REVIEW 2 major objections 4 minor 12 references
A Vafa-Intriligator formula for semi-positive quotients of linear spaces
T0 review · 2 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash
Pith's one-line read For semi-positive quotients $V/\!/G$, the genus-zero quasimap generating series is a rational function given by a finite sum over the fiber of a toric period map.
desk verdict Solid abelianization-to-toric reduction, but the Vafa-Intriligator theorem rests on a shaky comparison with Szenes-Vergne's Morrison-Plesser version; deserves refereeing with a mandate for major revision. 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 carrying mechanism is abelianisation: an invariant of $V/\!/G$ with insertion $P$ equals $\frac{1}{|W|}$ times the sum over lifts of the degree to a maximal torus $T$ of the corresponding invariant of $V/\!/T$ multiplied by a root-correction factor $\prod_{\alpha\in\Delta} D(\tilde\delta,\alpha)$. This reduces non-abelian targets to toric targets, where the quasimap moduli are toric quotients of linear spaces. There the paper uses the period map $p(u)=\sum_{\rho\in A}\frac{\log(\rho(u))}{2\pi i}[\rho]$ and its Jacobian determinant $D_A$ to apply the toric Vafa–Intriligator theorem: the generating series converges when the ray of the stability parameter meets the convergence domain, and on that domain it coincides with a finite sum over the fiber of $p$, twisted by the involution $\sigma$.
What would settle it
Take a rank-two semi-positive toric example whose momentum chamber splits into two simplicial cones, compute the coefficient growth of the abelian generating series in a $c$-positive basis to bound its convergence domain $D$, and check whether the ray spanned by the stability parameter intersects $D$; Lemma 5.4 predicts it does, and a miss would invalidate Theorem 5.3 and the equality in Theorem 5.4. Alternatively, evaluate the claimed identity at a point $q$ of the dual group where the defining sum has a denominator $D_A$ vanishing on the fiber; Remark 12 asserts the rational extension still exists there, so a computed pole would disprove the formula.
Extended reading notes
Core claim
The central claim is Theorem 5.4: for a representation $V$ of a reductive group $G$ with a stability parameter $\xi$ for which the action on the semistable locus is free and $V/\!/T$ is proper and semi-positive, the generating series $\langle P\rangle_y^G(q)$ of $\mathbb{C}^*$-equivariant genus-zero quasimap invariants extends to a rational function and equals $\langle P\rangle_y^B(\sigma(q))$ on its domain of convergence. Here $\langle P\rangle_y^B$ is the finite sum over $w\in p^{-1}(q)$ of explicit terms built from the weights of $T$ on $V$ and the roots of $G$, with the Jacobian determinant of the period map $p$ in the denominator, and $\sigma$ is an involution of the dual torus. If this is right, the paper's earlier abelianisation and residue results imply that the residue conjecture from [Kim+20] holds for all targets of the form $V/\!/G$, and that genus-zero quasimap counts for semi-positive quotients are finite and explicitly computable.
Load-bearing premise
The whole argument assumes that the ray spanned by the stability parameter inside the character lattice meets the convergence domain of the abelian generating series; that lemma is asserted with a compressed proof, and the paper also relies on a reduction step that removes an auxiliary correction term from the toric formula without a fully detailed justification.
Editorial extensions
If this is right
- Genus-zero quasimap invariants of semi-positive quotients $V/\!/G$ are finite and explicitly computable from the weights and roots of $V$, without summing infinitely many residue contributions.
- The residue conjecture posed in [Kim+20] holds for every target of the form $V/\!/G$ covered by the hypotheses.
- The non-equivariant limit $z=0$ gives a Jeffrey–Kirwan residue formula for ordinary quasimap invariants of $V/\!/G$.
- In the Grassmannian case, with $V=\mathrm{Mat}_{r\times n}(\mathbb{C})$ and $G=\mathrm{GL}_r(\mathbb{C})$, the formula specialises to the known Vafa–Intriligator formula for Quot schemes.
- The abelian series converges on a nonempty open region of the dual torus, and the non-abelian series inherits convergence on an analytic open subset of the dual group under semi-positivity.
Reading between the lines
- Editorial extension: the right-hand side of Theorem 5.4 is independent of the stability parameter except through the domain where the left series converges, so wall-crossing between stability chambers reduces to choosing which region of the dual torus the series represents; this could yield a direct wall-crossing proof for these invariants.
- Editorial extension: the finite-sum structure suggests a higher-genus analogue for these quotients, but the paper proves only genus zero and does not claim such an extension.
