On uniform large genus asymptotics of Witten's intersection numbers

arXiv:2603.15233 · math-ph, math.AG, math.CO, math.MP, nlin.SI · Submitted 2026-03-16 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Mathematical Physics. The day in mathematical physics: rigorous models of physical systems.

Kai: We're looking at the paper: "On uniform large genus asymptotics of Witten's intersection numbers".

Mira: Here's the gist of the paper. Following ideas from

14: , we give a uniform large genus asymptotics for primitive psi-class intersection numbers on the moduli space of stable algebraic curves, and extend this result including insertions of zeros in a certain uniform way.

Kai: Application to a particular formal solution of the Painlevé I equation is given.

Mira: We also use a method from

14: to give a new proof of the polynomiality conjecture on large genus asymptotic expansions of psi-class intersection numbers.

Kai: Witten's Intersection Numbers and Conjectures The paper deals with Witten’s intersection numbers, which are defined as partition functions related to the moduli space of stable algebraic curves, and explores their large genus asymptotics.

Title and authors: Kai: So to recap this paper, "On uniform large genus asymptotics of Witten's intersection numbers," Guo, Yang, and Zagier provide a uniform description for the large genus behavior of primitive psi-class intersection numbers on the moduli space of stable algebraic curves, extending results from prior work and proving a new polynomiality conjecture.

Mira: They are focusing on these numbers called Witten's intersection numbers. These play a big role in studying Weil–Petersson volumes and Masur–Veech volumes for moduli spaces of curves and quadratic differentials.

Lev: The key idea they introduce is this uniform large genus asymptotics, which means they’re finding the limit behavior as the genus gets very high, not just some local approximation.

Kai: And on top of that, they extend this result to include insertions of zeros in a certain uniform way, which adds another layer of structure to what's being measured.

Mira: The real substance here is using a method from earlier work to prove the polynomiality conjecture on those large genus asymptotic expansions.

Lev: That polynomiality part is significant because it suggests the coefficients aren't just random noise or complicated functions; they follow a predictable algebraic rule based on these multiplicities.

Kai: And they also have an application to a particular formal solution of the Painlevé one equation, which gives us some concrete results about how these intersection numbers behave in that context.

Mira: That connection to the Painlevé one equation is what makes this paper really interesting for us theorists, because it links the geometry of curves to solutions of differential equations.

Lev: For practical implementation, if we could nail down these asymptotic bounds with a uniform result, it gives us a better target for what we might expect from complex systems in our hardware.

The paper's summary: Kai: To summarize the core of "On uniform large genus asymptotics of Witten's intersection numbers," they are establishing a uniform large genus asymptotics for primitive psi-class intersection numbers on the moduli space of stable algebraic curves, and they extend this result including insertions of zeros in a certain uniform way.

Mira: They then apply this to a particular formal solution of the Painlevé one equation to give an application. This connects the intersection numbers to that specific differential equation solution.

Lev: The paper is using a method from earlier work, and they use it again to provide a new proof of the polynomiality conjecture on large genus asymptotic expansions of psi-class intersection numbers.

Kai: So, the main thrust is combining this uniform asymptotic result with that new proof of polynomiality, which tells us something fundamental about the structure of these coefficients.

Mira: Essentially, they are showing that these complicated intersection numbers have a very well-behaved asymptotic structure that is governed by polynomials in their multiplicities.

Lev: This suggests that even in the large genus limit, there's still an underlying algebraic simplicity to what we are measuring.

Kai: It’s about taking a result about asymptotic limits and showing it holds uniformly, which is a strong requirement when you’re dealing with complex moduli spaces.

The paper's improvements: Mira: One of the important parts they point out is the improvement they make to previous results by extending the original uniform large genus asymptotics to include insertions of zeros in a certain uniform way.

Kai: That extension is key because it allows them to handle more general cases, not just the basic intersection numbers without any special insertions.

Lev: From an engineering view, handling those zero insertions uniformly means we aren't restricted to simple, isolated points on the moduli space; we can deal with structured modifications.

Mira: And then they use a method from earlier work to give a new proof of the polynomiality conjecture, which is a significant step forward in understanding those expansions.

Kai: That proof shows that the coefficients of the large genus expansion are indeed polynomials in these multiplicities, which is what we hoped for when we looked at this conjecture.

Lev: If that polynomiality holds as they prove it, it solidifies the idea that the structure of these asymptotic coefficients is entirely determined by those fundamental combinatorial quantities.

Mira: The real improvement they bring is combining a uniform asymptotic result with this proof of polynomiality, giving us a much more complete picture.

Conclusion: Kai: So, to wrap up on "On uniform large genus asymptotics of Witten's intersection numbers," they provide a uniform large genus asymptotics for primitive psi-class intersection numbers on the moduli space of stable algebraic curves, and extend this result including insertions of zeros in a certain uniform way.

Mira: They also give an application to a particular formal solution of the Painlevé one equation, and they use that method from earlier work to give a new proof of the polynomiality conjecture on large genus asymptotic expansions of psi-class intersection numbers.

Lev: The implication is that we have this uniform asymptotic result for these intersection numbers, which is backed up by a proof showing their coefficients are polynomials in those multiplicities.

Kai: For someone who only listens to the show, it means these complicated geometric quantities have a predictable large genus limit, and they follow an algebraic rule dictated by the structure of the input parameters.

