REVIEW 2 major objections 5 minor 10 references
A remark on a theorem of Narasimhan and Ramanan
T0 review · 2 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash
Pith's one-line read The paper proves that for every point of the moduli space $\mathrm{SU}_X(2)$, the anticanonical Seshadri constant is at least $4$, so $\mathrm{SU}_X(2)\cong\mathbb{P}^3$.
desk verdict Coherent but flawed attempt to re-prove Narasimhan-Ramanan via Seshadri constants; the Quot scheme gap is real and blocks the proof. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The machinery is the generalized $\theta$ divisor $\Theta$, the ample generator of $\operatorname{Pic}(\mathrm{SU}_X(2))$, together with the canonical identification $-\mathcal{K}_{\mathrm{SU}_X(2)}\cong\Theta^4$. The proof first establishes that $\Theta$ is base point free, using Beauville's base-locus criterion and a degeneration argument, then for any curve $D$ through $x$ chooses an effective divisor in $|\Theta|$ through $x$ not containing $D$, and uses $D\cdot\Theta \ge \operatorname{mult}_x D$ to get $\varepsilon(-\mathcal{K}_{\mathrm{SU}_X(2)}, x)\ge 4$. The Bauer–Szemberg criterion converts this into the isomorphism with $\mathbb{P}^3$.
What would settle it
Compute $h^0(E)$ for a stable rank-2 trivial-determinant bundle $E$ on a genus-2 curve; Riemann–Roch gives $h^0(E)=0$, so no surjection $V\otimes\mathcal{O}_X\to E$ can induce an isomorphism $V\to H^0(E)$. Checking this one number shows whether the preimage used in Case 2 of Proposition 2.2 exists.
Extended reading notes
Core claim
On its own terms, the central discovery is that the anticanonical Seshadri constant of $\mathrm{SU}_X(2)$ saturates the maximal possible value $4$ at every point. Since $\dim \mathrm{SU}_X(2)=3$, the Bauer–Szemberg criterion then forces $\mathrm{SU}_X(2)\cong\mathbb{P}^3$. The proof reaches the value $4$ by showing the generalized $\theta$ divisor $\Theta$ is base point free, using the relation $-\mathcal{K}_{\mathrm{SU}_X(2)}\cong\Theta^4$, and applying the intersection inequality $D\cdot\Theta \ge \operatorname{mult}_x D\,\operatorname{mult}_x\Theta$ for any curve $D$ through $x$.
Load-bearing premise
The degeneration step assumes that a stable bundle has a preimage in the open set $R^{ss}$ used to build the moduli space, even though a stable bundle has no nonzero sections and so cannot be represented in that open set; without such a preimage the curve connecting the stable bundle to $L\oplus L^{-1}$ may not exist.
Editorial extensions
If this is right
- The anticanonical Seshadri constant at every point of $\mathrm{SU}_X(2)$ is exactly $4$, saturating the universal upper bound $\varepsilon(-\mathcal{K}_Y,x)\le \dim Y+1$ for smooth Fano varieties.
- By the Bauer–Szemberg criterion, $\mathrm{SU}_X(2)$ is isomorphic to $\mathbb{P}^3$, recovering the Narasimhan–Ramanan theorem.
- Under this isomorphism the generalized theta divisor $\Theta$ corresponds to the hyperplane class $\mathcal{O}_{\mathbb{P}^3}(1)$, so the base-point-free linear system $|\Theta|$ gives the isomorphism explicitly.
- The proof shows that no semistable rank-2 trivial-determinant bundle on a genus-2 curve has $H^0(M\otimes E)\ne 0$ for every $M\in\operatorname{Pic}^1(X)$; this is exactly the base-point-freeness statement.
Reading between the lines
- Implicit in the proof is a general criterion: if $Y$ is a smooth Fano $n$-fold with $-\mathcal{K}_Y=(n+1)H$ for an ample globally generated $H$, then $\varepsilon(-\mathcal{K}_Y,y)\ge n+1$ for all $y$, so $Y\cong\mathbb{P}^n$.
- The same Seshadri-constant strategy could be used on other moduli spaces where the anticanonical bundle is known to be a multiple of a base-point-free theta-like divisor; the required value is just the multiple, so one asks when that multiple equals $\dim+1$.
- The degeneration argument can be repaired, if needed, by working with quotients of $V\otimes\mathcal{O}_X(-n)$ for which $H^0(q(n))$ is an isomorphism; the rest of the proof would then run unchanged.
