Geometric realizations of non-symplectic involutions on the Hilbert square of a K3 surface

Ana Quedo

REVIEW 5 major objections 6 minor 21 references

Geometric realizations of non-symplectic involutions on the Hilbert square of a K3 surface

T0 review · 5 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Non-symplectic involutions on the Hilbert square of a K3 surface are built explicitly by conjugating natural involutions with Beauville involutions.

desk verdict The explicit Beauville-conjugated example in Theorem 40 is real and worth knowing; Theorem 39 is plausible but rests on two unproved Pell claims that should be supplied. read the letter →

arxiv 2412.12307 v1 pith:5GVEKA6F submitted 2024-12-16 math.AG math.DG

Ana Quedo This is my paper · ORCID

classification math.AGmath.DG MSC 14J5014J2814J35

keywords non-symplecticinvolutionsHilbertsquareK3surfacesirreducibleholomorphicsymplecticmanifoldsBeauvilleinvariantlatticesdeformationequivalencegeneralizedPellequations

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's goal is to turn cohomological existence statements about non-symplectic involutions on the Hilbert square $S^{[2]}$ of a K3 surface into maps one can write down and compute. It works with a special family of K3 surfaces whose Picard lattice is generated by two divisors $H,W$ with $H^2=4$, $W^2=2$, and $H\cdot W=8n$, so their Hilbert squares carry two Beauville involutions. The paper proves that the abstract involution $\sigma$ deforms to either composition $\kappa_1=i_2i_1i_2$ or $\kappa_2=i_1i_2i_1$ through Hilbert squares of K3 surfaces exactly when the polarization half-square $t$ is $1+(64n^2-7)^2$ or the companion quartic family in $k$. It then constructs a non-natural involution $\iota=i_1\phi^{[2]}i_1$ with invariant lattice $\langle2\rangle\oplus\langle-2\rangle$, giving an explicit geometric model for a classified case that previously had only non-geometric or natural examples. The payoff is that these involutions can be evaluated on subschemes, not just described by their action on $H^2$.

What carries the argument

Beauville involutions and anti-reflections. A Beauville involution on the Hilbert square of a smooth quartic K3 surface sends a length-2 subscheme to the residual intersection of its line with the quartic, and it acts on $H^2(S^{[2]},\mathbb{Z})$ as the anti-reflection $v\mapsto\langle v,D\rangle D-v$ in an ample square-2 class $D=[H]-\delta$. Conjugating any automorphism by a Beauville involution transports invariant lattices by that isometry (Proposition 36). On the auxiliary surfaces $S_n$, the square-2 divisor $W$ gives a natural involution $\phi^{[2]}$, while $[H]-\delta$ and $[8nW-H]-\delta$ give two Beauville involutions $i_1,i_2$; their compositions $\kappa_1,\kappa_2$ have rank-1 invariant lattices generated by explicit square-2 classes $D_1,D_2$. The Pell-equation criterion from the existence theorem selects the $t$-values for which the abstract $\sigma$ has the same divisor data as $D_1$ or $D_2$, making the deformation possible.

What would settle it

Check the two families numerically: find integers $(x,y)$ solving $x^2-4ty^2=5$ for $t=(64n^2-7)^2+1$ or $t=2^{12}\cdot5^2k^4-2^7\cdot7k^2+2$ with small $n,k$. One solution would violate the hypothesis on which Theorem 39 rests; a proof of non-solubility, or a verified finite search, would remove the gap.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that two lattice-theoretically predicted involutions on Hilbert squares admit explicit geometric models. Theorem 39 states that for $(S,L)$ with $\operatorname{Pic}S=\mathbb{Z}L$ and $L^2=2t$, the unique non-symplectic involution $\sigma$ on $S^{[2]}$ deforms to $\kappa_1=i_2i_1i_2$ or $\kappa_2=i_1i_2i_1$ on $S_n^{[2]}$, with every fibre a Hilbert square of a $2t$-polarized K3 surface, exactly when $t=1+(64n^2-7)^2$ or $t=2^{12}\cdot5^2k^4-2^7\cdot7k^2+2$. Theorem 40 exhibits $\iota=i_1\phi^{[2]}i_1$ as a non-natural non-symplectic involution whose invariant lattice is spanned by $8nH-W-8n\delta$ and $2H-3\delta$, therefore isometric to $\langle2\rangle\oplus\langle-2\rangle$, with orthogonal complement $U^{\oplus2}\oplus E_8(-1)^{\oplus2}\oplus\langle-2\rangle$.

