{"id":"9efdfe42-e045-49c5-ae71-b114dbf3c5ba","arxiv_id":"1908.07261","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For pairs of singular distributions on Riemannian manifolds, the paper derives a P-divergence theorem and an integral formula relating mixed scalar curvature to second fundamental forms, integrability tensors, and mean curvature vectors.","lead":"This paper proves a divergence theorem and an integral formula for pairs of singular distributions, which are subspaces that change dimension from point to point, defined as images of smooth matrix-valued maps. The result generalizes a classical formula for regular foliations and may yield splitting theorems for such geometric structures.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 2 and Proposition 4 omit the ‘allowed’ condition that the proof of identity (19) actually uses; self-adjointness plus div(P^2)=0 does not imply it, so the central formula is not proven as stated.","rationale":"The reader identified the allowed condition as the weakest assumption, and this is exactly where the proof has a concrete gap. Section 2 begins by assuming P is allowed, and the proofs of Proposition 1 and Lemma 3 rely on the forms b(i)_j vanishing; Proposition 4 and Theorem 2 do not restate this condition. The explicit self-adjoint example with P^2=I shows that allowedness is not a consequence of the stated hypotheses, so the theorem is under-specified. Because the same example still satisfies (19), the issue could be either a missing hypothesis or a repairable proof gap; the proposed z-dependent reflection family is designed to discriminate between those two cases. Thus the reader’s CONDITIONAL verdict is appropriate: the central claim needs either an additional hypothesis or a strengthened proof.","tokens_in":15549,"tokens_out":36811,"duration_ms":350580,"concrete_test":"On flat T^3 with orthonormal frame V=(cos z, sin z, 0), U=(-sin z, cos z, 0), E=∂_z, define P2=V⊗V and P1=A∘(U⊗U+E⊗E), where A is the z-dependent reflection A=cos(2kz)(U⊗U−E⊗E)+sin(2kz)(U⊗E+E⊗U). Then P1,P2 are self-adjoint, P1P2=0, and P^2=I, so div(P^2)=0; for k≠0 the Definition-3 forms b(1)_1 fail to vanish. Compute both sides of identity (19) for k=1 by symbolic/numeric differentiation in the flat metric. If LHS−RHS is nonzero, Theorem 2 is false as stated; if it vanishes identically in k, the allowed condition is unnecessary for the formula, and the proof, not the statement, must be repaired.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The load-bearing step is Proposition 4, identity (19). Its proof traces the Codazzi equation (4), which is Proposition 1 and assumes P = P1 + P2 is allowed (Definition 3), and Lemma 3’s derivation of (15a)–(15d) explicitly uses b(1)_2(e_t,e_t) = 0. Theorem 2 then integrates (19) via Theorem 1. Thus the theorem as printed has a hidden hypothesis: the pair must be allowed. Self-adjointness and div(P^2)=0 do not imply allowedness. For example, on flat T^3 with V=(cos z, sin z, 0), U=(-sin z, cos z, 0), E=∂_z, set P1=U⊗E+E⊗U and P2=V⊗V. Then P1,P2 are self-adjoint, P1P2=0, and P^2=I, so (24) holds, yet b(1)_1(U,U)=V≠0, violating Definition 3. In this constant-reflection example the two sides of (19) happen to cancel, so the example shows the proof does not cover the stated generality rather than disproving the formula outright. The correct fix is either to add ‘allowed’ to the statements of Proposition 4 and Theorem 2, or to prove (19) under a weaker condition directly. As printed, the central claim rests on an unstated premise.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript develops tools for singular distributions defined as images of smooth endomorphisms on a Riemannian manifold. Section 1 introduces structural tensors for a pair (P1,P2), defines an 'allowed' condition on the pair, and derives a Codazzi-type equation (4). Section 2 proves a P-divergence theorem (Theorem 1) under the condition div(PP*)=0, together with an open-manifold analogue (Proposition 3). Section 3 uses these ingredients to prove the pointwise identity (19) relating divP(H1+H2) to the mixed scalar curvature and to P-squared norms of the second fundamental forms, integrability