Defect of projective hypersurfaces with isolated singularities

Morihiko Saito; Seung-Jo Jung

REVIEW 4 major objections 4 minor 1 cited by

Defect of projective hypersurfaces with isolated singularities

T0 review · 4 major / 4 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read For a projective hypersurface with isolated singularities, the defect—the gap between two middle Betti numbers—is shown to be the same object as the cokernel of a map into intersection cohomology, the rank of a vanishing-cycle map, and the

desk verdict The main theorem is a clean Hodge-module proof of the four-way equality for the defect; the numerical outputs are genuinely new but conditional on E2-degeneration cited to an unpublished preprint and on an unstated 'Theorem 2', so the concrete defect values should be treated as provisional. read the letter →

arxiv 2512.23522 v4 pith:3234I7ZY submitted 2025-12-29 math.AG

Seung-Jo Jung , Morihiko Saito This is my paper · ORCID

classification math.AG MSC 14J7014B0532S3532S40

keywords defectprojectivehypersurfaceisolatedsingularityintersectioncohomologyMilnorfiberpoleorderspectralsequencevanishingcyclesunipotentmonodromy

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

The paper establishes that a single number associated to a projective hypersurface with isolated singularities—the defect def(X)=h^{n+1}(X)-h^{n-1}(X)—is in fact four different invariants at once: the dimension of the cokernel of the natural map from ordinary to intersection cohomology, the dimension of the cokernel of a map in the smoothing exact sequence, the rank of the vanishing-cycle map, and the dimension of the unipotent monodromy part of the Milnor fiber cohomology of the defining polynomial. A sympathetic reader would care because the defect controls the failure of self-duality of the cohomology and, for threefolds, is tied to Q-factoriality. The paper also shows that when the singularities are weighted homogeneous, this defect equals the E2-term of the pole order spectral sequence, which can be computed explicitly with a computer algebra system. It gives explicit defect values for several known nodal quintic threefolds and for a hypersurface with a single singular point.

What carries the argument

The central object is the defect and the exact sequences of mixed Hodge modules that link it to vanishing cycles and intersection cohomology. The load-bearing identity is the short exact sequence 0 → ⊕_x (i_x)_* N V_{x,1}(1) → Q_{h,X}[n] → IC_X → 0, where IC_X is the intersection complex, whose cohomology yields the equality between the cokernel of ι^{n−1} and the rank of the vanishing-cycle map. For computations, the pole order spectral sequence—the spectral sequence of the double complex (Ω^•, df∧, d)—degenerates at E2 for weighted homogeneous isolated singularities, so the dimension of the relevant E2-term equals the defect.

What would settle it

Compute the E2 and E∞ terms of the pole order spectral sequence for a specific weighted homogeneous isolated singularity; if they differ, the E2-degeneration fails and the identification with the defect collapses. Alternatively, for one of the reported examples—say a nodal quintic with 118 nodes and defect 19—compute the defect independently by smoothing and direct cohomology; a mismatch would disprove the equality.

Watch

Extended reading notes

Core claim

The core discovery is the equality of four invariants: def(X) = dim Coker ι^{n−1} = dim Coker ρ = dim Im σ = dim H^n(F_f)_1. Here ι^{n−1} is the inclusion of ordinary cohomology into intersection cohomology, ρ and σ are maps in the exact sequence attached to a one-parameter smoothing, and H^n(F_f)_1 is the unipotent monodromy part of the Milnor fiber cohomology of the defining polynomial. The proof uses a short exact sequence of mixed Hodge modules relating the constant Hodge module on X to the intersection complex, whose kernel is the unipotent part of the vanishing cycles. For weighted homogeneous isolated singularities, the paper asserts that the pole order spectral sequence—the spectral

Load-bearing premise

The numerical claims rest on the assumption that the pole order spectral sequence degenerates at E2 for isolated weighted homogeneous singularities (a theorem cited from an earlier preprint, not proved here) and, in at least one example, on the code's assumption that the computed dimension of N^{(2)}_{3d} equals the defect without independent verification.

