha+1})$ in the Neumann case or $R(D_\\alpha;B_\\alpha)\\subseteq N(D_{\\alpha+1})$ in the Dirichlet case, and that $D_\\alpha$ differs from $A_\\alpha$ by a zero-order, zero-class Green operator. The cohomology spaces of the lifted complex coincide with kernels of the original systems: $H_N^{\\alpha+1}=\\mathrm{Ker}(A_{\\alpha+1},A^*_\\alpha,B^*_\\alpha)$ and $H_D^{\\alpha+1}=\\mathrm{Ker}(A_{\\alpha+1},A^*_\\alpha,B_{\\alpha+1})$. Consequently, for $\\eta$ in the appropriate section space, the system $A_\\alpha\\psi=\\eta$ with homogeneous gauge and boundary conditions is solvable exactly when the next compatibility operator annihilates $\\eta$ and $\\eta$ is $L^2$-orthogonal to the relevant cohomology space; the solution is unique modulo the cohomology at the previous level.","pith_inferences":["The paper leaves implicit a general recipe: any linearized overdetermined geometric equation becomes accessible once one finds compatible boundary operators and verifies order-reduction, suggesting applications to other curvature or gauge problems where no Calabi complex was previously known.","Because the cohomology spaces are expressed through elliptic boundary-value problems, the theorems imply a stability statement under continuous variation: finite-dimensionality and dimension jumps should be controlled by the Fredholm index of those systems, a connection the paper only begins to explore through Euler characteristics.","The disrupted Ricci case hints that a formal weakening of overdetermined ellipticity, controlled at a single segment, can still yield uniqueness and solvability formulas; a testable extension would identify exactly how much ellipticity loss the cohomology reduction can tolerate."],"forward_implications":["For the Killing equation on a compact Riemannian manifold with boundary, a symmetric tensor $\\sigma$ arises as $\\tfrac{1}{2}L_Yg$ exactly when $H_g\\sigma=0$ and $\\sigma$ is $L^2$-orthogonal to $\\mathrm{Ker}(H_g,\\delta_g,P_n^g)$, with uniqueness modulo Killing fields.","For the Hessian equation, $\\sigma=H_gf$ is solvable exactly when $d_g\\sigma=0$ and $\\sigma\\perp\\mathrm{Ker}(d_g,H^*_g,B^*_H)$, up to the affine kernel of $H_g$.","For the linearized Riemann curvature equations with Dirichlet Cauchy data, the obstruction space is $B^2_D(M,g,\\Gamma)=\\mathrm{Ker}((D_\\Gamma\\mathrm{Rm}_g)^*,d_g,P_t)$, so solvability is $d_gT=0$, $P_tT=0$, $T\\perp B^2_D(M,g,\\Gamma)$.","Finite families of elliptic pre-complexes have constant Neumann and Dirichlet Euler characteristics, computed directly from the original operator kernels, and Sobolev versions of the Hodge decompositions hold for all $1