tensors, and mean curvature vectors, and then integrates this identity to obtain an integral formula (Theorem 2) that generalizes Walczak's regular formula. Theorems 3–5 give splitting and Liouville-type consequences. The main issue is that Proposition 4 and Theorem 2 omit the 'allowed' condition needed in the proofs of Proposition 1 and Lemma 3; as printed, the central formula is not proven under the stated hypotheses.","tokens_in":15825,"tokens_out":12860,"duration_ms":129138,"significance":"If the missing hypothesis is supplied, the paper delivers a clean, parameter-free generalization of Walczak's integral formula to singular distributions, together with a useful new P-divergence theorem. The analytic derivations are self-contained once the endomorphism representation of singular distributions from [12,13] is granted, and there is no circular reasoning or fitted parameters. The examples involving an Einstein product metric, almost contact structures, and f-structures are relevant and illustrate the scope. Because (19) is an explicit identity, the result is checkable in concrete examples. However, the omission of the 'allowed' condition from the main statements is a load-bearing gap, so the printed central theorem is not established as stated.","major_comments":[{"comment":"The identity (19) is proved by tracing the Codazzi equation (4) and using Lemma 3. Proposition 1 states (4) only under the assumption that P = P1 + P2 is allowed (Definition 3), and the proof of Lemma 3 explicitly invokes b(1)_2(e_t,e_t)=0, which is one of the allowed conditions. Neither Proposition 4 nor Theorem 2 states that (P1,P2) is allowed; they assume only self-adjointness and, in Theorem 2, div(P^2)=0. These hypotheses do not imply the allowed condition. For example, on flat T^3 with orthonormal U=(-sin z, cos z, 0), V=(cos z, sin z, 0), E=∂_z, set P1=U⊗E+E⊗U and P2=V⊗V. Then P1 and P2 are self-adjoint, P1P2=0, and P^2=I, so (24) holds, yet b(1)_1(U,U)=V≠0, contradicting Definition 3. In this example the two sides of (19) happen to cancel, so the example shows a proof gap rather than a counterexample to the formula, but it demonstrates that the stated hypotheses do not cover the proof. The fix is to add 'allowed' to the statements of Proposition 4 and Theorem 2 (and to the later theorems that rely on them) or to prove (19) under a weaker condition.","section":"§3, Proposition 4 and Theorem 2"},{"comment":"The splitting theorems inherit the missing allowed condition because they cite Proposition 4; hence they are not proven as stated. In addition, the proof of Theorem 3 applies Proposition 3 leaf-by-leaf, but the stated hypothesis div(P1^2)=0 is a global condition on M, and the paper does not explain how the leaf-restricted divergence condition required by Proposition 3 is obtained on the singular leaves. Please add the allowed condition to all results that use Proposition 4 and justify the leaf-by-leaf application of Proposition 3.","section":"§3, Theorems 3–5"}],"minor_comments":[{"comment":"In Eq. (20), the second line contains the term ⟨P2∇P1esP1es, ∇P1etP2et⟩, which appears to be a typo for ⟨P2∇P1esP1es, ∇P2etP2et⟩; Eq. (22) uses the latter expression.","section":"§3, proof of Proposition 4"},{"comment":"Definition 3 introduces forms b(i)_j with i,j∈{1,2} but says only 'the bilinear forms b(i)_1 and their dual b(i)_2'; the domain and codomain of each of the four forms should be spelled out explicitly.","section":"§1, Definition 3"},{"comment":"Lemma 1 and Proposition 1 use the adapted metric condition (1), but neither statement explicitly includes (1) as a hypothesis; this should be stated to avoid ambiguity.","section":"§1, Lemma 1 and Proposition 1"},{"comment":"In part (b) of Example 4, the text says 'Similarly to point b)' but clearly refers to point a); also, H is used there for the mean curvature vector of f(TM) without being defined in that example.","section":"§2, Example 4"},{"comment":"In Example 3, the coordinates are listed as (x,y,z,u,v), but the Christoffel symbol indices are said to range over {1,...,6} with (x1,...,x5); the index set should be {1,...,5}.","section":"§2, Example 3"},{"comment":"The proofs refer several times to 'underlined terms' (e.g., in the computation of (15a)), but no underlining is visible in the text; use equation labels or another visible device.","section":"§3, Lemma 3"}],"recommendation":"major_revision","confidential_remarks":"The paper is a credible extension of known foliation formulas, and the gap concerning the 'allowed' condition appears fixable by adding the missing hypothesis to the affected statements, rather than by a fundamentally new argument. I do not see grounds for rejection, but the central theorem as printed is under-specified, so the revision should be checked carefully."