Pith. sign in

REVIEW 3 major objections 4 minor 25 references

On the Chern classes of singular complete intersections

T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The total Milnor class of a complete intersection is determined by its components and the ambient tangent bundle.

desk verdict A useful but restricted product formula for Milnor classes of complete intersections, with an abstract that overstates novelty. read the letter →

arxiv 1909.01117 v1 pith:XZGYVYQJ submitted 2019-09-03 math.AG

Roberto Callejas-Bedregal , Michelle Morgado , Jose Seade This is my paper · ORCID
classification math.AG MSC 14C1755N4514M1014B0532S20
keywords completeintersectionsMilnorclassesSchwartz-MacPhersonFulton-JohnsonWhitneystratificationssingularvarietiesChernLê
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

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.

What carries the argument

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 $$\$\Delta$^!\big($c^{{FJ}}$(Z(s))\big)=c\big((TM|_{Z(\$\Delta$^*s)})^{\oplus r-1}\big)\cap $c^{{FJ}}$(Z(\$\Delta$^*s))$$ and 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.

What would settle it

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.

Watch

Extended reading notes

Core claim

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 $X=X_1\cap\cdots\cap X_r$, $$$c^{{SM}}$(X)=c((TM|_X)^{\oplus r-1})^{-1}\cap\big($c^{{SM}}$(X_1)\cdots $c^{{SM}}$(X_r)\big),$$ and the analogous identity holds for $c^{FJ}$. Consequently the total Milnor class satisfies $$\mathcal{M}(X)=(-1)^{\dim X}c((TM|_X)^{\oplus r-1})^{-1}\cap\big($c^{{FJ}}$(X_1)\cdots $c^{{FJ}}$(X_r)-$c^{{SM}}$(X_1)\cdots $c^{{SM}}$(X_r)\big).$$ The 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.

Load-bearing premise

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.

Editorial extensions

If this is right

  • 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.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • 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.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

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.

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 (3)
  1. [Abstract and §1 (Theorem 1.12)] 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.
  2. [§2, Theorem 1 assumptions and Remark 2.6] 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.
  3. [§2, Proposition 1.8] 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.
minor comments (4)
  1. [Example 2.5] 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.
  2. [Corollary 3.5 proof] 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.
  3. [Theorem 1 statement] 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.
  4. [Throughout] There are several typographical errors, including 'intesection' in Theorem 1 and 'stratification' in Proposition 1.8; a careful proofreading pass is recommended.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main formulas are derived from external Verdier-Riemann-Roch, product, and refined-Gysin theorems, not from assumptions equivalent to the conclusions.

full rationale

Theorem 1 is proved from Propositions 2.1 and 2.4, which in turn use Theorem 1.12 and Proposition 1.5. Proposition 1.5 is a direct computation using Fulton's refined Gysin formalism ([11, Prop. 14.1 and Prop. 6.3]) applied to the diagonal embedding. Theorem 1.12 is proved from Parusiński-Pragacz's formula for Schwartz-MacPherson classes (Remark 1.11(2)), the refined Gysin calculus, and Corollary 1.10. Corollary 1.10 rests on Proposition 1.8 and Lemma 1.9, which are proven internally from MacPherson's characteristic-cycle isomorphism and the vanishing-cycle computation of normal Morse indices; the only geometric input is the Whitney-transversality hypothesis, which is explicit and later shown to be necessary in Remark 2.6. Proposition 2.1 invokes Kwieciński's external product formula for Chern-Schwartz-MacPherson classes, and Proposition 2.4 uses the product formula for Fulton-Johnson classes proved in Lemma 2.3 directly from the definitions and refined intersection products. Theorem 3.1 uses the Ohmoto-Yokura product formula for Milnor classes ([17, Cor. 3.1]), and Corollary 3.5 uses the Parusiński-Pragacz hypersurface formula from [19] and induction, together with Proposition 2.1 and Lemma 3.4. No fitted parameter is introduced, no quantity is defined in terms of the target formula, and no theorem needed for the main result is justified by a self-citation: the unpublished self-citations [8] and [9] are background refinements, and [7] appears only in the final remark about global Lê classes. The abstract's 'first Verdier-Riemann-Roch type formulae' claim may collide with Schürmann's earlier work [20], but that is a novelty/correctness concern, not circularity. The transversality restriction, including its failure without the hypothesis (Remark 2.6), is openly stated and does not smuggle in the conclusion.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The paper introduces no new mathematical objects. It relies instead on standard constructible sheaf theory, stratified Morse theory, refined Gysin maps, and two external product formulas. The only domain-specific input is the transversality hypothesis, which is explicit and shown to be necessary.