- Because the theorem itself was already known, the contribution here is primarily a template for recognizing projective-space moduli by a numerical invariant rather than by constructing the isomorphism.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper claims an alternative proof of the Narasimhan–Ramanan theorem that the moduli space SU_X(2) of semistable rank-2 degree-0 trivial-determinant bundles on a smooth genus-2 curve X is isomorphic to P^3. The strategy is to apply the Bauer–Szemberg criterion for projective spaces to the Fano threefold SU_X(2) by showing that the Seshadri constant of -K_{SU_X(2)} at any point is at least 4. After reviewing the GIT construction of SU_X(2) as a quotient of an open subset Rss of a Quot scheme, the paper proves (Prop. 2.2) that every semistable rank-2 trivial-determinant bundle E admits a line bundle M of degree 1 with H^0(E⊗M)=0, which is used to show that the generalized theta divisor Θ is base point free (Cor. 2.3). A lemma (Lemma 3.1) then converts base point freeness into a lower bound on the Seshadri constant, giving ε(-K,x) ≥ 4 and hence SU_X(2) ≅ P^3.
Significance. If the proof were correct, it would give a concise new proof of a classical theorem and would usefully demonstrate a Seshadri-constant technique for moduli spaces. The paper is well organized and cites the necessary external results (Bauer–Szemberg, Drezet–Narasimhan, Newstead, Beauville) without invoking the theorem under proof, so there is no circularity. However, the proof of Prop. 2.2 for stable bundles is invalid because the chosen Quot open set Rss excludes stable bundles, and both the base point freeness and the Seshadri constant bound rely on that proposition. The overall approach may be salvageable with a modified Quot construction, but the manuscript as written does not establish the claimed theorem.
major comments (2)
- [§2, definition of Rss; Prop. 2.2, Case 2] Let E be a stable rank-2 degree-0 bundle with trivial determinant on X. Since any nonzero section of E would give an inclusion O_X→E whose saturation is a line subbundle of degree ≥0, contradicting stability, we have h^0(E)=0. Consequently, for any vector space V, there is no surjection q: V⊗O_X→E with H^0(q): V→H^0(E) an isomorphism: a nonzero V cannot map surjectively onto a zero vector space, and V=0 would force E=0. Thus stable bundles have no representatives in Rss. In Proposition 2.2, Case 2, the sentence 'We take a preimage of E in Rss' is therefore impossible for a stable E, so the curve C provided by Lemma 2.1 does not exist, and the semicontinuity/Grauert argument cannot begin. As a result, Prop. 2.2 is unproved for stable bundles, and the base point freeness of Θ (Cor. 2.3) and the inequality ε(-K,x) ≥ 4 (Thm. 1.2) are unsupported. The standard remedy—using quotients of V⊗O_X(-n) with H^0(q(n)) an isomorphism for n sufficiently large—is not present in the paper and would require reworking the proof.
- [§2, Rss and Lemma 2.1] Even for the strictly semistable endpoint L⊕L^{-1} of the intended degeneration, membership in Rss depends on h^0(L⊕L^{-1}). If L is nontrivial, h^0(L⊕L^{-1})=0; if L = O_X, h^0 = 2. Hence for a fixed nonzero V of dimension N, the endpoint lies in Rss only if N equals the relevant h^0, which cannot hold for all L simultaneously. The curve C in Rss connecting E to a bundle of the form L⊕L^{-1} therefore cannot be guaranteed by Lemma 2.1 for the Quot scheme as defined. This reinforces the conclusion that the chosen Rss is the wrong arena for the degeneration argument.
minor comments (5)
- [Abstract and Introduction] There are several typos, e.g., 'nota ble' in the abstract, and 'diviosr' and 'correspodence' in the introduction; these should be corrected.
- [Prop. 2.2] The statement says 'there exists a line bundle M ∈ Pic^1(C)' but the moduli space is Pic^1(X); this is presumably a typo.
- [Prop. 2.2, Case 2] In the Serre duality computation, the degree of M^{-3}⊗U_c^∨⊗K_X is -1, not -2 as stated; the conclusion h^0=0 still holds because the degree is negative.
- [Lemma 3.1] The symbol E is used both for the vector bundle in earlier sections and for an effective divisor in the linear system |Θ|; this notational clash should be avoided.
- [Lemma 3.1] The sentence 'otherwise the unique section s has to be a non vanishing section' is terse; it should explicitly say that, because Θ is globally generated, a single section would have to be nowhere vanishing, contradicting ampleness of Θ.
Circularity Check
No circularity: the Seshadri-constant proof uses independent external theorems; the likely flaw in Prop. 2.2 is a gap, not a circular step.
full rationale
The paper re-proves Narasimhan–Ramanan's theorem that SU_X(2) ≅ P^3 by establishing the Bauer–Szemberg Seshadri-constant criterion. Every substantive input is external to the claim being proved: the construction of SU_X(2) as a GIT quotient, Pic ≅ Z and −K ≅ Θ^4 from Drezet–Narasimhan, smoothness from Laszlo, irreducibility from Newstead, the curve lemma from Mumford–Ramanujam–Manin, and the Beauville description of the base locus of Θ. The paper's own Prop. 2.2 and Lemma 3.1 do not invoke the Narasimhan–Ramanan isomorphism, and no fitted parameter or self-citation is used to force the conclusion. The likely fatal gap noted by a close reading—that stable degree-0 bundles on a genus-2 curve may have h^0 = 0 or 1 and therefore need not admit a representative in Rss with H^0(q) an isomorphism—is a correctness concern, not a circularity concern. It shows the proof may fail on its own terms, not that it assumes its target. Thus no circular step is exhibited, and the score is 0.
