{"id":"332a9924-d9d8-4b55-9521-3399d9922aed","arxiv_id":"2508.20241","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Singular foliations of b^{k+1}-type are equivalent to k-th order foliations and are classified up to isotopy by fundamental group representations into the jet group G_{k,l}, with extension obstructed by a characteristic class.","lead":"The authors classify geometric objects called singular foliations that vanish near a submanifold to a specified order. Their main result ties this classification to representations of the fundamental group, giving moduli spaces and an obstruction to extending structures to higher order.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Local normal form in Prop 3.1 rests on an unverified application of [BBLM20]; if the hypotheses fail, the b^{k+1}/k-jet bijection and the RH classification collapse.","rationale":"The reader's weakest_assumption is exactly the imported [BBLM20] splitting theorem, so I agree. I considered whether Proposition 6.17 makes the concern non-load-bearing; however, as written, Proposition 3.1 is used before Proposition 6.17 and the text does not cross-reference an alternative proof of the normal form. Since the rest of the paper is well supported by detailed arguments (Maurer-Cartan data, topological groupoid fibrations, group-cohomology comparison), I would not reject or mark the paper unverdictable. The appropriate adjustment is to keep the reader's ACCEPT in substance but make it conditional on verifying the splitting-theorem hypotheses or replacing that step with Proposition 6.17. If the check passes, the ACCEPT is justified.","tokens_in":62586,"tokens_out":52290,"duration_ms":624069,"concrete_test":"Check the exact statement used: take the model case M=R^l×R^{n-l}, W={0}×R^{n-l}, A=Tan_k generated by z^I∂_{z_i} and ∂_{x_a}, and L={x=0}. Verify whether [BBLM20]'s splitting theorem applies to (A,L) at points z≠0; in particular, determine whether the theorem's 'transversal' requires T_xM=F_x⊕T_xL or only T_xM=F_x+T_xL. If the former, Proposition 3.1 lacks a hypothesis; if the latter, the application is sound. Alternatively, reorder the proof: derive the local normal form directly from Proposition 6.17 (for contractible W), which avoids [BBLM20] entirely.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central bijection (Theorem 6.19(1)↔(2)) and the codimension-1 classification both use the local model from Definition 1.5. Proposition 3.1(a) obtains it by applying the splitting theorem for singular foliations [BBLM20] to A=Tan_k(W,F) and an auxiliary leaf L of a foliation C complementary to W. The proof does not verify the hypotheses of the theorem: at points off W, A is all of TM, so L is contained in the regular leaf and is not transverse to A in the usual sense. If the intended statement requires transversality to all leaves, the theorem does not apply and the proof of the local model is incomplete. Since the local model is used both to show that Tan_k(W,F) is of b^{k+1}-type and to reconstruct a b^{k+1}-foliation from its jet, a failure here would invalidate Theorem 6.19 and, through it, Corollary 8.42 and the extension-class results. I am not asserting that [BBLM20] is false; it is a standard reference and the paper likely intends a leafwise/slice version. But the text should state and check the precise hypotheses. Proposition 6.17 later gives an independent normal form for simply connected W; if that proof is independent of Proposition 3.1, it may repair the gap.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces and classifies singular foliations of b^{k+1}-type, a class of singular foliations whose local model is the module generated by monomials of weight k+1 in the normal directions together with the tangent distribution of a submanifold W. The central result is a bijection between such foliations and k-th order foliations on the k-th order neighbourhood of W, and, after choosing a tubular neighbourhood, with Lie algebroid splittings of the k-th order Atiyah algebroid and flat G_{k,l}-connections on the k-th order frame bundle. Using topological groupoids and holonomy, the authors prove a Riemann-Hilbert correspondence, a