assumptions (6)
  • standard math Standard derived category and constructible sheaf formalism, including the Euler characteristic formula χ(X, F) = sum of χ(F_S)χ(S) over strata.
    Invoked throughout Section 1.1 and Proposition 1.8, citing Dimca [10, Theorem 4.1.22].
  • standard math Stratified Morse theory: normal Morse data, complex links, normal Morse indices, and their independence of choices.
    Used in Section 1.1 to define η(Sα, F) and to compute it via vanishing cycles, citing Goresky-MacPherson [12].
  • standard math Fulton's refined Gysin formalism, including base change, projection formula, and compatibility with Chern classes.
    Used in Propositions 1.5, 1.8, 2.1, 2.4 and Theorem 1.12, citing Fulton [11, Chapter 6].
  • standard math Kwiecinski's exterior product formula for Schwartz-MacPherson classes: cSM(X1 × ... × Xr) = cSM(X1) × ... × cSM(Xr).
    Used in the proof of Proposition 2.1, citing [13].
  • standard math Ohmoto-Yokura product formula for Milnor classes of products of singular varieties.
    Used in the proof of Theorem 3.1 to expand M(X1 × ... × Xr), citing [17, Corollary 3.1].
  • domain assumption Existence of Whitney stratifications of the Xi such that all intersections between strata are transversal, equivalently the diagonal embedding is transversal to the product stratification.
    This is the main hypothesis of Theorem 1 and Section 2; it is not proven to hold for arbitrary complete intersections and Remark 2.6 shows it cannot simply be omitted.

how reviews work

0 comments
Cite this review

Pith. "Pith review of On the Chern classes of singular complete intersections." pith.science (2026). https://pith.science/paper/XZGYVYQJ

@misc{pith2026190901117,
  author       = {Pith},
  title        = {Pith review of: On the Chern classes of singular complete intersections},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XZGYVYQJ}},
  note         = {Machine review of arXiv:1909.01117}
}
abstract

We consider two classical extensions for singular varieties of the usual Chern classes of complex manifolds, namely the total Schwartz-MacPherson and Fulton-Johnson classes, $c^{SM}(X)$ and $c^{FJ}(X)$. Their difference (up to sign) is the total Milnor class ${\mathcal M}(X)$, a generalization of the Milnor number for varieties with arbitrary singular set. We get first Verdier-Riemann-Roch type formulae for the total classes $c^{SM}(X)$ and $c^{FJ}(X)$, and use these to prove a surprisingly simple formula for the total Milnor class when $X$ is defined by a finite number of local complete intersection $X_1,\cdot \ldots \cdot,X_r$ in a complex manifold, satisfying certain transversality conditions. As applications we obtain a Parusi\'{n}ski-Pragacz type formula and an Aluffi type formula for the Milnor class, and a description of the Milnor classes of $X$ in terms of the global L\^e classes of the $X_i$.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 23 canonical work pages

  1. [20]

    Sch¨ urmann, J., A generalized Verdier-type Riemann-Roch theorem for Chern -Schwartz-MacPherson classes , arXiv:math/0202175

    arXiv

  2. [1]

    Aluffi, P., Chern classes for singular hypersurfaces. Trans. Amer. Math. Soc. (351) (1999), no. 10, 3989-4026

    work page 1999

  3. [2]

    Aluffi, P., Inclusion-exclusion and Segre classes. Comm. Algebra 31 (2003), no. 8, 3619–3630

    work page 2003

  4. [3]

    [J] Commun

    Aluffi, P., Marcolli, M., Feynman motives of banana graphs . [J] Commun. Number Theory Phys. 3, No. 1, 1-57 (2009)

    work page 2009

  5. [4]

    Ast´ erisque 82-83 (1981), 93–147

    Brasselet, J.-P., Schwartz, M.-H., Sur les classes de Chern d’un ensemble analytique complexe . Ast´ erisque 82-83 (1981), 93–147

    work page 1981

  6. [5]

    Brasselet, J.-P., Lehmann, D., Seade, J., Suwa, T., Milnor classes of local complete intersections . Trans. Amer. Math. Soc. 354 (2001), 1351-1371

    work page 2001

  7. [6]

    Springer Verlag L

    Brasselet, J.-P., Seade, J., Suwa, T., Vector fields on singular varieties . Springer Verlag L. N. M. 1987 (2009)

    work page 2009

  8. [7]