Load-bearing premise

The load-bearing assumption is the unproved assertion that the generalized Pell equation $x^2-4ty^2=5$ has no integer solution for $t=1+(64n^2-7)^2$ and for $t=2^{12}\cdot5^2k^4-2^7\cdot7k^2+2$; Theorem 39 needs this to invoke the existence and uniqueness of $\sigma$.

Editorial extensions

If this is right

For the two families of $t$ in Theorem 39, the abstract non-symplectic involution on $S^{[2]}$ is realized as a composition of two Beauville involutions on $S_n^{[2]}$, so it can be evaluated geometrically rather than only cohomologically.
The deformation connecting $(S^{[2]},\sigma)$ to $(S_n^{[2]},\kappa_i)$ can be chosen so every fibre is $\Sigma^{[2]}$ for a $2t$-polarized K3 surface $\Sigma$, so the model remains inside Hilbert squares throughout.
The involution $\iota=i_1\phi^{[2]}i_1$ is a concrete non-natural involution with invariant lattice $\langle2\rangle\oplus\langle-2\rangle$ and the natural-embedding orthogonal complement, providing a geometric representative of that lattice-theoretic case.
The explicit generators $8nH-W-8n\delta$ and $2H-3\delta$ of the invariant lattice make the action of $\iota^*$ on $\operatorname{NS}(S_n^{[2]})$ explicit through the displayed $3\times3$ matrix.

Reading between the lines

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

Editorial inference: the same conjugation recipe could produce geometric realizations of other classified non-symplectic automorphisms whenever an ample square-2 divisor and a Beauville involution coexist on a Hilbert square.
Editorial inference: the explicit invariant classes make a computation of the fixed locus of $\iota$ feasible; a description of that fixed locus would give an independent check of the realization.
Editorial inference: the two $t$-families are singled out by divisibility of the coefficients in $D_1,D_2$; analogous divisibility conditions for larger $b$ may yield further deformation families beyond the two stated.
Editorial inference: substituting other natural involutions for $\phi^{[2]}$ in the conjugation $i_1(\cdot)i_1$ would produce a family of non-natural involutions whose invariant lattices are the Beauville-images of the starting natural invariant lattices.

Share X Bluesky LinkedIn Reddit HN

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

Short version: this is a useful paper with one solid new construction and one conditional theorem. Theorem 40 gives an explicit non-natural involution on S_n^[2] with invariant lattice <2>⊕<-2>, by conjugating the natural involution with a Beauville involution. I checked the matrix and the lattice computation; they work. The invariant lattice is spanned by two classes with the right squares and zero mutual pairing, and the orthogonal complement matches the natural-involution case. This distinguishes the example from Ohashi–Wandel and gives people working on automorphisms of Hilbert squares an actual map to compute with. The construction does not assume the target result; it is an explicit composition, so there is no circularity. The reliance on the author's earlier paper [19] is also legitimate: the S_n family is prior published work used as input, not tuned to force the conclusion.

Theorem 39 is a different matter. The idea is sensible: use deformation equivalence to transport the abstract anti-reflection involution to a composition of Beauville involutions on the special surfaces S_n. The deformation framework is in place, and Corollary 38 computes the relevant divisors correctly. But the proof skips two number-theoretic checks that are load-bearing. For the second family, it is asserted that (a,b) = (1600k^2−7,5) is the minimal positive solution of P_t(−1); minimality is not shown and is not obvious, since one must rule out y<5. Separately, the condition that P_{4t}(5) has no solution is asserted with "since we can verify" and no verification. A mod-8 argument may indeed handle both stated families, as the stress-test note says, but the text does not give it. These are fixable gaps, not fatal flaws, but right now Theorem 39 is conditional on unstated arithmetic facts.