Editorial extensions

If this is right

The defect of any projective hypersurface with isolated singularities can be computed as the dimension of the unipotent Milnor fiber cohomology H^n(F_f)_1, giving a single invariant connecting topology, Hodge theory, and monodromy.
For weighted homogeneous isolated singularities, the defect is algorithmically computable from the E2-term of the pole order spectral sequence; the paper provides explicit values for several nodal quintic threefolds.
When X is a threefold with rational singularities (or more generally with no spectral number 1), Corollary 2 gives a formula for the Hodge number of intersection cohomology in terms of the smooth fiber, vanishing cycles, and the defect.
The equality def(X)=dim Coker ι^{n−1} ties the vanishing of the defect to the surjectivity of the natural map into intersection cohomology, which is the self-duality condition relevant to Q-factoriality in the rational-singularity threefold case.

Reading between the lines

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

If the equality holds, the defect can be interpreted as a measure of the 'gap' between ordinary and intersection cohomology; one might test whether similar equalities hold for other classes of singular varieties beyond hypersurfaces, such as complete intersections.
The computational method could be extended to non-weighted-homogeneous singularities if the degeneracy of the pole order spectral sequence can be proved or if higher differentials can be computed; the code could then be adapted to compute higher pages.
The relation defect = dim H^n(F_f)_1 suggests that defect is a unipotent-monodromy invariant; one could compare it with spectral-number distributions to seek new constraints or formulas for defect in terms of spectral numbers.
The paper's examples all have weighted homogeneous singularities; it remains open how far the E2-computation can be pushed, for instance for the non-projective cone example with a single singular point, or whether that example's defect can be reproduced by an independent algorithm.

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.

Desk Editor's Note

read the letter

The thing you should know: the main theorem holds up. Theorem 1 — def(X) equals dim Coker of the intersection-cohomology inclusion, dim Coker of the vanishing-cycle map ρ, dim Im σ, and dim H^n(F_f)_1 — is proved by a terse but standard mixed Hodge module argument. The last equality is well known, but the uniform proof is worth having. That part deserves serious referee time.

What is actually new is Section 3: a computational pipeline for weighted homogeneous isolated singularities, using the pole-order spectral sequence to compute H^3(F_f)_1. The explicit defect values for van Geemen–Werner quintics (19, 18), van Straten (29), Cheltsov (1), the Segre cubic (5), a one-point example (7), and the d=6 experiments (40, 30) are not in the literature. If those numbers are right, the paper is a useful extension of earlier defect computations.

The soft spots are all in Section 3. Theorem 3.1, the E2-degeneration that underpins the whole calculation, is cited only to [Sa 25], an earlier preprint by the same author. Worse, the code identifies dim N^(2)_{3d} with the defect, and Example 3.1 openly says the 'hypothesis of Theorem 2' is practically assumed — but Theorem 2 is never stated. That is a real gap. Formula (3.3) for the rank of d_1 is asserted without proof, and the mod-p rank method in Remark 3.1 is explicitly experimental. So the listed defect values are conditional: if the E2-degeneration or the unnamed hypothesis fails in any example, those numbers are unsupported. The paper hints at a way around the 'Theorem 2' assumption by computing dim N^(2)_{2d}, but it doesn't give those outputs, which is frustrating.

To be fair, the authors flag the assumption; it isn't hidden. But for a paper that advertises explicit computations, shipping code that depends on an unstated theorem is a genuine weakness.

Who is this for: singularity theorists working on defect, Q-factoriality, and nodal threefolds. A reader who wants the theorem gets a clean proof. A reader who wants to rely on the numbers should wait for clarification or verify them independently.

Recommendation: send it to peer review. The theoretical part is solid, and the computational pipeline is potentially valuable. The referee should ask for a precise statement of 'Theorem 2' and a proof or independent verification of the numerical results. Minor: clean up the code listing and report which primes were used.