topological refinement, and an isotopy classification. In codimension one this yields a homeomorphism between isotopy classes of hypersurface algebroids of b^{k+1}-type and a G^0_{k,1}-quotient of a representation variety of π_1(W). The paper also studies the extension problem from k-th to (k+1)-st order foliations, defining an extension class e(F) and showing it can be viewed as a section of a vector bundle over the k-jet character stack. Several explicit examples, including surface, Heisenberg, mapping-torus, and codimension-two cases, are computed.","tokens_in":62929,"tokens_out":47884,"duration_ms":557210,"significance":"If correct, the paper gives a substantial and coherent classification of a natural class of singular foliations and Lie algebroids, connecting them to jet geometry, Atiyah algebroids, and representation varieties of the fundamental group. The Riemann-Hilbert correspondence for k-th order foliations, the isotopy-level classification, and the extension-class theory are new and likely to be useful for later work on b^k-geometry and Poisson geometry. The paper is careful and detailed: the proofs are largely self-contained once standard splitting, Riemann-Hilbert, and Van Est inputs are accepted, and the examples in Section 7 provide concrete, checkable predictions. The explicit computation of extension classes and the description of the character stack are notable strengths.","major_comments":[{"comment":"The proof applies the splitting theorem for singular foliations [BBLM20] to a leaf L of an auxiliary foliation C complementary to W. As written, the hypotheses of the theorem are not stated or verified. At points of L that are not on W, Tan_k(W,F) is all of TM, so L is contained in a regular leaf and is not transverse to the leaves in the usual geometric sense. This step is load-bearing: it produces the local normal form that identifies Tan_k(W,F) as b^{k+1}-type and is used in the bijection (c) and in Theorem 6.19. I suspect the intended hypothesis is the algebraic transversality condition T_xM = F_x + T_xL for all x∈L, which does hold here (trivially off W, and by complementarity of C on W). The authors should state the precise version of [BBLM20] they use and verify this condition explicitly.","section":"Section 3, Proposition 3.1(a)"},{"comment":"The proof contains the assertion: 'Since V_t|_W=0, and assuming that k ≥ 1, the linear approximation of V_t vanishes, implying that ν(φ_t)=id.' This is false: in the local model, the vector field V = z ∂_y lies in Tan_k(W,F), vanishes on W, but has a nonzero first jet. The conclusion ν(φ_t)=id is nevertheless correct, and the missing justification is that the normal component of any section of Tan_k lies in I^{k+1}TM, so the induced map on the normal bundle is the identity. This step should be corrected, since Theorem 8.7 is used in the proof of Theorem 8.35 to identify the kernel of the Riemann-Hilbert fibration.","section":"Section 8.1.6, Theorem 8.7 proof"}],"minor_comments":[{"comment":"The proof uses the notation i^!A after introducing γ^!A; the former should presumably be γ^!A. Please fix this typo.","section":"Proposition 1.4 proof"},{"comment":"'Levy decomposition' should be 'Levi decomposition'.","section":"Section 6.2.1"},{"comment":"'cohology' should be 'cohomology'.","section":"Theorem 9.3"},{"comment":"The notation Rep(Frk(E),g) is used, but the framing is denoted φ elsewhere; use a consistent symbol, e.g. Rep(Fr_k(E),φ).","section":"Theorem 8.35 and Section 8.2.4"}],"recommendation":"major_revision","confidential_remarks":"The paper is strong and the main results are likely correct, but the two proof issues above affect load-bearing steps. Both are local and fixable, so I expect acceptance after revision. The overlap with [Fra24] and [FLG24] is openly discussed, and the isotopy-level classification and extension-class section go substantially beyond those works. No concerns about attribution or scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"One thing to know before you look at it: this is a serious classification paper, and the main result holds up. The authors introduce b^{k+1}-type singular foliations for a submanifold of arbitrary