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The paper's headline result, Theorem 2, is not proven as stated. The proof of Proposition 4 uses the 'allowed' condition from Definition 3 through Lemma 3, but neither Proposition 4 nor Theorem 2 states it. The stress-test example is convincing: on flat T^3 with U=(-sin z, cos z,0), V=(cos z, sin z,0), E=∂_z, set P1=U⊗E+E⊗U and P2=V⊗V. These are self-adjoint, P1P2=0, and P^2=I, so div(P^2)=0 holds; yet b(1)_1(U,U)=V≠0, so the pair is not allowed. The theorem's hypotheses are therefore insufficient for the given proof. Whether (19) still holds for all such pairs is open, since the example's two sides happen to cancel, but the paper doesn't settle it.\n\nCredit where due: the P-divergence operator and Theorem 1's divergence theorem are new and self-contained. The Codazzi-type equation (4) for singular distributions is a reasonable extension, and the integral formula (19) does generalize Walczak's regular-case formula. The definitions of structural tensors and P-norms are original, and the regular case P=id checks out. The self-citations [12,13] are used legitimately as a representation tool, not as a hidden assumption.\n\nSoft spots, in proportion. The missing hypothesis is the big one. Theorem 3 applies Proposition 3 on leaves, and the L1 condition on H2 is stated only on leaves; that may be fine but deserves a clarifying sentence. The 'allowed' condition gets few examples; it feels strong, and the reader can't tell how often it holds. The tensor computations are long and unverified by machine, so Lemma 3 and the trace in (23) deserve a careful independent check.\n\nBottom line: a serious technical paper with a fixable flaw. It deserves a serious referee who can put 'allowed' back into the statements or prove (19) under weaker hypotheses. I'd send it to review, not desk-reject.","headline":"A worthwhile but under-specified paper: the main integral formula requires an 'allowed' condition that is missing from the statements of Proposition 4 and Theorem 2.","tokens_in":16329,"tokens_out":4000,"would_cite":false,"duration_ms":37176,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C15"],"pacs":[],"model":"deepseek-v4-flash","headline":"For singular distributions on a closed Riemannian manifold, a generalized divergence theorem turns a combination of mixed scalar curvature and distribution invariants into a vanishing integral.","keywords":["singular distribution","P-divergence","mixed scalar curvature","integral formula","Codazzi equation","second fundamental form","mean curvature","Riemannian manifold"],"falsifier":"Construct a closed Riemannian manifold with a self-adjoint pair P1,P2 such that P=P1+P2 satisfies div($P^{2}$)=0 but the pair fails the allowed condition, and compute both sides of the integral formula directly; a nonzero discrepancy would show the central formula does not hold under the theorem's stated hypotheses. Alternatively, verify whether every pair satisfying those hypotheses is automatically allowed, since the proof would collapse if not.","tokens_in":15337,"feed_emoji":"🧮","tokens_out":6680,"duration_ms":60443,"temperature":0.7,"pith_summary":"The paper extends the classical integral formula that relates mixed scalar curvature to the second fundamental forms, integrability tensors, and mean curvature vectors of two complementary orthogonal distributions, so that it also covers singular distributions whose rank may vary pointwise. It models such distributions as images of the tangent bundle under smooth endomorphisms and introduces a modified divergence operator, the P-divergence, for