{"id":"b42866ca-e05b-4f57-9c2d-41bdf0bf018e","arxiv_id":"1909.01117","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For a singular complete intersection X = X1 ∩ ... ∩ Xr in a complex manifold satisfying a transversality condition, the total Milnor class of X is expressed as a combination of the total Milnor and Schwartz-MacPherson classes of the components Xi.","lead":"This mathematics paper proves formulas that break the total Milnor class of a singular complete intersection into the Chern-Schwartz-MacPherson and Fulton-Johnson classes of its individual component hypersurfaces or local complete intersections. It gives a practical way to compute Milnor classes, which measure how singular a space is, for intersections built from simpler pieces.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Mathematical argument holds, but the transversality hypothesis is even more restrictive than stated—it excludes isolated singularities—and the 'first' novelty claim conflicts with [20].","rationale":"I reviewed the proof chain: Proposition 1.5 and Lemma 2.3 for Fulton-Johnson classes, Proposition 1.8 and Lemma 1.9 for characteristic cycles, Theorem 1.12 from the Parusiński-Pragacz formula, and the product formulas in Section 2. The sign bookkeeping checks out, including the subtle coefficients in Proposition 1.8 where the unspecified 'd' must be the codimension of the stratum, and the excess-intersection geometry in Lemma 1.9 is consistent in homology. No internal inconsistency or false algebraic step was found. The central theorem is correct under its stated hypotheses. The real issue is scope and framing: the transversality assumption is not a minor technicality. It fails for any complete intersection whose singular locus has an isolated point on the diagonal of the product, which excludes many interesting singular examples. The authors do flag this via Remark 2.6, so it is not a hidden flaw, but it deserves a prominent caveat. The second concern is the overclaimed novelty of 'first Verdier-Riemann-Roch formulae' for Schwartz-MacPherson classes, since Schürmann's cited work already provides such a formula in greater generality. This does not affect the truth of the theorem but does affect the paper's contribution claim. The reader's CONDITIONAL verdict is therefore appropriate, and I do not propose changing it.","tokens_in":14803,"tokens_out":58663,"duration_ms":604087,"concrete_test":"Extract Schürmann's Verdier-RR theorem from arXiv:math/0202175 and instantiate it for the diagonal embedding Δ:M→M^r with the constant sheaf on a local complete intersection; if the resulting identity is exactly Theorem 1.12, the word 'first' in the abstract must be removed or qualified. Additionally, attempt to construct a diagonal-transverse Whitney stratification for X1 a quadric cone in P^3 with vertex p and X2 a plane through p yielding two lines: the 0-dimensional stratum at (p,p) makes transversality impossible, confirming that the theorem cannot cover isolated singularities.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 1 depends on Proposition 1.8 and Corollary 1.10, which require a Whitney stratification of the product X1×...×Xr with the diagonal Δ:M→M^r transverse to every stratum. This is not a harmless regularity condition: if X has an isolated singular point p that lies in the singular locus of one component, then (p,...,p) is a singular point of the product, so any Whitney stratification of the product must contain a 0-dimensional stratum at that point. A 0-dimensional stratum has tangent space {0}, so transversality T_pΔ + T_p(stratum) = T_pM^r is impossible (the diagonal is a proper subspace when r>1). Thus Theorem 1 and its Corollaries cannot be applied to complete intersections with isolated singularities, a central case for Milnor classes. The paper acknowledges the necessity of transversality in Remark 2.6, but the abstract's phrasing may mislead readers about the scope. Separately, the abstract claims 'first Verdier-Riemann-Roch type formulae'; Theorem 1.12 for Schwartz-MacPherson classes is a special case of Schürmann's generalized Verdier-RR theorem (arXiv:math/0202175), which the paper itself cites. The novelty should be explicitly limited to the explicit intersection formula and the Fulton-Johnson/Milnor parts.