codimension, show they are equivalent to k-th order foliations (jets of distributions involutive up to order k), and then, after choosing a tubular neighbourhood, to Lie algebroid splittings of the k-th order Atiyah algebroid and to flat G_{k,l}-connections on the k-th order frame bundle. From there they get a Riemann-Hilbert correspondence, both topological and up to isotopy, and an extension obstruction living in H^2(W, Sym^{k+1}(ν^*)⊗ν). The classification of hypersurface algebroids in terms of a character variety is the concrete payoff, and the examples—the surface quadric, the Heisenberg obstruction, the mapping torus—are explicit and checkable. The paper is honest about the overlap with Francis and with Fischer–Laurent-Gengoux; it does not oversell its set-theoretic part.\n\nNow the soft spot. Proposition 3.1, which gives the local normal form for Tan_k(W,F), invokes the splitting theorem [BBLM20] without checking its hypotheses. Your stress-test note is right that at points off W, the leaf L of a complementary foliation lies inside the regular leaf M\\W, and the intersection A∩TL jumps from TL to 0 as you hit W. I do not think this is fatal: the transversality condition that [BBLM20] actually needs—T_xM = A_x + T_xL at all points of L—does hold, and the jump in the intersection is exactly the kind of thing that splitting theorem is designed to handle. But the text should say so. The same applies to the proof of Proposition 1.4 for hypersurface algebroids: it cites the Lie algebroid splitting theorem without stating the precise transversality condition. This is a rigor-and-readability fix, not a collapse.\n\nIf you want to be sure, check the application of the splitting theorem in the k-jet bijection of Theorem 6.19; if the local model were false, the bijection and the subsequent RH classification would fall. I don't see evidence of that. The rest of the paper—the Maurer-Cartan description, the holonomy logic via Fr_k(E), the comparison of extension classes through Van Est—is coherent and detailed. There is a lot of material here, and it is well organized.\n\nWho is this for: anyone working on b^k geometry, singular foliations, or log-symplectic Poisson geometry. It is not light reading, but the character-variety viewpoint will probably become the standard way to think about these objects. I would send it to a serious referee—ideally someone who can check the [BBLM20] application and the Van Est comparison—rather than desk reject.","headline":"The b^{k+1}-type classification holds together; the only real soft spot is the unverified use of the [BBLM20] splitting theorem in the local normal form, which is fixable and not fatal.","tokens_in":63405,"tokens_out":8222,"would_cite":true,"duration_ms":93820,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53D17","57R30","58A20"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that b^{k+1}-type singular foliations are encoded by k-th order foliations — distributions involutive up to order k — and that, for a closed connected hypersurface, their isotopy classes form a character variety of the fund","keywords":["b^{k+1}-type singular foliations","hypersurface algebroids","k-th order foliations","Lie algebroids","Riemann-Hilbert correspondence","Atiyah algebroid","character variety","extension class"],"falsifier":"On W = S^3 (trivial π_1) with trivial normal bundle, Corollary 8.42 predicts a single isotopy class of b^{k+1}-type algebroids — all isotopic to Scott's bundle. Producing two non-isotopic such algebroids would refute the classification. Or, in the genus-g surface example with trivial line bundle, the paper predicts a b^4-algebroid extends to b^5 exactly when Σ(x_i z_i − y_i w_i) = 0 (Example 7.1); computing the extension class of a parameter choice with non-zero value and finding an extension anyway would refute Corollary 4.7. A third check: Corollary 9.9 says ∫_{S^2} e(F) = 0 for every sphere","tokens_in":62507,"feed_emoji":"🌀","tokens_out":20237,"duration_ms":170810,"temperature":0.7,"pith_summary":"This paper classifies singular foliations that are tangent to a submanifold W to order k — the b^{k+1}-type foliations, which include the logarithmic b-tangent