Minor issue: the theorem statement has "j=1,2" while referring to κ_i; the indexing is sloppy but the meaning is clear.

Who is this for? Specialists in hyperkähler automorphisms and K3 surfaces. Theorem 40 is a citable example; Theorem 39 is a reasonable claim backed by a proven deformation framework once the Pell checks are filled in. I would send this to a serious referee rather than desk-reject. The missing arithmetic arguments are exactly the kind of thing a referee can require in revision.

Referee Report

5 major / 6 minor

Summary. The paper studies non-symplectic involutions on the Hilbert square of K3 surfaces, focusing on explicit geometric constructions. For a rank-two Picard lattice generated by H and W with intersection matrix [[4, 8n], [8n, 2]], the author constructs Beauville involutions i1, i2 on S_n^[2] and studies the compositions κ1 = i2 i1 i2 and κ2 = i1 i2 i1. Theorem 39 claims that the unique non-symplectic involution on S^[2] for a 2t-polarized K3 surface deforms into these κi exactly when t belongs to two explicit families. Theorem 40 constructs the non-natural involution ι = i1 φ[2] i1 with invariant lattice ⟨2⟩ ⊕ ⟨-2⟩ and orthogonal complement U^2 ⊕ E8(-1)^2 ⊕ ⟨-2⟩. The paper presents explicit matrix computations and relies on prior classification results; the main issue is a sign error in the formula for D1 and several unproved number-theoretic assertions in the proof of Theorem 39.

Significance. If the sign correction and the missing number-theoretic verifications are supplied, the paper gives genuinely new explicit geometric realizations of non-natural non-symplectic involutions, complementing the existence results of Boissière-Cattaneo-Nieper-Wiesskirchen-Sarti and Boissière-Camere-Sarti. The computation in Theorem 40 is transparent and can be checked by hand, and the construction by conjugating a natural involution with a Beauville involution is elegant. Theorem 39 is conditional on elementary Pell-equation checks that are likely easy to fill; the main obstruction is the sign error and the terse 'only if' part of the proof.

major comments (5)

[Section 3, Corollary 38] The formula for D1 is incorrect. For n = 1, the stated D1 = 59H + 464W - 57δ has Beauville-Bogomolov square 876034, not 2. The correct class is D1 = -(64n^2-5)H + 8n(64n^2-6)W - (64n^2-7)δ, which has square 2. This error propagates to the definition of L1 in the proof of Theorem 39, where the positive H coefficient is used; with the corrected sign, one obtains L1^2 = 2(1 + (64n^2-7)^2) as required for the first family t = (64n^2-7)^2 + 1. As written, the proof of Theorem 39 is invalid, though the theorem statements may remain true after the correction.
[Section 3, proof of Theorem 39] The assertion 'since we can verify that P_{4t}(5) does not have a solution' is not substantiated. For both claimed families one has t ≡ 2 mod 8, so 4t ≡ 0 mod 8, and the congruence x^2 ≡ 5 mod 8 has no solution. This one-line verification should be included because the absence of solutions to P_{4t}(5) is a hypothesis of Theorem 29 and Proposition 34 and is load-bearing for the existence of the involution σ.
[Section 3, proof of Theorem 39] For the b = 5 family, the minimality of (a,b) = (1600k^2-7, 5) as the minimal positive solution of x^2 - t y^2 = -1 is not proved. The text only says that the only possible value for a is 64n^2-7, but this does not rule out a smaller solution with y < 5. Proposition 34 requires the minimal solution, so a verification must be supplied, for instance by checking the finitely many possible y = 1,2,3,4 via a congruence or descent argument.
[Section 2.1, Proposition 5 and Section 3, Proposition 37] Proposition 5 is stated for d_n = 4(n^2-2) with n a multiple of 4, but Proposition 37 needs the non-existence of solutions to x^2 - 8(8n^2-1)y^2 = -8 for every n > 1. The stated proposition does not apply to this d_n. The proof of Proposition 5 actually works verbatim for d_n = 8(8n^2-1) without a congruence condition on n, so the statement should be corrected; as written, the absence of (-2)-curves, and hence the existence of the Beauville involutions i1 and i2, is not established.
[Section 3, Theorem 39] The 'only if' direction is only sketched. The proof does not explicitly justify that deformation equivalence of (S^[2],σ) with (S_n^[2],κ_i) forces the invariant lattice generator of κ_i to coincide, up to the natural isometry, with b[L_i] - aδ for the minimal solution (a,b) of P_t(-1). Without this identification, the conclusion that the two listed families are the only possible t-values is not fully supported; the argument should spell out how the minimal solution enters and why no other t can occur.

