Pith. sign in

REVIEW 2 major objections 3 minor 30 references

Combinatorially Determined Zeroes of Bernstein--Sato Ideals for Tame and Free Arrangements

T0 review · 2 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read Central free hyperplane arrangements have Bernstein–Sato zero loci that are cut out by a fully combinatorial product of linear factors indexed by the indecomposable edges of the intersection lattice; for tame arrangements the roots in…

desk verdict A substantial extension of Maisonobe's program to non-reduced and tame/free arrangements; the main formulas are credible, but the referee should verify the imported annihilator-generation theorem from the author's preprint [3]. read the letter →

arxiv 1909.00547 v3 pith:JD3ZDQLJ submitted 2019-09-02 math.AG math.ACmath.CV

classification math.AGmath.ACmath.CV MSC 14F1032S4032S0532S2232C38
keywords Bernstein–SatoidealpolynomialhyperplanearrangementfreedivisortameD-moduleintersectionlatticelogarithmicderivations
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

Bernstein–Sato polynomials and ideals package the vanishing of powers of a function into algebraic data; for a general hypersurface they are hard to compute and are not determined by combinatorial data. This paper shows that central hyperplane arrangements behave much better: when the arrangement is free (its logarithmic derivations form a locally free module), the zero locus of the multivariate Bernstein–Sato ideal for any factorization is cut out by an explicit product of linear polynomials indexed by the indecomposable edges of the intersection lattice, and the same product gives the roots of the Bernstein–Sato polynomial of every power of a reduced free arrangement. For merely tame arrangements, the paper proves that all roots in $[-1,0)$ are combinatorially determined. It also uses generalized ideals $B^g_{f'F}$ to show that certain small roots force a lower bound on the number of hyperplanes one must add to a tame arrangement to make it free. These are among the few cases in which D-module invariants reduce to counting degrees of subarrangements.

What carries the argument

The load-bearing object is the extended Spencer co-complex $\mathrm{Sp}^g_{f'F,x}$, built from the logarithmic derivations $\mathrm{Der}_{X,x}(-\log f)$ together with the principal ideal $(g)$. Freeness makes this a finite free complex, and its $D_{X,x}[S]$-dual is computed in Theorem 3.9: the dual of $D_{X,x}[S]f'F^S / D_{X,x}[S]gf'F^S$ is a shift of the analogous module with $F^S$ replaced by $(gf'f_{\mathrm{red}})^{-1}F^{-S}$. From this duality one obtains the symmetry involution $\phi(s_k)=-s_k-\frac{1}{m_k}-\frac{2d'_k}{d_k}-\frac{d''_k}{d_k}$, which mirrors the Bernstein–Sato variety. On the estimate side, the theorem that the annihilator of $f'F^S$ is generated by derivations under tameness, Euler-homogeneity, and local finiteness of the logarithmic stratification lets a right-constant-term argument produce containment in a product of linear factors; the symmetry closes the gap into the exact formula (1.2). The linear polynomials $P^g_{f'F,X}$ encode the degrees of the induced factorization on each indecomposable edge, so all data in the final formula are read off from the intersection lattice.

What would settle it

Compute the Bernstein–Sato polynomial of $f^2$ for a concrete central reduced free arrangement such as $f=xyz(x+y)(x+z)(y+z)$ with a computer algebra system, and compare each root with the finite set predicted by the product formula (4.25); a single root outside that set, or a missing predicted root, would refute the paper's central formula.