structures used across geometry. The central claim is that such a foliation is fully encoded by its k-jet along W: a k-th order foliation, a distribution on the k-th order neighbourhood of W that is involutive up to order k. After choosing a tubular neighbourhood, the data is equivalently a flat connection on the bundle of k-jets of frames of the normal bundle, so the classification becomes a Riemann-Hilbert correspondence: foliations up to isomorphism are representations of the fundamental group of W in the jet group G_{k,l}, up to conjugation. For a closed connected hypersurface W, isotopy classes of the resulting b^{k+1}-type algebroids form the character variety Hom_{ν_W}(π_1(W), G_{k,1})/G^0_{k,1}, with Scott's b^{k+1}-tangent bundles as the trivial representation. The paper also resolves the extension problem: a k-th order foliation extends to order k+1 exactly when a characteristic class e(F) in H^2(W, Sym^{k+1}(ν*)⊗ν) vanishes, and this class ties together the character varieties of different orders.","feed_headline":"Fundamental-group representations classify b^{k+1}-type foliations","feed_subtitle":"The whole variety of b^{k+1}-types is a character variety, and one cohomology class controls when k-jets extend.","key_machinery":"The load-bearing objects are: (1) the k-th order Atiyah algebroid at_k(E), the Lie algebroid of k-jets of projectable vector fields on a tubular neighbourhood E → W, whose Lie algebroid splittings are exactly the k-th order foliations; (2) the jet group G_{k,l} = J^k Diff(R^l, 0), the group of k-jets of diffeomorphisms of R^l fixing the origin, with Levy decomposition K_{k,l} ⋊ GL(R^l); (3) the k-th frame bundle Fr_k(E), principal G_{k,l}-bundle of k-jets of frames of the normal bundle, whose Atiyah algebroid is at_k(E) — so each k-th order foliation is a flat connection and holonomy is a representation of π_1(W); (4) the extension class e(F), the obstruction to extending to order k+1, given","core_discovery":"Singular foliations of b^{k+1}-type — vector fields tangent to W to order k — correspond one-to-one with k-th order foliations (distributions involutive up to order k). With a tubular neighbourhood, these are equivalently Lie algebroid splittings of the k-th order Atiyah algebroid at_k(E) or flat connections on the k-th frame bundle Fr_k(E). Holonomy yields a Riemann-Hilbert correspondence: classes of foliations are conjugacy classes of ρ: π_1(W) → G_{k,l}; for a closed connected hypersurface, isotopy classes form Hom_{ν_W}(π_1(W), G_{k,1})/G^0_{k,1}, with Scott's bundles as the trivial class. Extension to order k+1 is obstructed by an extension class e(F) in H^2(W, Sym^{k+1}(ν*)⊗ν).","pith_inferences":["Since G_{k,l} has the homotopy type of GL(R^l), the entire b^{k+1}-moduli space is assembled from cohomology classes of W with coefficients in symmetric powers of the normal bundle; the smooth structure of M near W enters only through the isomorphism class of the normal bundle and the flat connection it induces — a much coarser invariant than the full foliation data.","Lemma 10.3 implies the linear model (the trivial representation) lies in the closure of every component of the character variety; read as a deformation principle, every b^{k+1}-type algebroid degenerates to Scott's bundle, which could make the trivial representation a universal starting point for constructing new structures by unfolding along paths in the moduli space.","A testable boundary case: run the bijection for non-compact W, manifolds with corners, or normal bundles without any flat connection — cases where the imported splitting theorems' hypotheses are strained; the dictionary should break exactly where the local monomial model fails.","The extension class viewed as a section of a vector bundle over the character stack makes extendability a zero-set condition; one could seek analogous sections for higher-order deformations of other singular geometric structures built on b-type tangent bundles, transferring the paper's obstruction theory wholesale."],"forward_implications":["Hypersurface algebroids of b^{k+1}-type — the generalizations of logarithmic and b-tangent bundles used in index theory, Poisson