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves product formulas for the total Schwartz-MacPherson and Fulton-Johnson classes of a local complete intersection X = X1 ∩ ... ∩ Xr in a compact complex manifold, under a transversality hypothesis on Whitney stratifications of the Xi. The method is to embed the product X1 × ... × Xr into M^{(r)}, apply refined Gysin pullback along the diagonal, and combine known product formulas (Kwiecinski for cSM, Ohmoto-Yokura for Milnor classes) with Verdier-Riemann-Roch type statements for cSM and cFJ. The main result, Theorem 1, gives cSM(X) and cFJ(X) in terms of the classes of the components and the ambient tangent bundle, and hence a simple formula for the total Milnor class M(X). Section 3 derives an Aluffi-type formula and a Parusinski-Pragacz-type formula for line-bundle complete intersections, and a description via global Lê classes.","tokens_in":15065,"tokens_out":20693,"duration_ms":228026,"significance":"If correct, the main formula is a striking and useful simplification: the total Milnor class of a transverse complete intersection is determined by the total classes of its components and the restriction of the ambient tangent bundle. The proof is genuinely external and coherent: it uses refined Gysin maps, exterior product results, constructible-sheaf vanishing-cycle formalism, and Goresky-MacPherson stratified Morse theory, with no fitted parameters or ad hoc normalization. The example in §2.5 is consistent with the formula. The main caveats are that the transversality hypothesis excludes some natural singularity configurations, and that the abstract's novelty claim about Verdier-Riemann-Roch formulae needs to be calibrated against Schürmann's earlier work [20].","major_comments":[{"comment":"The abstract's claim of 'first Verdier-Riemann-Roch type formulae' for cSM is not supported by the cited literature. Schürmann's generalized Verdier-Riemann-Roch theorem for Chern-Schwartz-MacPherson classes (arXiv:math/0202175, reference [20]) already contains a formula of this type, and Theorem 1.12 is essentially the special case of the diagonal embedding. The authors should remove 'first' from the abstract and instead state the genuinely new content: the explicit diagonal factor for cFJ and for Milnor classes, together with the intersection-product formula in Theorem 1.","section":"Abstract and §1 (Theorem 1.12)"},{"comment":"The transversality hypothesis is more restrictive than the Introduction suggests, and this limitation should be stated explicitly. For example, if one component Xi has an isolated singular point p (so its stratum at p is 0-dimensional), and r ≥ 2 with all di ≥ 1, then the product stratum containing (p,...,p) has tangent space {0} × ... × T_pS_i × ...; together with the diagonal tangent space this cannot span T_p M^{(r)} because the other components are proper subvarieties. Thus Theorem 1 does not apply to many, though not all, complete intersections with isolated singularities. The authors should add a remark quantifying which isolated-singularity cases are excluded, and discuss whether a degeneration or limiting argument could extend the formula.","section":"§2, Theorem 1 assumptions and Remark 2.6"},{"comment":"The proof of Proposition 1.8 uses the fact that pulling back a Whitney stratification by a submanifold transverse to all strata yields a Whitney stratification, and that normal slices are preserved by the diagonal embedding. This is standard, but it is the only place where the main transversality assumption is used, so it deserves a brief statement or reference. As written, the sentence 'Since the stratification {Tγ} is transversal to ∆(M), we have that {∆^{-1}(Tγ)} is a Whitney stratification of M' is asserted without justification.","section":"§2, Proposition 1.8"}],"minor_comments":[{"comment":"The displayed formula for cSM(Z1) writes '2c(T P3) - c(T P2)' without fundamental classes; the numerical polynomial only matches after intersecting with [P3] and [P2]. Please clarify the notation to avoid confusing readers.","section":"Example 2.5"},{"comment":"The notation γ^0 and the abbreviation 'S ≠ X' are not defined; in particular, γ^0 should be set to 1. The induction proof is terse and would benefit from spelling out how the three terms combine after the displayed computation.","section":"Corollary 3.5 proof"},{"comment":"The phrase 'all intersections amongst strata in the various Xi are transversal' should be made precise: it should mean that for every choice of strata S_i ∈ S_i, the intersection S_1 ∩ ... ∩ S_r is transversal in M, so that the product stratification is transverse to the diagonal. This would remove any ambiguity about the hypothesis used in §2.","section":"Theorem 1 statement"},{"comment":"There are several typographical errors, including 'intesection' in Theorem 1 and 'stratification' in Proposition 1.8; a careful proofreading pass is recommended.