- Editorial extension: applying the formula to quotients beyond Grassmannians, such as quiver flag varieties or partial flag varieties, should give explicit finite-sum intersection numbers that can be checked against known quantum-cohomology results.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This paper studies genus-zero C*-equivariant quasimap invariants of GIT quotients V//G of a linear representation V by a reductive group G. The main results are: an abelianisation formula (Theorem 3.2) expressing invariants of V//G as a Weyl-group average of invariants of V//T with an explicit root correction; a Jeffrey-Kirwan residue formula for fixed degree (Theorem 4.1, Corollary 4.1); and, under a semi-positivity assumption, a Vafa-Intriligator formula (Theorem 5.4) expressing the generating series as a finite sum over preimages of a map p on the dual torus. The proof strategy is to abelianise, apply the toric Vafa-Intriligator theorem of Szenes and Vergne, and then descend back to the dual group. The paper also recovers the Marian-Oprea formula for Grassmannians as a special case.
Significance. The abelianisation and residue formulae are detailed and give a substantial new computational tool: they reduce nonabelian quasimap invariants to explicit toric residues, with no fitted constants, and the specialization to Grassmannians reproduces a known formula. If Theorem 5.4 is fully established, it would provide a general Vafa-Intriligator formula for semi-positive quotients of linear spaces and settle the Kim-Oh-Ueda-Yoshida residue conjecture for this class of targets. However, the proof of Theorem 5.4 relies on two points that are not fully supported: the convergence lemma 5.4 and the no-Morrison-Plesser version of the Szenes-Vergne theorem in Appendix B. These points are load-bearing, so the main theorem is currently conditional.
major comments (2)
- [Appendix B, Theorem B.1 and Remark 16] Theorem 5.4 is reduced in its proof to Theorem B.1, but Theorem B.1 is not actually established from the cited Szenes-Vergne result. Remark 16 concedes that in [SV04] every coefficient carries an additional Morrison-Plesser class, and the assertion that this class "doesn't appear" for V//T is not justified: Example 2.1 shows that the virtual class of the toric quasimap space involves the nontrivial factor prod_{rho: <delta,rho><0} rho^{-1-<delta,rho>}. Furthermore, the descent step at the end of Appendix B invokes [Vis04, Theorem 4.33] to conclude that xPy_B is rational on qT from rationality of xPy_B o u, but it does not check invariance of the function under the kernel of u, nor does it explain why descent for rational functions follows from descent for quasi-coherent sheaves. Because the proof of Theorem 5.4 is a direct appeal to Theorem B.1, this gap is load-bearing.
- [Lemma 5.4] The proof of Lemma 5.4 is compressed and relies on the assertion that D is convex "by the general theory of power series". While convergence domains of power series are logarithmically convex, the convexity of the set D in the additive coordinates on chi(T)_R is not immediate from this phrase, because D is defined as a preimage under the exponential map and the relevant power series coordinates depend on a c-positive basis (Theorem 5.2). Since Lemma 5.4 is the step that produces a point in the intersection of the ray spanned by xi with the convergence domain, Theorem 5.3 and the equality in Theorem 5.4 on the stated convergence domain depend on it. A complete proof or a precise citation should be supplied, and the induction over simplicial cones should justify the claim that the auxiliary open set U intersects some top-dimensional cone in an interior point.
minor comments (4)
- [Remark 10] There is a likely typo: the text says "if delta is not a xi-effective class, then the moduli space Q(V//T,delta) is nonempty and delta does not contribute to the power series"; the intended statement is presumably that the moduli space is empty.
- [Definition 5.2 and Lemma 5.3] The sign conventions for the coordinate maps should be reconciled: Definition 5.2 sets q_i([psi]) = e^{2 pi i <lambda_i,psi>}, while Lemma 5.3 uses the map x -> (e^{-2 pi x_1}, ..., e^{-2 pi x_r}) with no reference to that formula.
- [Theorem 5.2] The statement of Theorem 5.2 is grammatically tangled ("For every integral basis ... the dual basis lambda is called a c-positive basis. Let lambda be a c-positive basis. If lambda in chi(T)^* is a degree...") and should be rewritten so that the quantifiers over bases and degrees are explicit.
- [Equation (1) and Definition 5.4] The notation D is used both for the rational function D(delta,w) in (1) and for the Jacobian determinant D_A in (24); this is potentially confusing and should be disambiguated.