Watch

Extended reading notes

Core claim

The central discovery is Theorem 1.4. Let $f$ be a central, possibly non-reduced, free hyperplane arrangement with factorization $f=f_1\cdots f_r$, let $f'$ divide $f$, and write $g=f/f'$. If $(f',F)$ is an unmixed pair up to units and $\deg f'\leq 4$, then the reduced zero locus of the Bernstein–Sato ideal $B^g_{f'F}$ is exactly the variety of $$\prod_{X\text{ indecomposable}}\prod_{j=0}^{d_{X,\mathrm{red}}+d_X-2r(X)-d'_X}\left(P^g_{f'F,X}+j\right),$$ where $X$ runs over indecomposable edges of the intersection lattice, $r(X)$ is the rank of the edge, and $P^g_{f'F,X}=\sum_k d_{X,k}s_k+r(X)+d'_X$ is an explicit linear form. Setting $f'=1$ and $f$ reduced removes the degree bound and gives $V(B_F)$ for every factorization; in particular the roots of the Bernstein–Sato polynomial of any power of a central reduced free arrangement are given by the same combinatorial product. For tame arrangements the paper establishes that the roots in $[-1,0)$ are combinatorial, complementing known examples where roots outside that interval are not. The formulas rest on a duality theorem for the D-module $D_{X,x}[S]f'F^S$ that generalizes the free-divisor symmetry previously proved in the univariate case.

Load-bearing premise

The load-bearing premise is that the arrangement is free—the module of logarithmic derivations is locally free—because freeness makes the extended Spencer co-complex a finite free resolution and yields the duality that powers the symmetry; without it only the tame result for roots in $[-1,0)$ is obtained.

Editorial extensions

If this is right

  • For a central reduced free arrangement and any factorization, the zero locus of the multivariate Bernstein–Sato ideal $V(B_F)$ is a hypersurface cut out by formula (1.3), so the entire zero locus is read off from the intersection lattice.
  • For every power $f^k$ of a central reduced free arrangement, the roots of the Bernstein–Sato polynomial are combinatorial, lie in $(-1-1/k,0)$, and accumulate at $-1$ as $k\to\infty$.
  • For a tame arrangement that is not free, all roots of $B^g_{f'f}$ in $[-1,0)$ are exactly the finite union of rational numbers $-j_X/d_X$ dictated by indecomposable edges; any non-combinatorial root must lie outside that interval.
  • If a tame reduced arrangement of rank $n$ and degree $d$ has a root $-(2d+v)/d$ with $1<v\le n-1$ and $\gcd(d,v)=1$, then every central arrangement that frees it has degree at least $n-v$.
  • Appendix B verifies the conjectured link between exponentiating $V(B_F)$ and the cohomology support locus of the complement for central reduced free arrangements.

Reading between the lines

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

  • Remark 4.28 suggests that the restrictions "unmixed pair up to units" and $\deg f'\le4$ are proof artifacts: if the announced vanishing criterion is applied, formula (1.2) should hold for all free arrangements and all compatible $f'$.
  • Because freeness is the operative hypothesis rather than hyperplane-specific geometry, the same Spencer duality may compute Bernstein–Sato zero loci for other free divisors, though the paper's Remark 3.10 warns that the relevant Koszul complex can fail to be a resolution when the auxiliary ideal is non-principal.
  • The freeing bound of Theorem 5.4 could be tested sharply: take tame arrangements with the required small roots, construct a freeing arrangement from the intersection lattice alone, and compare its degree with $n-v$; such experiments would show whether the bound is tight.
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

2 major / 3 minor

Summary. The paper studies generalized Bernstein–Sato ideals B^g_{f'F} attached to a central hyperplane arrangement f, a factorization f = f_1 ... f_r, and a divisor f' of f. The main results are: for tame arrangements, a combinatorial description of the roots lying in [-1,0) of the univariate Bernstein–Sato polynomial; for free arrangements paired with an unmixed factorization and deg(f') ≤ 4, a closed formula for the zero locus V(B^g_{f'F}) as a product of explicit linear factors indexed by indecomposable edges of the intersection lattice; and for free reduced arrangements an independent proof of Budur's conjecture. The paper also uses the generalized ideals to give lower bounds on the degree of an arrangement needed to free a tame arrangement. The methods extend Maisonobe's Spencer-complex and symmetry arguments, with a duality theorem for D_{X,x}[S]f'F^S proved via the extended Spencer co-complex, a trace-of-adjoints calculation deferred to Appendix A, and Budur's conjecture treated in Appendix B.