Referee Report

4 major / 4 minor

Summary. The paper studies the defect def(X)=h^{n+1}(X)-h^{n-1}(X) of a projective hypersurface with isolated singularities. Theorem 1 identifies the defect with the dimension of the cokernel of the intersection-cohomology inclusion, with the cokernel/rank of the vanishing-cycle map in the smoothing exact sequence, and with dim H^n(F_f)_1, the unipotent part of the Milnor fiber cohomology. The proof is via mixed Hodge modules, using the exact sequence (1.8) and its dual. The paper then gives a computational method, based on the pole order spectral sequence and a stated E2-degeneration result cited to [Sa25], and reports numerical defects for several nodal quintic threefolds and other examples, computed by Singular and C code.

Significance. Theorem 1, if correct, is a useful and conceptually clean unification: it ties the classical defect to several a priori independent invariants and to Hodge-module-theoretic objects. The corollaries, especially the relation to intersection cohomology and to Q-factoriality for n=3, are potentially valuable. The paper also provides explicit computational code, which is commendable. However, the computational part is not currently self-contained: it depends on an unproved rank formula, on an unpublished preprint for E2-degeneration, and on an unstated hypothesis identifying a computed dimension with the defect. Thus the numerical results are conditional, while the core theoretical identity appears well supported by standard mixed Hodge module arguments.

major comments (4)

[Section 3, Theorem 3.1] All of Section 3 relies on the assertion that the pole order spectral sequence degenerates at E2 for isolated weighted homogeneous singularities. This is stated as Theorem 3.1 and cited only to [Sa25], an arXiv preprint by one of the authors. The degeneration is not proved or even sketched in the present paper. If this input fails, the dimensions M^(r), N^(r) are not linked to the Milnor fiber cohomology in the way used to compute the defect. Please provide a proof, a published reference, or clearly mark the numerical results as depending on a conjecture.
[Example 3.1, code] The code computes dim N^(2)_{3d}, and the reported values def(X)=19, 18, 29, 1, 5, 7, 40, 30 are obtained by identifying this dimension with the defect. The text states that 'the hypothesis of Theorem 2 is practically assumed' for this identification, but Theorem 2 is nowhere stated in the paper. Without a precise statement of that hypothesis and a verification that each of the listed examples satisfies it, these numerical claims are unsupported. The suggested alternative of also computing dim N^(2)_{2d} is not carried out. Please supply the missing hypothesis and the N^(2)_{2d} outputs, or replace the numerical section with one that does not depend on an unnamed theorem.
[Eq. (3.3)] The rank formula rk d1 = rk φ_k - (rk Gr^G_0 φ_k + rk Gr^G_1 φ_k) is introduced with 'We can verify' but no proof or reference is given. This formula is a load-bearing input in the code: it is used to compute rk d1 and hence the reported Milnor fiber dimensions and defects. A full derivation or a precise citation to a proof must be included before the computations can be accepted.
[Remark 3.1, mod-p rank computation] The C program in Remark 3.1 computes the rank of the matrices over a single prime chosen by the user, and the text explicitly calls this method experimental. A single mod p rank gives only a lower bound for the rational rank, so it does not certify the rank over Q. Since the d=6 example (and possibly others) relies on this computation, please either run several primes and report agreement, or use an exact rank algorithm with certification.

minor comments (4)

[p.2, after Eq. (8)] The phrase 'X has an rational singularity' should be 'X has a rational singularity'.
[Section 3, before Eq. (3.1)] The sentence 'M, N are also denoted as M^(1), N^(1)' is confusing because the superscript notation is never used afterward. Either remove it or state where it is needed.
[Remark 3.1, code comments] The instruction 'Return must be added after .h>' refers to a text-file artifact; please clarify the intended editing step. Also, the input file preprocessing by running the first eight lines of the Singular program could be described more cleanly for reproducibility.
[Example 3.6] The line 'dim V^1 = def(X) = 7' is stated without explaining how V^1 is computed from the code. A one-sentence explanation would help.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: Theorem 1 is derived from independent Hodge-module exact sequences; the numerical examples are conditional on a self-cited E2-degeneration and an unstated 'Theorem 2' identification, but these are assumptions rather than circular reductions.