which a divergence theorem holds under div(PP*)=0. A new Codazzi-type equation for a pair of transverse singular distributions is derived and traced to a pointwise identity expressing the P-divergence of the sum of mean curvature vectors through curvature and norm terms. Integrating that identity on a closed manifold yields the paper's main integral formula. If correct, the result gives a common extension in which the classical formula is recovered when P is the identity endomorphism.","feed_headline":"Singular distributions get an integral curvature formula","feed_subtitle":"A P-divergence theorem and a Codazzi equation make the mixed curvature integral vanish when div(P^2)=0.","key_machinery":"The mechanism that carries the argument is the P-divergence operator div_P X = trace(Y -> P*∇_{PY}X), together with the condition div(PP*)=0, which identifies div_P X with div(PP*(X)) and makes (div_P X)dvol an exact form. The trace of the Codazzi-type equation (4), valid when the pair (P1,P2) is allowed in the sense that certain bilinear forms vanish, produces identity (19): div_P(H1+H2) equals S^P_mix plus the P-norm terms. Theorem 1 then integrates this identity to zero on closed manifolds.","core_discovery":"The central claim is Theorem 2: for self-adjoint endomorphisms P1 and P2 of the tangent bundle of a closed Riemannian manifold, with P=P1+P2 and div($P^{2}$)=0, the integral over the manifold of S^P_mix + ||h1||^2_P + ||h2||^2_P - ||T1||^2_P - ||T2||^2_P - ||H1||^2_P - ||H2||^2_P vanishes. Here S^P_mix is the mixed scalar curvature of the pair of singular distributions D_i=P_i(TM), h_i and T_i are their second fundamental forms and integrability tensors, H_i are their mean curvature vectors, and ||·||_P is the P-weighted norm introduced in the paper. The proof uses the P-divergence theorem together with the traced Codazzi-type equation, which yields the pointwise identity (19). When P is the identity, the formula reduces to the classical integral formula for a pair of complementary orthogonal regular distributions.","pith_inferences":["The 'allowed' condition on (P1,P2) appears to be load-bearing for identity (19), even though Theorem 2 states only self-adjointness and div(P^2)=0; finding a self-adjoint pair that satisfies div(P^2)=0 but violates the allowed condition would test whether the theorem statement needs amendment.","The P-divergence framework likely extends to other integral curvature identities whenever a natural divergence-free endomorphism PP* is available, such as from divergence-free Einstein tensors, as the paper's own example suggests.","Because the P-norm is not necessarily positive for general endomorphisms, comparing sign conditions on S^P_mix with positivity properties of P may yield refined rigidity statements beyond the regular case."],"forward_implications":["On closed manifolds the vanishing integral acts as an obstruction: the mixed scalar curvature and the distribution invariants must balance, so configurations where the signed combination is one-sided cannot occur.","Under sign conditions on S^P_mix, the splitting results in the paper force the singular distributions to be autoparallel, meaning both the second fundamental forms and integrability tensors vanish, which in the regular case gives local product structure.","Setting P to the identity endomorphism recovers the classical integral formula for complementary orthogonal regular distributions, so the new result is an extension of that known statement.","On complete open manifolds, the paper's modification of Stokes' theorem yields Liouville-type conclusions: if the P-divergence of a suitable vector field has constant sign and an integrability condition holds, then it vanishes identically."],"supporting_citations":[{"why":"Supplies the construction of self-adjoint endomorphisms whose images are orthogonal singular distributions, used in the final section.","marker":"[13]"},{"why":"Provides the regular-case Codazzi equation, structural tensors, and mixed scalar curvature framework that the paper generalizes.","marker":"[14]"},{"why":"States the classical integral formula for two orthogonal complementary distributions, which Theorem 2 reduces to when