Significance. If the main results are correct, they represent a substantial advance: they give combinatorial formulas for Bernstein–Sato zero loci in situations where previously only the linear-factorization case was understood, and they prove combinatorial determinacy of the small roots for all tame arrangements. The manuscript is unusually explicit about its hypotheses and limitations, and it supplies detailed arguments for the Spencer duality and the Budur-conjecture case, including a self-contained proof of the trace formula in Appendix A. The main open risk is not an internal contradiction but the reliance of the central chain on results from the author's unpublished preprint [3] and on the sketchy proof of Proposition 2.26; these points need to be addressed before the paper can be considered fully self-contained.

major comments (2)
  1. [§2.2, Theorem 2.21 and Theorem 2.20] The generated-by-derivations theorem is a load-bearing premise: it is used in Theorem 3.9 to identify the Spencer co-complex as a resolution, and Theorem 3.9 is in turn used for the symmetry Theorem 3.16 and for the zero-locus computations in Corollary 4.27/Theorem 1.4. The proof of Theorem 2.20 relies on Theorem 2.23 and Corollary 2.28 of the author's unpublished preprint [3] for the primality, dimension, and annihilator equality of the generalized Liouville ideal. This is a dependency rather than a circularity, but the present text does not state or prove those inputs. Since a failure of any of these inputs would propagate to the main theorem, the manuscript should either include the relevant statements and proofs or clearly delimit which parts of the argument are contingent on [3].
  2. [§2.3, Proposition 2.26] Proposition 2.26 is what converts the element of B^g_{f'L} obtained for the linear factorization L into an element of B^g_{f'F} for an arbitrary factorization F; it is used in Corollary 4.19 and again in the proof of Theorem 4.26, and through Theorem 4.26 it supports the 'any factorization' claim in Theorem 1.4. The proof is only sketched as 'essentially the same as the proof of Proposition 5.3.' Since a gap here would invalidate all results for non-linear factorizations, the full argument should be supplied or the proposition should be stated as a theorem with a complete proof.
minor comments (3)
  1. [Example 2.27] The displayed formula for B_F has an unmatched parenthesis; the expression should be checked for typographical correctness.
  2. [Theorem 4.18] The theorem statement does not assume f is essential, but the proof explicitly reduces to the essential case after the inductive step. The statement should either include essentiality or explain how the non-essential case is deduced, for example by adding dummy variables or by treating the induced arrangement in the span of its hyperplanes.
  3. [Definition 4.10 and Theorem 1.4] The formulas involve products indexed by j_X = 0,..., d_{X,red}+d_X-2r(X)-d'_X; it would help the reader if the text stated explicitly whether an empty product is interpreted as 1 in the (apparently impossible) case where the upper limit is negative, and why the hypotheses on f' and X guarantee nonnegative upper limits.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular reduction; zero-locus formula is obtained by independent upper/lower inclusions, with external dependencies on the author's preprint [3] as the main risk.

full rationale

The claimed zero-locus computation is not circular. Theorem 1.4 is proved by sandwiching: Theorem 4.26 gives lower inclusion (4.19) and upper inclusion (4.20); when (f',F) is unmixed up to units and deg(f') ≤ 4 the two products have the same variety, so Corollary 4.27 concludes equality. The lower inclusion comes from an explicit element produced in Theorem 4.18/Corollary 4.19, and the upper inclusion from Theorem 4.11 together with the symmetry of Theorem 3.16. Neither step assumes the target variety; the combinatorial product is formed from intersection-lattice data and factorization degrees, not from the zero locus being computed. The symmetry itself is a consequence of the duality Theorem 3.9, which uses the Spencer co-complex and the annihilator-generation theorem. The least secured part is Theorem 2.21: its proof depends on lemmas from the author's earlier preprint [3] ('By Theorem 2.23 of loc. cit., ~L_{F,x} is a prime ideal of dimension n+r'), and Proposition 2.26's proof is only sketched ('The argument is essentially the same as the proof of Proposition 5.3 of this paper'). These are external or deferred dependencies, not circular reductions: [3] is a parameter-free statement whose assumptions do not include the zero-locus theorem, and no equation in this paper is redefined or fitted to produce the answer. The paper also flags its own limitations (Remark 3.10 on non-principal I; Remark 4.28(c) on possibly unnecessary unmixed-pair hypothesis), which further confirms the derivation is not manufactured. Overall: no significant circularity, score 1.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The central formulas rest on standard D-module theory plus three domain facts specific to arrangements (Saito-holonomicity, Euler-homogeneity, and freeness among them). No free parameters or invented entities are introduced.