Circularity Check
No significant circularity: the nonabelian Vafa-Intriligator formula is derived from external toric results (Webb and Szenes-Vergne) via abelianization, with no fitted or self-defined quantity.
full rationale
The derivation chain is a reduction rather than a circle. Theorem 3.2 is imported from Webb's abelianization theorem [Web23], and Appendix A explicitly says 'No argument here is really original: we just check that Webb's proof goes through without obstructions'; so the abelianization input is an external result, not the paper's own conclusion. The toric Vafa-Intriligator equality xPyT(q) = xPyB(q) is cited to Szenes and Vergne [SV04, Theorem 4.1], and Theorem 5.4 combines that external toric formula with Theorem 5.1 (a direct consequence of the external abelianization) and Theorem 5.3. The right-hand side (25) is an explicit finite sum over the finite fibers of the map p, not a restatement of the generating series being computed; no parameter is fitted to the left-hand side. The only self-citation is [Ont23], used for the C*-equivariant Jeffrey-Kirwan localization theorem in Section 4; the paper itself describes the same residue formula as obtainable by an alternative route 'for example in the form of [Ont23]' (Section 1.1.2), and the main convergence/Vafa-Intriligator theorem does not depend on that intermediate theorem. Appendix B's adaptation of Szenes-Vergne is admittedly not verbatim: Remark 16 concedes that in [SV04] each coefficient carries an additional Morrison-Plesser class, and the fpqc descent step from u-rationality to rationality on qT is asserted rather than proved. That is a correctness gap in an external input, not a circularity, because Theorem B.1 is not defined in terms of the paper's own target invariants and no assumption is equivalent to the conclusion.
Assumptions & free parameters
assumptions (9)
- standard math Quasimap moduli spaces Q(V//G, delta) have a perfect obstruction theory and virtual fundamental class, as constructed in CKM14, Theorem 7.2.2.
- standard math The C*-fixed locus of Q(V//G, delta) decomposes as the disjoint union of F_{delta1,delta2}, as in CK13, Section 4, recalled as Proposition 2.1.
- standard math Martin-Maddock integration formula (Theorem 3.1): if g*alpha = j*beta then the integral over V//G equals 1/|W| times the integral over V//T of beta times the product of root line bundle classes.
- standard math Webb's abelian-nonabelian correspondence for I-functions (Web23, Theorem 1.1.1) and its fixed-locus version in Lemma A.1.
- standard math Szenes-Vergne toric Vafa-Intriligator formula (SV04, Theorem 4.1), in the form recalled as Theorem B.1.
- standard math Shoemaker's identification NE(xi) = c^vee for the effective cone of stable quasimap degrees (Sho22, Theorem 1.7), used in Lemma 5.1.
- domain assumption The actions of G and T are free on the semistable loci and V//T, hence V//G, is proper.
- domain assumption Semi-positivity: the anticanonical character kappa = sum of weights lies in the closure of the chamber c, and is in the interior in the positive case.
- standard math The convergence domain D of a multivariable power series is logarithmically convex, used in Lemma 5.4.
Cite this review
Pith. "Pith review of A Vafa-Intriligator formula for semi-positive quotients of linear spaces." pith.science (2026). https://pith.science/paper/H7EF2M3V
@misc{pith2026250517845,
author = {Pith},
title = {Pith review of: A Vafa-Intriligator formula for semi-positive quotients of linear spaces},
year = {2026},
howpublished = {\url{https://pith.science/paper/H7EF2M3V}},
note = {Machine review of arXiv:2505.17845}
}
abstract
We consider genus zero quasimap invariants of smooth projective targets of the form $V/\!/G$, where $V$ is a representation of a reductive group $G$. In particular we consider integrals of cohomology classes arising as characteristic classes of the universal quasimap. In this setting, we provide a way to express the invariants of $V/\!/G$ in terms of invariants of $V/\!/T$, where $T$ is a maximal subtorus of $G$. Using this, we obtain residue formulae for such invariants as conjectured by Kim, Oh, Yoshida and Ueda. Finally, under some positivity assumptions on $V/\!/G$, we prove a Vafa-Intriligator formula for the generating series of such invariants, expressing them as finite sums of explicit contributions.
Figures
Reference graph
Works this paper leans on
-
[1]
Towards a Schubert calculus for maps from a Riemann surface to a Grassmannian
[Ber94] A. Bertram. “Towards a Schubert calculus for maps from a Riemann surface to a Grassmannian”. In:International Journal of Mathemat- ics 5.06 (1994). arXiv:alg-geom/9403007, pp. 811–825. [BF97] K. Behrend and B. Fantechi. “The intrinsic normal cone”. In:Inven- tiones Mathematicae128 (1997). arXiv:alg-geom/9601010, p
work page Pith review arXiv 1994
-
[5]
arXiv:math/9903178. 1999, p
arXiv 1999
-
[45]
Toric Residues and Mirror Symmetry
[BM02] V. V. Batyrev and E. N. Materov. “Toric residues and mir- ror symmetry”. In: Moscow Mathematical Journal 2.3 (2002). arXiv:math/0203216, pp. 435–475. [Bri96] M. Brion. “The push-forward and Todd class of flag bundles”. In: Banach Center Publications36 (1996), pp. 45–50. [BV99] M. Brion and M. Vergne. “Arrangement of hyperplanes. I. Rational functio...