full rationale

The core derivation (Theorem 1) is self-contained: it uses the exact sequences (2) and (4), the duality (2.1), and standard mixed Hodge-module identifications to show def(X)=dim Coker ι^{n-1}=dim Coker ρ=dim Im σ=dim H^n(F_f)_1. None of these quantities is fitted or defined in terms of another, so the main equality does not reduce to its input. The computational section (Section 3) relies on the E2-degeneration of the pole-order spectral sequence, cited to [Sa 25] (an earlier preprint by the same author); this is a self-citation and is load-bearing for the numerical examples, but it is a stated theorem with independent assumptions (isolated weighted homogeneous singularities) and is not a restatement of the defect equality. The paper itself flags the main limitation: 'the hypothesis of Theorem 2 is practically assumed in order that this dimension coincide with the defect', so the reported values 19,18,29,1,5,7,40,30 are conditional on an unstated identification between dim N^(2)_{3d} and def(X), and are not independently verified. This is a missing assumption / unsupported numerical claim, not a circular derivation: the numbers are not fitted to the target or defined by it. The score 2 reflects one potentially load-bearing self-citation and an explicit unproved identification in the examples, while the central theorem remains independent.

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

No free parameters are fitted: all inputs are fixed polynomials from cited literature and the computations are exact linear algebra (with mod-p reduction for speed). The main uncharged inputs are the standard Hodge-module machinery and the E2-degeneration from [Sa 25], plus two unproved or underspecified computational identifications.

assumptions (5)

standard math Standard theory of mixed Hodge modules: vanishing cycle functors, polarizability, intersection cohomology Hodge modules, and self-duality of intersection cohomology.
Used throughout Section 1 and the proofs in Section 2, cited to [Sa 88], [Sa 90], and [BBD 82]; not proved in this paper.
domain assumption E2-degeneration of the pole order spectral sequence for isolated weighted homogeneous singularities.
Theorem 3.1, cited to [Sa 25] (same author's prior preprint); it is the bridge from spectral sequence E2-terms to the defect.
domain assumption Isomorphism H^k(F_f)_1 ≅ H^k(U) for U = P^{n+1} \ X.
Used in the proof of Theorem 1, Eq. (2.3), cited to [Di 92] and [BuSa 10].
ad hoc to paper Rank formula for d1: rk ϕ_k − (rk Gr^G_0 ϕ_k + rk Gr^G_1 ϕ_k), Eq. (3.3).
Stated as 'We can verify' with no proof or reference; it underlies the Singular code.
ad hoc to paper Computational identification of dim N^{(2)}_{3d} with the defect for the examples.
Code comment in Example 3.1: 'hypothesis of Theorem 2 is practically assumed' — an undefined theorem/assumption needed for the output values.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Defect of projective hypersurfaces with isolated singularities." pith.science (2026). https://pith.science/paper/3234I7ZY

@misc{pith2026251223522,
  author       = {Pith},
  title        = {Pith review of: Defect of projective hypersurfaces with isolated singularities},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3234I7ZY}},
  note         = {Machine review of arXiv:2512.23522}
}

abstract

Let $X$ be a hypersurface with isolated singularities defined by $f$ in ${\bf P^{n+1}}$ with $n>1$. The difference ${\rm def}(X):=h^{n+1}(X)-h^{n-1}(X)$ is called the defect of $X$ (for self-duality of the cohomology of $X$). It is known that its vanishing is closely related to ${\bf Q}$-factoriality of $X$ without assuming rational singularities when $n=3$. This number coincides with the dimension of the cokernel of the inclusion $H^{n-1}(X)\to{\rm IH}^{n-1}(X)$, the rank of the morphism from the vanishing cohomologies of $X$ to $H^{n+1}(X)$ for a one-parameter smoothing of $X$ with total space smooth, and also with the dimension of the unipotent monodromy part of the Milnor fiber cohomology of $f$ with degree $n$. In the case $X$ has only weighted homogeneous isolated singularities, the defect ${\rm def}(X)$ is then given by the $E_2$-term of the spectral sequence of the double complex with differentials ${\rm d}f\wedge$ and $\rm d$ by the $E_2$-degeneration of the pole order spectral sequence. It can be calculated explicitly using a computer even for analogues of the Hirzebruch quintic threefold with more than one hundred ordinary double points found by B.\ van Geemen and J.\ Werner in a compatible way with their computation. We give also an example with ${\rm def}(X)>0$ and $|{\rm Sing}\,X|=1$ in the non-projective cone case where $n=3$.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Roots of Bernstein-Sato polynomials for projective hypersurfaces with ordinary double points
math.AG 2026-07 conditional novelty 7.0 of 10

The upper bound p_f ≤ q_s on the Bernstein-Sato root range is proved sharp for nodal projective hypersurfaces under two explicit combinatorial inequalities, with explicit constructed examples.