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The mathematical core appears sound under the stated transversality hypothesis, and I found no circularity or fitted parameters. The main concerns are framing: the 'first VRR' statement overclaims priority relative to Schürmann's known theorem, and the paper should be more candid about how restrictive the transversality assumption is for isolated singularities. Both are fixable in revision. I would not recommend rejection on the mathematics itself."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper proves a clean product formula for the total Milnor class of a complete intersection in terms of the component classes and the ambient tangent bundle, under a transversality hypothesis on the diagonal. That formula is the real contribution. The abstract overstates the novelty: the cSM Verdier-RR half is a special case of Schürmann's theorem [20], which the paper itself cites.\n\nWhat's actually new: Theorem 1's formulas for cSM, cFJ, and the Milnor class of X = X1∩...∩Xr under diagonal transversality. The cFJ part appears genuinely new, and the applications to Parusiński-Pragacz and Aluffi type formulas for complete intersections are a useful packaging. The proofs are careful, built from Kwiecinski's product formula, Fulton's refined Gysin formalism, and the Ohmoto-Yokura product formula for Milnor classes; the sign check in Theorem 3.1 for r=2 works, and Example 2.5 is consistent. Remark 2.6 correctly shows the formula fails without transversality.\n\nSoft spots. First, the transversality assumption is not harmless. If X has an isolated singular point, the relevant product stratum at that point has tangent space that, together with the diagonal, cannot span T_pM^r—the tangent spaces of the component strata sum to a proper subspace of T_pM. So the theorem excludes isolated singularities, a central case for Milnor classes. The paper says \"certain transversality conditions\" but does not advertise how restrictive this is; a referee should ask for a precise statement of which singularities are admissible and whether the condition is checkable. Second, the abstract's \"first Verdier-Riemann-Roch type formulae\" is not accurate for cSM; [20] already gives a generalized Verdier-RR for cSM, and the proof here is a special case. The novelty claim should be limited to cFJ and to the explicit intersection formula.\n\nNeither issue sinks the paper. The main theorem is new and the argument is coherent. The authors are candid about the need for transversality in Remark 2.6. The paper is worth a serious referee, but the abstract and the statement of Theorem 1 need revision for precision.\n\nFor whom: anyone working on Milnor classes or characteristic classes of singular complete intersections will find the formula useful in the non-isolated case. I'd bring it to reading group. I'd cite it if I were doing intersection formulas under transversal setups.","headline":"A useful but restricted product formula for Milnor classes of complete intersections, with an abstract that overstates novelty.","tokens_in":15600,"tokens_out":11055,"would_cite":true,"duration_ms":114704,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14C17","55N45","14M10","14B05","32S20"],"pacs":[],"model":"deepseek-v4-flash","headline":"The total Milnor class of a complete intersection is determined by its components and the ambient tangent bundle.","keywords":["complete intersections","Milnor classes","Schwartz-MacPherson classes","Fulton-Johnson classes","Whitney stratifications","singular varieties","Chern classes","Lê classes"],"falsifier":"Recompute the Milnor class of the intersection in Example 2.5, where $X_1=\\{x_0x_1=0\\}$ and $X_2=\\{x_3=0\\}$ in $\\mathbb{P}^4$, by a direct stratified-Morse or vanishing-cycle computation that does not use Theorem 1; the theorem predicts $-H^3$, so any other class in $H_*(X)$ would settle the claim negatively. The paper's Remark 2.6 is the negative control: a smooth quadric and a tangent plane in $\\mathbb{P}^3$ violate the transversality hypothesis, and there the component-wise formula gives zero while the true Milnor class is the class of a point.","tokens_in":14604,"feed_emoji":"📐","tokens_out":12982,"duration_ms":95009,"temperature":0.7,"pith_summary":"This paper proves that, for a complete intersection $X=X_1\\cap\\cdots\\cap X_r$ inside a compact complex manifold, the total Milnor class $\\mathcal{M}(X)$ is determined by the Schwartz-MacPherson and Fulton-Johnson classes of the individual components and by the ambient tangent bundle. Under a transversality condition on the product $X_1\\times\\cdots\\times X_r$, the total Schwartz-MacPherson class and the total Fulton-Johnson class of $X$ are each equal to the corresponding product of component classes, capped with the inverse Chern class of $(TM|_X)^{\\oplus r-1}$. The