assumptions (5)
  • domain assumption Hyperplane arrangements are Saito-holonomic and strongly Euler-homogeneous (Examples 2.7 and 2.10).
    This places arrangements into the class where the annihilator-generation theorem (Theorem 2.20) and the duality theorem (Theorem 3.9) apply; it is a structural fact about arrangements, not derived in this paper.
  • standard math Theorem 2.23 of [3]: for tame, strongly Euler-homogeneous, Saito-holonomic f, the generalized Liouville ideal is prime of dimension n+r, giving annihilator generation.
    This is the engine behind Theorem 2.21 and hence all later inclusions. It is cited from the author's earlier preprint [3] rather than reproved here.
  • standard math Saito's freeness criterion for rank at most 2 arrangements and the equivalence indecomposable iff mdr(f) >= 2 (Remark 4.7).
    Used in Theorem 4.25 to reduce the free rank-2 case and in Theorem 4.8 to choose the polynomial v; classical results from Saito's theory of logarithmic forms.
  • standard math Standard D-module facts: side-changing functor between left and right D-modules, grade and purity of holonomic modules, Bjork's characterization of characteristic ideals.
    Used in Section 3.2 to prove principality of the radical of B^g_{f'F,x} and in Theorem 3.16 for the symmetry involution.
  • standard math Castro-Jimenez-Ucha trace of adjoints formula (Theorem 4.1.4 of [9]); the paper gives an alternative proof in Appendix A.
    Used in Proposition 3.6 to identify the image of the dual Spencer differential; the formula itself is a known background result.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Combinatorially Determined Zeroes of Bernstein--Sato Ideals for Tame and Free Arrangements." pith.science (2026). https://pith.science/paper/JD3ZDQLJ

@misc{pith2026190900547,
  author       = {Pith},
  title        = {Pith review of: Combinatorially Determined Zeroes of Bernstein--Sato Ideals for Tame and Free Arrangements},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JD3ZDQLJ}},
  note         = {Machine review of arXiv:1909.00547}
}
abstract

For a central, not necessarily reduced, hyperplane arrangement $f$ equipped with any factorization $f = f_{1} \cdots f_{r}$ and for $f^{\prime}$ dividing $f$, we consider a more general type of Bernstein--Sato ideal consisting of the polynomials $B(S) \in \mathbb{C}[s_{1}, \dots, s_{r}]$ satisfying the functional equation $B(S) f^{\prime} f_{1}^{s_{1}} \cdots f_{r}^{s_{r}} \in \text{A}_{n}(\mathbb{C})[s_{1}, \dots, s_{r}] f_{1}^{s_{1} + 1} \cdots f_{r}^{s_{r} + 1}.$ Generalizing techniques due to Maisonobe, we compute the zero locus of the standard Bernstein--Sato ideal in the sense of Budur (i.e. $f^{\prime} = 1)$ for any factorization of a free and reduced $f$ and for certain factorizations of a non-reduced $f$. We also compute the roots of the Bernstein--Sato polynomial for any power of a free and reduced arrangement. If $f$ is tame, we give a combinatorial formula for the roots lying in $[-1,0).$ For $f^{\prime} \neq 1$ and any factorization of a line arrangement, we compute the zero locus of this ideal. For free and reduced arrangements of larger rank, we compute the zero locus provided $\text{deg}(f^{\prime}) \leq 4$ and give good estimates otherwise. Along the way we generalize a duality formula for $\mathscr{D}_{X,\mathfrak{x}}[S]f^{\prime}f_{1}^{s_{1}} \cdots f_{r}^{s_{r}}$ that was first proved by Narv\'aez-Macarro for $f$ reduced, $f^{\prime} = 1$, and $r = 1.$ As an application, we investigate the minimum number of hyperplanes one must add to a tame $f$ so that the resulting arrangement is free. This notion of freeing a divisor has been explicitly studied by Mond and Schulze, albeit not for hyperplane arrangements. We show that small roots of the Bernstein--Sato polynomial of $f$ can force lower bounds for this number.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 29 canonical work pages