Reference graph

Works this paper leans on

36 extracted references · 3 linked inside Pith · cited by 1 Pith paper

[1]

Beilinson, A., Bernstein, J., Deligne, P., Faisceaux pervers, Ast\'erisque 100 (1982)

1982
[2]

Budur, N., Saito, M., Jumping coefficients and spectrum of a hyperplane arrangement, Math.\ Ann.\ 347 (2010), 545--579

2010
[3]

Alg.\ Geom.\ 19 (2010), 781--791

Cheltsov, I., Factorial threefold hypersurfaces, J. Alg.\ Geom.\ 19 (2010), 781--791

2010
[4]

Decker, W., Greuel, G.-M., Pfister, G., Sch\"onemann, H., Singular 4.3.2 --- A computer algebra system for polynomial computations, available at http://www.singular.uni-kl.de (2023)

2023
[5]

Dimca, A., Singularities and Topology of Hypersurfaces, Springer, 1992

1992
[6]

Dimca, A., Saito, M., Some consequences of perversity of vanishing cycles, Ann.\ Inst.\ Fourier 54 (2004), 1769--1792

2004
[7]

Dimca, A., Saito, M., Koszul complexes and spectra of projective hypersurfaces with isolated singularities (arxiv:1212.0436v5)

arXiv
[8]

Dimca, A., Sticlaru, G., Computing Milnor fiber monodromy for some projective hypersurfaces, Contemp.\ Math.\ 742 (2020), 31–52

2020

Show all 36 references

[9]

Fern\'andez de Bobadilla, J., Pallar\'es, I., Saito, M., Hodge modules and cobordism classes, J.\ Eur.\ Math.\ Soc.\ 27 (2025), 773--800

2025
[10]

van Geemen, B., Werner, J., Nodal quintics in ^4 , Lect.\ Notes Math.\ 1399, Springer (1989), 48--59

1989
[11]

Grauert, H., \"Uber Modifikationen und exzeptionelle analytische Mengen, Math.\ Ann.\ 146 (1962), 331--368

1962
[12]

Griffiths, Ph., On the period of certain rational integrals I, II, Ann.\ Math.\ 90 (1969), 460--541

1969
[13]

Hartshorne, R., Algebraic Geometry, Springer, New York, 1977

1977
[14]

Hirzebruch, F., Topological Methods in Algebraic Geometry, Springer, Berlin, 1966

1966
[15]

Hirzebruch, F., Some examples of threefolds with trivial canonical bundle, Collected papers II, Springer 1987, 757--770

1987
[16]

Kerr, M., Laza, R., Saito, M., Deformation of rational singularities and Hodge structure, Alg.\ Geom.\ 9 (2022), 476--501

2022
[17]

Kov\'acs, S.J., Rational, log canonical, Du Bois singularities: on the conjectures of Koll\'ar and Steenbrink, Compos.\ Math.\ 118 (1999), 123--133

1999
[18]

Lindner, N., Hypersurfaces with defect, J.\ Algebra 555 (2020), 1--35 (see also arxiv:1610.04077)

2020 arXiv
[19]

Milnor, J., Singular Points of Complex Hypersurfaces, Princeton Univ.\ Press, 1968

1968
[20]

Namikawa, Y, Steenbrink, J.H.M., Global smoothing of Calabi-Yau threefolds, Inv.\ Math.\ 122 (1995), 403--419

1995
[21]

Maxim, L., Saito, M., Sch\"urmann, J., Spectral Hirzebruch-Milnor classes of singular hypersurfaces, Math.\ Ann.\ 377 (2020), 281–315

2020
[22]