minor comments (6)

[Throughout] There are several typographical issues: 'Acknowlegments' should be 'Acknowledgments', 'lenght' should be 'length', and 'sympletic' should be 'symplectic'.
[Section 3, Theorem 39] The indices in the phrase 'κi, j=1,2' should read 'κ_i, i=1,2'.
[Section 3, Theorem 39] The second family is printed with ambiguous superscripts; the derivation in the proof indicates it should read t = 2^12 · 5^2 k^4 - 2^7 · 7 k^2 + 2, not the variant with 2^5 that the current rendering suggests.
[Section 4, Theorem 2] The notation '< −2 > ⊕< 2>' should use the standard lattice notation '⟨−2⟩ ⊕ ⟨2⟩'.
[Section 3, Proposition 37] The claim that d_n is not a square and hence there are no nontrivial 0-divisors is true, but it would be helpful to include the short congruence argument (squares mod 64 are not congruent to -8 mod 64).
[References] The reference [19] is listed with a 2025 volume/year; please update to the final published version if it has appeared.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the geometric constructions are explicit conjugations by Beauville involutions, and the unproved Pell verifications in Theorem 39 are correctness gaps, not circular reductions.

full rationale

The paper's derivations are explicit and do not assume the target results. Theorem 2/40 constructs ι = i1 φ[2] i1 as a conjugate of the natural involution φ[2] by the Beauville involution i1; Proposition 36 gives Inv(ι) = i1*(<W,δ>) = <8nH−W−8nδ, 2H−3δ>, and the squares and the orthogonal complement are then read off from the ⟨2⟩⊕⟨−2⟩ isometry. This is a direct computation, with no fitted parameter and no pre-supposed invariant lattice. Theorem 1/39 uses the Boissière–Cattaneo–Nieper-Wisskirchen–Sarti existence theorem (Theorem 29) and Beri-type deformation Proposition 34 as inputs; the proof solves D_i = b[L_i] − aδ for t, which is reverse design of the permitted t-values, not a prediction forced by a fit. The only self-citation, [19] by Paiva and Quedo, supplies the K3 family S_n, an elementary lattice lemma, and the square-two involution φ; these are prior results whose stated assumptions do not include Hilbert-square involutions, so under the review rules they count as independent support and do not raise the circularity score. I nonetheless flag, per the reviewing rule, two omitted proofs in the proof of Theorem 39: the sentence 'since we can verify that P4t(5) does not have a solution' supplies no verification, and the minimality of (a,b)=(1600k^2−7,5) as the minimal positive solution of P_t(−1) is asserted without argument. These are correctness gaps in the conditional chain, not circular reductions: the target claim is not used as an input. Overall no significant circularity; score 0.