    Callejas-Bedregal, R., Morgado, M. F. Z., Seade, J., Lˆ e cycles and Milnor classes. Inventiones Mathematicae: Volume 197, Issue 2 (2014), 453–482 and 483–489

    work page 2014

Show all 25 references
  1. [8]

    Callejas-Bedregal, R., Morgado, M. F. Z., Seade, J., On the Milnor classes of local complete intersections . Preprint 2012, arXiv:1208.5084

    2012 arXiv

  2. [9]

    Callejas-Bedregal, R., Morgado, M. F. Z., Seade, J., On the total Milnor class of complete intersection . Preprint 2018

    2018

  3. [10]

    Universitext

    Dimca, A., Sheaves in topology . Universitext. Springer-Verlag, Berlin, (2004)

    2004

  4. [11]

    Ergebnisse der Mathematik und ihrer Grenzgebiete, Spring er-Verlag, Berlin, (1984)

    Fulton, W., Intersection theory. Ergebnisse der Mathematik und ihrer Grenzgebiete, Spring er-Verlag, Berlin, (1984)

    1984

  5. [12]

    Goresky, M., MacPherson, R., Stratified Morse theory . Ergeb. Math. Grenzgeb. (3) 14, Berlin, Springer-Verlag, ( 1988)

    1988

  6. [13]

    Kwieci´ nski, M., Formule du produit pour les classes caract´ eristiques de Ch ern- Schwartz-MacPherson et homologie din- tersection. C. R. Acad. Sci. Paris S´ er. I Math. 314 (1992), no.8, 625–62 8

    1992

  7. [14]

    B., Lˆ e Cycles and Hypersurface Singularities

    Massey, D. B., Lˆ e Cycles and Hypersurface Singularities. Lecture Notes in Mathematics 1615 (1995), Springer-Verla g

    1995

  8. [15]

    Annals of Math

    MacPherson, R., Chern classes for singular algebraic varieties . Annals of Math. 100 (1974), n. 2, 423–432

    1974

  9. [16]

    Hirzebruch-Milnor classes of complete intersections

    Maxim, L., Saito, M., Sch¨ urmann, J. Hirzebruch-Milnor classes of complete intersections . Adv. Math. 241 (2013), 220–245

    2013

  10. [17]

    Bulletin of the Polish Acad

    Ohmoto, T.,Yokura, S., Product Formulas for the Milnor Class . Bulletin of the Polish Acad. of Sciences Math. Vol. 48 (2000), no. 4, 387-401. 12 ROBERTO CALLEJAS-BEDREGAL, MICHELLE F. Z. MORGADO, AND J OS ´E SEADE

    2000

  11. [18]

    Journal of the American Math

    Parusi´ nski, A., Pragacz, P., Chern-Schwartz-MacPherson Classes and the Euler Characte ristic of Degeneracy Loci and Special Divisors. Journal of the American Math. Society, vol. 8, no. 4 (1995), 793–817

    1995

  12. [19]

    Parusi´ nski, A., Pragacz, P., Characteristic classes of hypersurfaces and characterist ic cycles . J. Alg. Geo. 10 (2001), 63-79

    2001

  13. [21]

    Sch¨ urmann, J., Topology of singular spaces and constructible sheaves , Monografie Matematyczne 63 (New Series), Birkhauser, Basel, 2003

    2003

  14. [22]

    arXiv: 1510.01986v3 [math

    Sch¨ urmann, J., Chern classes and transversality for singular spaces . arXiv: 1510.01986v3 [math. AG] 6 Jan 2016

    2016 arXiv

  15. [23]

    H., Classes caract´ eristiques d´ efinis par une stratification d’une vari´ et´ e analytique complexe

    Schwartz, M. H., Classes caract´ eristiques d´ efinis par une stratification d’une vari´ et´ e analytique complexe. C.R. Acad. Sci. Paris 260 (1965), 3262–3264 and 3535–3537

    1965

  16. [24]

    Internat

    Seade, J., Suwa, T., An adjunction formula for local complete intersections . Internat. J. Math. 9 (1998), 759–768

    1998

  17. [25]

    Suwa, T., Classes de Chern des intersections compl` etes locales . C. R. Acad. Sci. Paris I Math., 324 (1996), 67–70. Centro de Ci ˆencias Exatas e da Natureza, Universidade Federal da Para ´ıba-UFPb, Jo˜ao Pessoa, PB - Brasil. E-mail address : roberto@mat.ufpb.br Instituto de...

    1996

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.