work page Pith review arXiv 2002
-
[291]
Residue Mirror Symmetry for Grassmannians
[Kim+20] B. Kim et al. “Residue Mirror Symmetry for Grassmannians”. In: Springer Proceedings in Mathematics & Statistics332 (2020). arXiv:1607.08317. [Mad14] Z. Maddock. “A ratio of integration between quotients in geo- metric invariant theory”. In:Transformation Groups 19.1 (2014). arXiv:1210.4253, pp. 131–158. [Mar00] S. Martin. “Symplectic quotients by...
arXiv 2020
-
[453]
NotesonGrothendiecktopologies,fiberedcategoriesand descent theory
[Vis04] A.Vistoli.“NotesonGrothendiecktopologies,fiberedcategoriesand descent theory”. In:arXiv preprint(2004). arXiv:math/0412512. [Web19] R. Webb. “Quasimaps and some examples of stacks for everybody”. arXiv:1910.07262
arXiv 2004
-
[715]
Moduli stacks of stable toric quasimaps
[CK10] I. Ciocan-Fontanine and B. Kim. “Moduli stacks of stable toric quasimaps”. In: Advances in Mathematics 225.6 (2010). arXiv:0908.4446, pp. 3022–3051. [CK13] I. Ciocan-Fontanine and B. Kim. “Wall-crossing in genus zero quasimap theory and mirror maps”. In:Algebraic Geometry1 (Apr. 2013). arXiv:1304.7056. [CKM14] I. Ciocan-Fontanine, B. Kim, and D. Ma...
work page Pith review arXiv 2010
-
[2000]
Counts of maps to Grassmannians and intersections on the moduli space of bundles
[MO07a] A. Marian and D. Oprea. “Counts of maps to Grassmannians and intersections on the moduli space of bundles”. In:Journal of Differ- ential Geometry76.1 (2007). arXiv:math/0602335, pp. 155–175. [MO07b] A. Marian and D. Oprea. “Virtual intersections on the Quot scheme and Vafa-Intriligator formulas”. In: Duke Math. J. 136.1 (2007). arXiv:math/0505685,...
work page Pith review arXiv 2007
-
[2013]
A mirror theorem for toric complete intersections
[Giv98] A. Givental. “A mirror theorem for toric complete intersections”. In: Topological field theory, primitive forms and related topics . arXiv:alg-geom/9701016. Springer, 1998, pp. 141–175. [GP97] T. B. Graber and R. Pandharipande. “Localization of virtual classes”. In: Inventiones mathematicae 135 (1997). arXiv:alg- geom/9708001, pp. 487–518. [JK95] ...
work page Pith review arXiv 1997
Show all 12 references
-
[2019]
The abelian–nonabelian correspondence for I-functions
[Web23] R. Webb. “The abelian–nonabelian correspondence for I-functions”. In: International Mathematics Research Notices 2023.3 (2023). arXiv:1804.07786, pp. 2592–2648. [Zho22] Y. Zhou. “Quasimap wall-crossing for GIT quotients”. In: Inven- tiones mathematicae227.2 (2022). arX...
2023 arXiv
-
[2022]
On quantum cohomology rings of Fano manifolds and a formula of Vafa and Intriligator
[ST97] B. Siebert and G. Tian. “On quantum cohomology rings of Fano manifolds and a formula of Vafa and Intriligator”. In:Asian Journal of Mathematics 1.4 (1997). arXiv:alg-geom/9403010, pp. 679–695. 43 [Sta25] T. Stacks project authors. The Stacks project. https://stacks. mat...
1997 arXiv
-
[2023]
Log Calabi-Yau surfaces and Jeffrey- Kirwan residues
[OS24] R. Ontani and J. Stoppa. “Log Calabi-Yau surfaces and Jeffrey- Kirwan residues”. In: Mathematical Proceedings of the Cambridge Philosophical Society176.3 (2024). arXiv:2109.13048, pp. 547–592. [Sho22] M. Shoemaker. “A Kleiman criterion for GIT stack quotients”. arXiv:2211.09218
2024 arXiv
-
[2025]
Toric reduction and a conjecture of Batyrev and Materov
[SV04] A. Szenes and M. Vergne. “Toric reduction and a conjecture of Batyrev and Materov”. In:Inventiones Mathematicae158.3 (2004). arXiv:math/0306311, p
2004 arXiv
Reviewed August 7, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.