  1. [3]

    Bernstein-Sato Varieties and Annihilation of Powers

    Daniel Bath. Bernstein–Sato Varieties and Annihilatio n of Powers. arXiv e-prints , page arXiv:1907.05301, Jul 2019

  2. [1]

    Plus-one generated and next to free arrangements of hyperplanes

    Takuro Abe. Plus-one generated and next to free arrangem ents of hyperplanes. arXiv e-prints, page arXiv:1808.04697, Aug 2018

  3. [2]

    Local Bernstein-Sa to ideals: algorithm and exam- ples

    Rouchdi Bahloul and Toshinori Oaku. Local Bernstein-Sa to ideals: algorithm and exam- ples. J. Symbolic Comput. , 45(1):46–59, 2010

  4. [4]

    Analytic D-modules and applications , volume 247 of Mathematics and its Applications

    Jan-Erik Bjork. Analytic D-modules and applications , volume 247 of Mathematics and its Applications . Kluwer Academic Publishers Group, Dordrecht, 1993

  5. [5]

    Cohen-Macaulay rings , volume 39 of Cambridge Studies in Advanced Mathematics

    Winfried Bruns and J¨ urgen Herzog. Cohen-Macaulay rings , volume 39 of Cambridge Studies in Advanced Mathematics . Cambridge University Press, Cambridge, 1993

  6. [6]

    New free divisors from old , 5(1):17–47, 2014

    Buchweitz, Ragnar-Olaf and Conca, Aldo. New free divisors from old , 5(1):17–47, 2014

  7. [7]

    Bernstein-Sato ideals and local systems

    Nero Budur. Bernstein-Sato ideals and local systems. Ann. Inst. Fourier (Grenoble) , 65(2):549–603, 2015

  8. [8]

    Zero loci of Bernstein-Sato ideals

    Nero Budur, Robin van der Veer, Lei W u, and Peng Zhou. Zero loci of Bernstein–Sato ideals. arXiv e-prints , page arXiv:1907.04010, Jul 2019

Show all 30 references
  1. [9]

    F. J. Castro-Jim´ enez and J. M. Ucha. Free divisors and du ality for D-modules. Tr. Mat. Inst. Steklova , 238(Monodromiya v Zadachakh Algebr. Geom. i Differ. Uravn. ):97–105, 2002

  2. [10]

    Solvable groups, free divis ors, and nonisolated matrix singularities I: Towers of free divisors

    James Damon and Brian Pike. Solvable groups, free divis ors, and nonisolated matrix singularities I: Towers of free divisors. Ann. Inst. Fourier (Grenoble) , 65(3):1251–1300, 2015

  3. [11]

    Computing Milno r fiber monodromy for some projective hypersurfaces

    Alexandru Dimca and Gabriel Sticlaru. Computing Milno r fiber monodromy for some projective hypersurfaces. arXiv e-prints , page arXiv:1703.07146, Mar 2017

  4. [12]

    Linear free di- visors and the global logarithmic comparison theorem

    Michel Granger, David Mond, Alicia Nieto-Reyes, and Ma thias Schulze. Linear free di- visors and the global logarithmic comparison theorem. Ann. Inst. Fourier (Grenoble) , 59(2):811–850, 2009

  5. [13]

    On the formal struc ture of logarithmic vector fields

    Michel Granger and Mathias Schulze. On the formal struc ture of logarithmic vector fields. Compos. Math. , 142(3):765–778, 2006

  6. [14]

    Cohomologie ´ e vanescente p-adique: calculs lo- caux

    Michel Gros and Luis Narv´ aez-Macarro. Cohomologie ´ e vanescente p-adique: calculs lo- caux. Rend. Sem. Mat. Univ. Padova , 104:71–90, 2000

  7. [15]

    Kashiwara

    M. Kashiwara. Vanishing cycle sheaves and holonomic sy stems of differential equations. In Algebraic geometry (Tokyo/Kyoto, 1982) , volume 1016 of Lecture Notes in Math. , pages 134-142. Springer, Berlin, 1983