9 (1961), 5--22

Mumford, D., The topology of normal singularities of an algebraic surface and a criterion for simplicity, Publ.\ Math.\ I.H.E.S. 9 (1961), 5--22

1961
[23]

Navarro Aznar, V., On the Chern classes and the Euler characteristic for nonsingular complete intersections, Proc.\ Am.\ Math.\ Soc.\ 78 (1980), 143--148

1980
[24]

Park, S.G., Popa, M., -factoriality and Hodge-Du Bois theory (arxiv:2508.17748)

arXiv
[25]

Reichelt, T., Saito, M., Walther, U., Dependence of Lyubeznik numbers of cones of projective schemes on projective embeddings, Selecta Math.\ (N.S.) 27 (2021), Paper No.\ 6, 22 pp

2021
[26]

Saito, M., Modules de Hodge polarisables, Publ.\ RIMS, Kyoto Univ.\ 24 (1988), 849--995

1988
[27]

Saito, M., Duality for vanishing cycle functors, Publ.\ RIMS, Kyoto Univ.\ 25 (1989), 889--921

1989
[28]

RIMS, Kyoto Univ.\ 26 (1990), 221--333

Saito, M., Mixed Hodge modules, Publ. RIMS, Kyoto Univ.\ 26 (1990), 221--333

1990
[29]

Providence, RI, 1991, pp 509--517

Saito, M., On Koll\'ar's conjecture, Proc.\ Sympos.\ Pure Math., 52, Part 2, A.M.S. Providence, RI, 1991, pp 509--517

1991
[30]

Saito, M., Mixed Hodge complexes on algebraic varieties, Math.\ Ann.\ 316 (2000), 283–331

2000
[31]

Saito, M., On b -function, spectrum and rational singularity, Math.\ Ann.\ 295 (1993), 51--74

1993
[32]

Saito, M., Roots of Bernstein-Sato polynomials of certain homogeneous polynomials with two-dimensional singular loci, Pure Appl.\ Math.\ Q.\ 16 (2020), 1219--1280

2020
[33]

Saito, M., Bernstein-Sato polynomials for projective hypersurfaces with weighted homogeneous isolated singularities (arxiv:1609.04801v11)
[34]

Steenbrink, J.H.M., Limits of Hodge Structures, Inv.\ Math.\ 31 (1976), 229--257

1976
[35]

525--563

Steenbrink, J.H.M., Mixed Hodge structure on the vanishing cohomology, in Real and complex singularities, Sijthoff and Noordhoff, Alphen aan den Rijn, 1977, pp. 525--563

1977
[36]

van Straten, D., A quintic hypersurface in ^4 with 130 nodes, Topology 32 (1993), 857--864

1993

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.

[1] [1]

Beilinson, A., Bernstein, J., Deligne, P., Faisceaux pervers, Ast\'erisque 100 (1982)

1982

[2] [2]

Budur, N., Saito, M., Jumping coefficients and spectrum of a hyperplane arrangement, Math.\ Ann.\ 347 (2010), 545--579

2010

[3] [3]

Alg.\ Geom.\ 19 (2010), 781--791

Cheltsov, I., Factorial threefold hypersurfaces, J. Alg.\ Geom.\ 19 (2010), 781--791

2010

[4] [4]

Decker, W., Greuel, G.-M., Pfister, G., Sch\"onemann, H., Singular 4.3.2 --- A computer algebra system for polynomial computations, available at http://www.singular.uni-kl.de (2023)

2023

[5] [5]

Dimca, A., Singularities and Topology of Hypersurfaces, Springer, 1992

1992

[6] [6]

Dimca, A., Saito, M., Some consequences of perversity of vanishing cycles, Ann.\ Inst.\ Fourier 54 (2004), 1769--1792

2004

[7] [7]

Dimca, A., Saito, M., Koszul complexes and spectra of projective hypersurfaces with isolated singularities (arxiv:1212.0436v5)

arXiv

[8] [8]

Dimca, A., Sticlaru, G., Computing Milnor fiber monodromy for some projective hypersurfaces, Contemp.\ Math.\ 742 (2020), 31–52

2020