P is the identity.","marker":"[16]"},{"why":"Gives earlier examples of singular distributions of the type used here, motivating the present definitions.","marker":"[12]"},{"why":"Supplies the regular-case Liouville-type result for complete open manifolds that Proposition 3 extends to the P-divergence setting.","marker":"[6]"},{"why":"Provides the result generalized by Theorem 5 for complete Riemannian almost product manifolds.","marker":"[15]"},{"why":"Supplies the exhaustion argument used in Proposition 3 to conclude that a sign-constant P-divergence vanishes.","marker":"[17]"}],"fun_headline_variants":["Mixed curvature integral vanishes for singular distributions under div(P²)=0","Integral formula for singular distributions via P-divergence and Codazzi","Singular distributions: integral identity from traced Codazzi","Generalized integral formula for pairs of singular distributions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof depends on the technical 'allowed' condition for the pair (P1,P2), requiring certain bilinear forms to vanish, although the printed version of Theorem 2 lists only self-adjointness and div($P^{2}$)=0.","fun_headline_variants_meta":{"raw":{"variants":["Mixed curvature integral vanishes for singular distributions under div(P²)=0","Integral formula for singular distributions via P-divergence and Codazzi","Singular distributions: integral identity from traced Codazzi","Generalized integral formula for pairs of singular distributions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000798,"raw_usage":{"total_tokens":3461,"prompt_tokens":845,"completion_tokens":2616,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":461,"completion_tokens_details":{"reasoning_tokens":2547}},"tokens_in":461,"tokens_out":2616,"duration_ms":20542,"temperature":1.0,"reasoning_tokens":2547,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:21:48.837591+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct a closed Riemannian manifold with a self-adjoint pair P1,P2 such that P=P1+P2 satisfies div($P^{2}$)=0 but the pair fails the allowed condition, and compute both sides of the integral formula directly; a nonzero discrepancy would show the central formula does not hold under the theorem's stated hypotheses. Alternatively, verify whether every pair satisfying those hypotheses is automatically allowed, since the proof would collapse if not.","supporting_citations":[{"cited_title":"Popescu and M","cited_arxiv_id":null,"evidence_quote":"Supplies the construction of self-adjoint endomorphisms whose images are orthogonal singular distributions, used in the final section."},{"cited_title":"Rovenski, Foliations on Riemannian Manifolds and Submanifolds , Birkh¨ auser, 1998","cited_arxiv_id":null,"evidence_quote":"Provides the regular-case Codazzi equation, structural tensors, and mixed scalar curvature framework that the paper generalizes."},{"cited_title":"Walczak, An integral formula for a Riemannian manifold with tw o orthogonal complementary distributions, Colloq","cited_arxiv_id":null,"evidence_quote":"States the classical integral formula for two orthogonal complementary distributions, which Theorem 2 reduces to when P is the identity."},{"cited_title":"Popescu and M","cited_arxiv_id":null,"evidence_quote":"Gives earlier examples of singular distributions of the type used here, motivating the present definitions."},{"cited_title":"Caminha, P","cited_arxiv_id":null,"evidence_quote":"Supplies the regular-case Liouville-type result for complete open manifolds that Proposition 3 extends to the P-divergence setting."},{"cited_title":"Stepanov, Liouvile-type theorems for some classes of comp lete Riemannian almost product manifolds and for special mappings of complete Riemannian m anifolds, J","cited_arxiv_id":null,"evidence_quote":"Provides the result generalized by Theorem 5 for complete Riemannian almost product manifolds."},{"cited_title":"Yau, Some function-theoretic properties of complete Riem annian manifolds and their applications to geometry","cited_arxiv_id":null,"evidence_quote":"Supplies the exhaustion argument used in Proposition 3 to conclude that a sign-constant P-divergence vanishes."}],"review_version":1}