  8. [16]

    Filtration Relative, l’Id´ eal de Bernstein et ses pentes

    Philippe Maisonobe. Filtration Relative, l’Id´ eal de Bernstein et ses pentes. arXiv e-prints , page arXiv:1610.03354, Oct 2016

  9. [17]

    Id´ eal de Bernstein d’un arrangement central g´ en´ erique d’hyperplans

    Philippe Maisonobe. Id´ eal de Bernstein d’un arrangement central g´ en´ erique d’hyperplans. arXiv e-eprints , page arXiv:1610.03357, Oct 2016

  10. [18]

    L’id´ eal de Bernstein d’un arrang ement libre d’hyperplans lin´ eaires

    Philippe Maisonobe. L’id´ eal de Bernstein d’un arrang ement libre d’hyperplans lin´ eaires. arXiv e-prints , page arXiv:1610.03356, Oct 2016. 42 DANIEL BATH

  11. [19]

    Malgrange

    B. Malgrange. Le polynˆ ome de Bernstein d’une singular it´ e isol´ ee. InFourier integral operators and partial differential equations (Colloq. Inte rnat., Univ, Nice, Nice, 1974) , pages 98-119. Lecture Notes in Math., Vol. 459. Springer, Be rlin, 1975

  12. [20]

    Adjoint divisors and fr ee divisors

    David Mond and Mathias Schulze. Adjoint divisors and fr ee divisors. J. Singul., 7:253–274, 2013

  13. [21]

    A duality approach to the symmetry of Bernstein-Sato polynomials of free divisors

    Luis Narv´ aez Macarro. A duality approach to the symmetry of Bernstein-Sato polynomials of free divisors. Adv. Math. , 281:1242–1273, 2015

  14. [22]

    Theory of logarithmic differential forms a nd logarithmic vector fields

    Kyoji Saito. Theory of logarithmic differential forms a nd logarithmic vector fields. J. Fac. Sci. Univ. Tokyo Sect. IA Math. , 27(2):265–291, 1980

  15. [23]

    Bernstein-Sato polynomials of hyperp lane arrangements

    Morihiko Saito. Bernstein-Sato polynomials of hyperp lane arrangements. Selecta Math. (N.S.), 22(4):2017–2057, 2016

  16. [24]

    Roots of Bernstein–Sato polynomials o f certain homogeneous polynomials with two-dimensional singular loci

    Morihiko Saito. Roots of Bernstein–Sato polynomials o f certain homogeneous polynomials with two-dimensional singular loci. arXiv e-prints , page arXiv:1703.05741, Mar 2017

  17. [25]

    Tohˇ aneanu Homology of homogeneous divisors

    Aron Simis and S ¸tefan O. Tohˇ aneanu Homology of homogeneous divisors. Israel J. Math , 200(1):449–487, 2014

  18. [26]

    Computing homomorphism s between holonomic D- modules

    Harrison Tsai and Uli W alther. Computing homomorphism s between holonomic D- modules. J. Symbolic Comput. 32(6):597-617, 2001. Effective methods in rings of dif- ferential operators

  19. [27]

    Derivation degree sequences of non-free arrangements

    Max W akefield. Derivation degree sequences of non-free arrangements. Comm. Algebra , 47(7):2654–2666, 2019

  20. [28]

    Bernstein-Sato polynomial versus cohomo logy of the Milnor fiber for generic hyperplane arrangements

    Uli W alther. Bernstein-Sato polynomial versus cohomo logy of the Milnor fiber for generic hyperplane arrangements. Compos. Math. , 141(1):121-145, 2005

  21. [29]

    The Jacobian module, the Milnor fiber, and t he D-module generated by f s

    Uli W alther. The Jacobian module, the Milnor fiber, and t he D-module generated by f s. Invent. Math. , 207(3):1239–1287, 2017

  22. [30]

    personal communication

    Masahiko Yoshinaga. personal communication. Department of Mathematics, Purdue University, West Lafayet te, IN, USA. E-mail address : dbath@purdue.edu

Pith tools

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