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

The paper introduces no fitted constants and no new entities. Its input consists of integers n, k, t and variables a, b in Pell equations; all are exact. The main external inputs are the surfaces S_n from [19] and the classification theorems from [5], [6], [7], [8]. The unproved Pell non-solvability is the most fragile imported assumption.

assumptions (5)

domain assumption Existence of the K3 surfaces S_n with Picard lattice described by Q_n, including the ample square-2 divisor W and the automorphism φ with H^2(S_n,Z)^φ=<W>, from [19, Example 1 and Proposition 12].
Section 3 and 4 use these surfaces as input; the paper cites [19] and does not re-prove their construction.
domain assumption Classification and existence results: [6, Theorem 5.5] (existence and uniqueness of the involution σ), [5, Theorem 3.1], [7, Theorem 8.5 and Proposition 8.2], and [8, Theorem 4.5 and 5.6] (moduli and deformation equivalence).
These prior theorems supply the abstract existence and the deformation framework that the new geometric realizations plug into.
domain assumption No (-2)-divisors and no nontrivial 0-divisors on S_n, so the positive cone equals the ample cone (from [19, Lemma 9], Proposition 5, Lemma 11).
Needed in Proposition 37 to conclude H and 8nW-H are very ample and define Beauville involutions.
ad hoc to paper The assertion that P_{4t}(5) has no solution for t=(64n^2-7)^2+1 and for t=2^12·5^2k^4-2^7·7k^2+2 is used without proof.
It is stated as 'we can verify' in the proof of Theorem 39 and is required to apply [6, Theorem 5.5]; no modular argument is given.
standard math Verbitsky's Global Torelli theorem and the Beauville-Bogomolov form on H^2.
Standard background used throughout for identifying automorphisms with lattice isometries.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Geometric realizations of non-symplectic involutions on the Hilbert square of a K3 surface." pith.science (2026). https://pith.science/paper/5GVEKA6F

@misc{pith2026241212307,
  author       = {Pith},
  title        = {Pith review of: Geometric realizations of non-symplectic involutions on the Hilbert square of a K3 surface},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5GVEKA6F}},
  note         = {Machine review of arXiv:2412.12307}
}

read the original abstract

We give new examples of geometric constructions of non-natural non-symplectic involutions of IHS manifolds whose existence is guaranteed by previous results of Bossi\`ere-Cattaneo-Nieper-Wiesskirchen-Sarti in arXiv:1410.8387 and Bossi\'ere-Camere-Sarti in arXiv:1402.5154.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 21 canonical work pages

[2]

Beauville

A. Beauville. Variétés Kähleriennes dont la première cl asse de Chern est nulle. J. Diﬀerential Geom. , 18(4):755–782, 1983

work page 1983
[19]

Paiva and A

D. Paiva and A. Quedo. Automorphisms of quartic surface s and Cremona transformations. Journal of Pure and Applied Algebra, 229(1):107850, 2025

work page 2025
[1]

Beauville

A. Beauville. Some remarks on Kähler manifolds with c1 = 0 . Classiﬁcation of algebraic and analytic manifolds. PM, 39:1–26, 1983

work page 1983
[3]

P. Beri. On birational transformations and automorphisms of some Hy perkahler manifolds . PhD thesis, 2020

work page 2020
[4]

Beri and A

P. Beri and A. Cattaneo. On birational transformations o f Hilbert schemes of points on K3 surfaces. Mathematische Zeitschrift, 301(2):1537–1554, 2022

work page 2022
[5]

Boissiére, A

S. Boissiére, A. Cattaneo, D. Markushevich, and A. Sarti . On the non-symplectic involutions of the Hilbert square of a K3 surface. Izvestiya: Mathematics , 83(4):731

work page
[6]

Boissiére, A

S. Boissiére, A. Cattaneo, M. Nieper-Wisskirchen, and A . Sarti. The Automorphism Group of the Hilbert Scheme of Two Points on a Generic Projective K3 Surface , pages 1–15. Springer International Publishing, Cham, 201 6

work page
[7]

Boissière, C

S. Boissière, C. Camere, and A. Sarti. Classiﬁcation of a utomorphisms on a deformation family of hyperkähler four-f olds by p-elementary lattices. Kyoto J. Math. , 56:465–499, 2016

work page 2016

Show all 21 references

[8]

Boissière, C

S. Boissière, C. Camere, and A. Sarti. Complex ball quoti ents from manifolds of K3[n]-type. J. Pure Appl. Algebra , 223(3):1123–1138, 2019

2019
[9]