Milnor class, defined as the signed difference of those two totals, therefore inherits a compact product formula. This matters because Milnor classes generalize the Milnor number to varieties with arbitrary singular locus, and complete intersections have resisted the simple hypersurface treatments. As applications, the paper derives a Parusiński-Pragacz-type formula and an Aluffi-type formula in the line-bundle case, and a description of the Milnor class through global Lê classes.","feed_headline":"Milnor classes of complete intersections reduce to their pieces","feed_subtitle":"When the pieces meet transversely, the total Milnor class is a Chern-corrected product of their characteristic classes.","key_machinery":"The mechanism is the diagonal embedding $\\Delta:M\\to M^{(r)}$ together with the refined Gysin map $\\Delta^!$ from intersection theory. The exterior product of the sections $s_i$ defines a section of $E=p_1^*E_1\\oplus\\cdots\\oplus p_r^*E_r$ on $M^{(r)}$, whose zero scheme is $X_1\\times\\cdots\\times X_r$; pulling back by $\\Delta$ gives $X$. The paper proves Verdier-Riemann-Roch-type identities\n$$\\$\\Delta$^!\\big($c^{{FJ}}$(Z(s))\\big)=c\\big((TM|_{Z(\\$\\Delta$^*s)})^{\\oplus r-1}\\big)\\cap $c^{{FJ}}$(Z(\\$\\Delta$^*s))$$\nand the same for $c^{SM}$, using Fulton's intersection-theoretic properties of regular embeddings and the normal-Morse-index description of conormal cycles. Exterior product formulas for $c^{SM}$ and $c^{FJ}$ then convert $\\Delta^!$ of the product class $c(X_1)\\times\\cdots\\times c(X_r)$ into the intersection product $c(X_1)\\cdots c(X_r)$, and inverting the Chern factor yields the theorem.","core_discovery":"The central discovery is Theorem 1: if $X_i\\subset M$ is the zero scheme of a regular section of a holomorphic vector bundle of rank $d_i$, and if the product $X_1\\times\\cdots\\times X_r$ carries a Whitney stratification to which the diagonal embedding $\\Delta:M\\to M^{(r)}$ is transverse, then for $X=X_1\\cap\\cdots\\cap X_r$,\n$$$c^{{SM}}$(X)=c((TM|_X)^{\\oplus r-1})^{-1}\\cap\\big($c^{{SM}}$(X_1)\\cdots $c^{{SM}}$(X_r)\\big),$$\nand the analogous identity holds for $c^{FJ}$. Consequently the total Milnor class satisfies\n$$\\mathcal{M}(X)=(-1)^{\\dim X}c((TM|_X)^{\\oplus r-1})^{-1}\\cap\\big($c^{{FJ}}$(X_1)\\cdots $c^{{FJ}}$(X_r)-$c^{{SM}}$(X_1)\\cdots $c^{{SM}}$(X_r)\\big).$$\nThe load-bearing point is that no higher correction terms from the singularities of the intersection appear: all ambient data enter only through the inverse Chern factor of the tangent bundle. The paper shows the transversality hypothesis cannot simply be dropped, since tangential intersections can make the component-wise product give the wrong class.","pith_inferences":["Beyond the paper: since the only ingredients are $\\Delta^!$ and product behavior of the two classes, the same diagonal-embedding argument could plausibly yield Verdier-Riemann-Roch-type formulas for other constructible-function characteristic classes, such as Hirzebruch-Milnor classes; the paper does not pursue this.","Beyond the paper: the formula suggests a divide-and-conquer computational strategy—compute hypersurface Milnor classes of the components with existing algorithms, then combine them with the inverse Chern factor—rather than computing the singular locus of the full intersection.","Beyond the paper: the factor $(TM|_X)^{\\oplus r-1}$ makes the formula asymmetric-looking in $r$, and testing associativity (grouping components in different orders) could reveal an identity that the component classes must satisfy; this is an immediate, checkable consequence not stated in the paper.","Beyond the paper: the transversality requirement might be replaceable in many geometric situations by a generic-perturbation or limiting argument, but the paper neither asserts nor rules this out; its Remark 2.6 shows the failure is real for tangential intersections."],"forward_implications":["The total Milnor class of any complete intersection satisfying the transversality condition is computable from the total classes of its components and the Chern classes of the ambient tangent bundle, with no separate analysis of the singular locus of $X$.","When each $X_i$ is the zero scheme of a section of a line bundle, the Milnor class admits a Parusiński-Pragacz-type formula built only from Schwartz-MacPherson classes of the strata, confirming the description anticipated by Ohmoto and Yokura.","In the same line-bundle setting, the Milnor class can be written through Aluffi's $\\mu$-classes of the singular loci of the components.","For $r=2$, the general formula specializes to a closed expression mixing $\\mathcal{M}(X_1)$, $\\mathcal{M}(X_2)$, $c^{SM}(X_1)$, and $c^{SM}(X_2)$ with one inverse Chern factor, making two-component intersections as tractable as hypersurfaces.","The Milnor class of $X$ can also be expressed via the global Lê classes of the hypersurfaces $X_i$, linking the complete-intersection invariant to existing local Lê cycle computations."],"supporting_citations":[{"why":"Supplies the refined Gysin homomorphism, the regular-embedding properties, and the projection formula used to prove the Verdier-Riemann-Roch identities for cSM and cFJ.","marker":"[11]"},{"why":"Provides the exterior product formula for Schwartz-MacPherson classes that turns the component classes into the class of X1×...