geometry, and integrable systems — are classified by the topology of W: up to isotopy they are exactly the points of Hom_{ν_W}(π_1(W), G_{k,1})/G^0_{k,1}.","Scott's b^{k+1}-tangent bundles, whose definition requires a choice of defining-function jet, all represent the same isotopy class; the auxiliary choice is immaterial.","The extension problem is decidable from a cohomology class: a k-th order foliation extends to order k+1 if and only if e(F) = 0, and when an extension exists the space of extensions is a torsor for H^1_∇(W, Sym^{k+1}(E*)⊗E) modulo the stabilizer of the holonomy.","Because e(F) is aspherical (its integral over every map of S^2 into W vanishes), the existence of higher-order extensions is governed by H^1 and H^2 of W with local coefficients — a cohomological, not homotopical, constraint."],"supporting_citations":[{"why":"Splitting theorem for singular foliations; supplies the local normal form x^I ∂_{x_i} on which the bijection between b^{k+1}-type foliations and k-th order foliations rests (Proposition 3.1).","marker":"[BBLM20]"},{"why":"Introduced the b^{k+1}-tangent bundles that hypersurface algebroids of b^{k+1}-type generalize; the classification identifies them with the trivial representation and shows they are all isotopic.","marker":"[Sco16]"},{"why":"Foundational theory of singular foliations (sheaves of vector fields, holonomy groupoids) that frames the paper's objects and their local-freeness analysis.","marker":"[AS09]"},{"why":"Splitting theorems for Lie algebroids giving the local model x^{k+1}∂_{x_1} + ∂_{x_2} + ... for hypersurface algebroids (Proposition 1.4).","marker":"[Duf01, Fer02, Wei00, BLM19]"},{"why":"Supplies the Van Est map used to identify the differential-form extension class e(F) with the group-cohomology extension class for lifting holonomy (Proposition 9.8).","marker":"[WX91]"},{"why":"Lie's second theorem; integrates algebroid splittings into holonomy morphisms from the fundamental groupoid, the engine of the Riemann-Hilbert functor.","marker":"[MX00, MM02]"},{"why":"Used in Section 9.3 to show the Van Est image of the groupoid 2-cocycle is exactly the extension class e(F).","marker":"[Cra03]"}],"fun_headline_variants":["b^{k+1}-type foliations equal k-th order jets","Riemann-Hilbert maps foliation jets to character stacks","Holonomy classifies b^{k+1}-type foliation jets","Foliation jet extension obstructed by a cohomology class"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The dictionary rests on one imported splitting theorem: locally around W, every b^{k+1}-type foliation is isomorphic to the model spanned by monomials x^I ∂_{x_i} of weight k+1 in the normal coordinates together with the tangent directions of W. If this normal form fails for some germ, the bijection with k-th order foliations collapses.","fun_headline_variants_meta":{"raw":{"variants":["b^{k+1}-type foliations equal k-th order jets","Riemann-Hilbert maps foliation jets to character stacks","Holonomy classifies b^{k+1}-type foliation jets","Foliation jet extension obstructed by a cohomology class"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000842,"raw_usage":{"total_tokens":3537,"prompt_tokens":808,"completion_tokens":2729,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":552,"completion_tokens_details":{"reasoning_tokens":2654}},"tokens_in":552,"tokens_out":2729,"duration_ms":21963,"temperature":1.0,"reasoning_tokens":2654,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T15:13:48.606301+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On W = S^3 (trivial π_1) with trivial normal bundle, Corollary 8.42 predicts a single isotopy class of b^{k+1}-type algebroids — all isotopic to Scott's bundle. Producing two non-isotopic such algebroids would refute the classification. Or, in the genus-g surface example with trivial line bundle, the paper predicts a b^4-algebroid extends to b^5 exactly when Σ(x_i z_i − y_i w_i) = 0 (Example 7.1); computing the extension class of a parameter choice with non-zero value and finding an extension anyway would refute Corollary 4.7. A third check: Corollary 9.9 says ∫_{S^2} e(F) = 0 for every sphere","supporting_citations":[],"review_version":1}