Boissière and A

S. Boissière and A. Sarti. A note on automorphism and bira tional transformations of holomorphic symplectic manifol ds. Proc. Amer. Math. Soc. , 12:4053–4062, 2012

2012
[10]

K. Conrad. Pell’s equation, ii. https://kconrad.math.uconn.edu/blurbs/ugradnumthy/pelleqn2.pdf
[11]

Huybrechts

D. Huybrechts. Lectures on K3 surfaces, volume 158 of Camb. Stud. Adv. Math. Cambridge: Cambridge University Press, 2016

2016
[12]

M. Joumaah. Non-symplectic involutions of irreducibl e symplectic manifolds of K3[n]-type. Mathematische Zeitschrift , 283(3-4):761–790, 2016

2016
[13]

Lagrange

J. Lagrange. Solution d’un problème d’arithmétique,Oeuvres de Lagrang e, Tome I . Gauthier Villars, 1867
[14]

S. Mori. On degrees and genera of curves on smooth quarti c surfaces in P3. Nagoya Math. J , 96:127–132, 1984

1984
[15]

Morrison

D. Morrison. On K3 surfaces with large Picard number. Invent Math , 75:105–121, 1984

1984
[16]

K. O’Grady. A new six-dimensional irreducible symplec tic variety. J. Algebraic Geom , 12(3):435–505, 2003

2003
[17]

Ohashi and M

H. Ohashi and M. W andel. Non-natural non-symplectic in volutions on symplectic manifolds of K3[2]. arXiv:1305.6353v2, 2013

2013 arXiv
[18]

K. O’Grady. Involutions and linear systems on holomorp hic symplectic manifolds. Geom. Funct. Anal .15 , 6:1223–1274, 2005. 12 ANA QUEDO

2005
[20]

Saint-Donat

B. Saint-Donat. Projective models of K3 surfaces. Am. J. Math. , 96:602–639, 1974

1974
[21]

Verbitsky

M. Verbitsky. Mapping class group and a global Torelli t heorem for hyperkähler manifolds. Duke Math J. , 162:2929–2986, 2013. Dipartimento di Matematica - Università di Bologna Piazza d i Porta S. Donato, 5, 40126 Bologna BO Email address : ana.martinsquedo@unibo.it

2013

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [2]

Beauville

A. Beauville. Variétés Kähleriennes dont la première cl asse de Chern est nulle. J. Diﬀerential Geom. , 18(4):755–782, 1983

work page 1983

[2] [19]

Paiva and A

D. Paiva and A. Quedo. Automorphisms of quartic surface s and Cremona transformations. Journal of Pure and Applied Algebra, 229(1):107850, 2025

work page 2025

[3] [1]

Beauville

A. Beauville. Some remarks on Kähler manifolds with c1 = 0 . Classiﬁcation of algebraic and analytic manifolds. PM, 39:1–26, 1983

work page 1983

[4] [3]

P. Beri. On birational transformations and automorphisms of some Hy perkahler manifolds . PhD thesis, 2020

work page 2020

[5] [4]

Beri and A

P. Beri and A. Cattaneo. On birational transformations o f Hilbert schemes of points on K3 surfaces. Mathematische Zeitschrift, 301(2):1537–1554, 2022

work page 2022

[6] [5]

Boissiére, A

S. Boissiére, A. Cattaneo, D. Markushevich, and A. Sarti . On the non-symplectic involutions of the Hilbert square of a K3 surface. Izvestiya: Mathematics , 83(4):731

work page

[7] [6]

Boissiére, A

S. Boissiére, A. Cattaneo, M. Nieper-Wisskirchen, and A . Sarti. The Automorphism Group of the Hilbert Scheme of Two Points on a Generic Projective K3 Surface , pages 1–15. Springer International Publishing, Cham, 201 6

work page

[8] [7]

Boissière, C

S. Boissière, C. Camere, and A. Sarti. Classiﬁcation of a utomorphisms on a deformation family of hyperkähler four-f olds by p-elementary lattices. Kyoto J. Math. , 56:465–499, 2016

work page 2016