×Xr.","marker":"[13]"},{"why":"Gives the product formula for Milnor classes used in Theorem 3.1 to expand the Milnor class of the product X1×...×Xr.","marker":"[17]"},{"why":"Gives the Parusiński-Pragacz description of Schwartz-MacPherson classes via projectivized conormal cycles and the hypersurface Milnor class formula used as base case.","marker":"[19]"},{"why":"Introduces the µ-classes and the formula for Milnor classes of singular hypersurfaces that Corollary 3.3 extends to complete intersections.","marker":"[1]"},{"why":"Provides the stratified Morse theory notions of complex link and normal Morse index used to compute normal data and compare slices under the diagonal.","marker":"[12]"}],"fun_headline_variants":["Milnor class of transverse complete intersection is a product","Chern-corrected product gives Milnor class of intersection","Transverse pieces yield Milnor class via Chern factor","Milnor class of complete intersections: a product formula","Milnor class from pieces: a simple formula"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the pieces $X_i$ meet in general position in the strong stratified sense: the product $X_1\\times\\cdots\\times X_r$ must admit a Whitney stratification to which the diagonal embedding of $M$ is transverse, and the paper's own example shows the formula is false when this fails.","fun_headline_variants_meta":{"raw":{"variants":["Milnor class of transverse complete intersection is a product","Chern-corrected product gives Milnor class of intersection","Transverse pieces yield Milnor class via Chern factor","Milnor class of complete intersections: a product formula","Milnor class from pieces: a simple formula"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000587,"raw_usage":{"total_tokens":2788,"prompt_tokens":1005,"completion_tokens":1783,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":621,"completion_tokens_details":{"reasoning_tokens":1706}},"tokens_in":621,"tokens_out":1783,"duration_ms":278236,"temperature":1.0,"reasoning_tokens":1706,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:26:34.052629+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the Milnor class of the intersection in Example 2.5, where $X_1=\\{x_0x_1=0\\}$ and $X_2=\\{x_3=0\\}$ in $\\mathbb{P}^4$, by a direct stratified-Morse or vanishing-cycle computation that does not use Theorem 1; the theorem predicts $-H^3$, so any other class in $H_*(X)$ would settle the claim negatively. The paper's Remark 2.6 is the negative control: a smooth quadric and a tangent plane in $\\mathbb{P}^3$ violate the transversality hypothesis, and there the component-wise formula gives zero while the true Milnor class is the class of a point.","supporting_citations":[{"cited_title":"Ergebnisse der Mathematik und ihrer Grenzgebiete, Spring er-Verlag, Berlin, (1984)","cited_arxiv_id":null,"evidence_quote":"Supplies the refined Gysin homomorphism, the regular-embedding properties, and the projection formula used to prove the Verdier-Riemann-Roch identities for cSM and cFJ."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the exterior product formula for Schwartz-MacPherson classes that turns the component classes into the class of X1×...×Xr."},{"cited_title":"Bulletin of the Polish Acad","cited_arxiv_id":null,"evidence_quote":"Gives the product formula for Milnor classes used in Theorem 3.1 to expand the Milnor class of the product X1×...×Xr."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the Parusiński-Pragacz description of Schwartz-MacPherson classes via projectivized conormal cycles and the hypersurface Milnor class formula used as base case."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the µ-classes and the formula for Milnor classes of singular hypersurfaces that Corollary 3.3 extends to complete intersections."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the stratified Morse theory notions of complex link and normal Morse index used to compute normal data and compare slices under the diagonal."}],"review_version":1}