Assumptions & free parameters
assumptions (6)
- standard math Theorem 1.1 (Bauer-Szemberg): a smooth Fano n-fold with ε(-K_Y, y)=n+1 for some y is P^n.
- standard math Pic(SU_X(2)) ≅ Z, with ample generator Θ and -K_SU_X(2) ≅ Θ^4.
- standard math Beauville's criterion: the base locus of Θ is {E : H^0(M⊗E) ≠ 0 for all M in Pic^1}.
- ad hoc to paper Rss is irreducible and contains every semistable bundle in SU_X(2) as a point.
- standard math Lemma 2.1: an irreducible curve through any two points in an irreducible quasi-projective variety.
- standard math Upper bound ε(-K_Y, y) ≤ n+1 for a smooth Fano n-fold.
Cite this review
Pith. "Pith review of A remark on a theorem of Narasimhan and Ramanan." pith.science (2026). https://pith.science/paper/V4MX52UG
@misc{pith2026241115774,
author = {Pith},
title = {Pith review of: A remark on a theorem of Narasimhan and Ramanan},
year = {2026},
howpublished = {\url{https://pith.science/paper/V4MX52UG}},
note = {Machine review of arXiv:2411.15774}
}
abstract
In this short note, we provide an alternative proof of a notable theorem by Narasimhan and Ramanan. The theorem states that the moduli space of $S$-equivalence classes of semistable rank $2$ vector bundles over a curve $X$ of genus $2$ with trivial determinant is isomorphic to $\mathbb{P}^3$. Our proof relies on a criterion by Bauer and Szemberg, which characterizes projective spaces among smooth Fano varieties using Seshadri constants.
Reference graph
Works this paper leans on
-
[1]
Vector bundles on curves and generalized theta functions: recent results and open problems
Arnaud Beauville. Vector bundles on curves and generalized theta functions: recent results and open problems, 1994. http://arxiv.org/abs/alg-geom/9404001 arXiv:alg-geom/9404001
work page Pith review arXiv 1994
-
[2]
Seshadri constants and the generation of jets
Thomas Bauer and Tomasz Szemberg. Seshadri constants and the generation of jets. Journal of Pure and Applied Algebra , 213(11):2134--2140, 2009. URL: https://www.sciencedirect.com/science/article/pii/S0022404909000656, https://doi.org/https://doi.org/10.1016/j.jpaa.2009.03.005 doi:https://doi.org/10.1016/j.jpaa.2009.03.005
-
[3]
Seshadri constants of abelian varieties
Olivier Debarre. Seshadri constants of abelian varieties. In The F ano C onference , pages 379--394. Univ. Torino, Turin, 2004
work page 2004
-
[4]
J.-M. Drezet and M. S. Narasimhan. Groupe de P icard des vari\' e t\' e s de modules de fibr\' e s semi-stables sur les courbes alg\' e briques. Invent. Math. , 97(1):53--94, 1989. https://doi.org/10.1007/BF01850655 doi:10.1007/BF01850655
-
[5]
3264 and all that---a second course in algebraic geometry
David Eisenbud and Joe Harris. 3264 and all that---a second course in algebraic geometry . Cambridge University Press, Cambridge, 2016. https://doi.org/10.1017/CBO9781139062046 doi:10.1017/CBO9781139062046
-
[6]
Local structure of the moduli space of vector bundles over curves
Yves Laszlo. Local structure of the moduli space of vector bundles over curves. Comment. Math. Helv. , 71(3):373--401, 1996. https://doi.org/10.1007/BF02566426 doi:10.1007/BF02566426
-
[7]
D. Mumford, C.P. Ramanujam, and I.U.I. Manin. Abelian Varieties . Studies in mathematics. Hindustan Book Agency, 2008. URL: https://books.google.co.in/books?id=J81jPwAACAAJ
work page 2008
-
[8]
P. E. Newstead. Introduction to moduli problems and orbit spaces , volume 51 of Tata Institute of Fundamental Research Lectures on Mathematics and Physics . Tata Institute of Fundamental Research, Bombay; Narosa Publishing House, New Delhi, 1978
1978
Show all 10 references
-
[9]
M. S. Narasimhan and S. Ramanan. Moduli of vector bundles on a compact R iemann surface. Ann. of Math. (2) , 89:14--51, 1969. https://doi.org/10.2307/1970807 doi:10.2307/1970807
1969 doi
-
[10]
Bauer, Seshadri constants and periods of polarized abelian varieties,
T. Bauer, Seshadri constants and periods of polarized abelian varieties,
Reviewed August 12, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.