Mira: It’s about understanding that underlying simplicity even when you're looking at high-genus curves or complex differential equations.

Lev: We get a solid result on how to predict the asymptotic behavior of these coefficients based on their structure, which is something we can use to benchmark expectations in other areas.

Kai: That’s the gist of what this paper is about, showing a very structured large genus limit for Witten's intersection numbers.

Mira: It gives us a clearer picture of the relationship between geometry and those integrable systems they are studying.

Lev: And it sets a new benchmark for how much structure we can expect in these kinds of complex asymptotic expansions when we look at high genus limits.

University of Science and Technology of China · Max Planck Institute for Mathematics

math-ph, math.AG, math.CO, math.MP, nlin.SI

Submitted: 2026-03-16

Updated: 2026-10-31

Comments: We corrected an error in the statement of Theorem 2, and updated its proof. We also corrected a few typos. 23 pages

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 98/100

The gist: Following ideas from [14], we give a uniform large genus asymptotics for primitive psi-class intersection numbers on the moduli space of stable algebraic curves…

Key concepts

Witten's Intersection Numbers
These are partition functions related to the moduli space of stable algebraic curves. They are crucial in studying geometric volumes like Weil–Petersson volumes and Masur–Veech volumes, and they are conjectured to satisfy integrable system constraints like the KdV hierarchy.
Large Genus Asymptotics
This refers to analyzing how these intersection numbers behave as the genus $g$ of the algebraic curve tends toward infinity. The paper establishes uniform bounds and specific asymptotic expansions for these numbers, showing they approach a limit of $1/ ext{π}$ plus lower-order corrections.
Polynomiality Conjecture
This conjecture suggests that the coefficients in the large genus expansion of Witten's intersection numbers are polynomials in certain parameters ($p_0, p_1, ext{etc.}$). The paper proves this conjecture by analyzing recursion relations and showing the functions belong to a specific polynomial ring.
Primitive Normalized Intersection Numbers $C(d)$
These are specific normalized Witten's intersection numbers that are primitive (cannot be simplified further). The paper conjectures they lie between bounds defined by $C(3g-2)$ and $C(23g-3)$, providing a structural constraint on their growth as genus increases.

Terminology

Summary

Following ideas from [14], we give a uniform large genus asymptotics for primitive psi-class intersection numbers on the moduli space of stable algebraic curves, and extend this result including insertions of zeros in a certain uniform way. p. 0 Application to a particular formal solution of the Painlevé I equation is given. p. 0 We also use a method from [14] to give a new proof of the polynomiality conjecture on large genus asymptotic expansions of psi-class intersection numbers. p. 0

Witten's Intersection Numbers and Conjectures

The paper deals with Witten’s intersection numbers, which are defined as partition functions related to the moduli space of stable algebraic curves, and explores their large genus asymptotics. These numbers play important roles in studying Weil–Petersson volumes of Mg,n [19, 22, 25, 26, 27] and Masur–Veech volumes of moduli space of quadratic differentials [1, 6]. p. 1 A key aspect is the conjecture that these numbers satisfy the KdV hierarchy or Virasoro constraints. p. 1 The paper introduces normalized intersection numbers C(d) and conjectures their nesting property, stating that for fixed genus g ≤ 13, primitive normalized Witten’s intersection numbers C(d) with d1,. p. 3.., dn ≥ 2 lie between the values C(3g−2) and C(23g−3). p. 3

Asymptotic Bounds and Theorems

The authors establish several key results concerning the asymptotic behavior of these numbers as the genus g tends to infinity. Theorem 1 states that for n ≥ 1 and d ∈ (Z≥1)n satisfying g(d) ∈ Z, we have C(d) = 1/π + O(g(d)) uniformly as g(d) → ∞. p. 4 This is supported by Conjecture 1, which suggests that the normalized intersection numbers G(d) tend to 1 uniformly if n = O(log g). p. 2 Theorem 2 provides a more refined asymptotic expansion for C(d), showing that C(d) = 1/π + O(g(d)) uniformly as g(d) → ∞. p. 6

Polynomiality in Large Genus Expansions

The paper addresses the polynomiality conjecture on large genus asymptotic expansions of psi-class intersection numbers. Theorem 3 proves this conjecture by showing that the coefficients of the large genus expansion of C(d) are polynomials in the multiplicities p0, p1,. p. 6.. The proof involves analyzing a recursion relation (29) and showing that the functions c˜k(p) belong to Q[p0,. p. 5.., p3k−2].

Applications to Integrable Systems

An application of Theorem 1 is given to the Painlevé I equation, where the asymptotic behavior of the coefficients cg is related to Witten’s intersection numbers. This leads to Theorem A, which determines a constant A = 1/(2π) for certain coefficients. p. 4 The analysis also yields Corollary 1, which shows that C(0k, 2/3g−3+k) → 0 as g → ∞ when k = O(√g). p. 5

Conclusion of Asymptotic Behavior

The analysis concludes by showing that for k/√g → ∞, the term in the asymptotic expansion tends to 0, confirming Corollary 1. This confirms the uniform large genus asymptotics for primitive Witten’s intersection numbers. p. 0 The proof relies on a recursion (85) and subsequent analysis of the resulting functions c˜k(p). p. 0 The final result is that Cbk(d′) = bck(p1(d′), p2(d′),. p. 0..).

References

[1] A. Aggarwal, Large genus asymptotics for intersection numbers and principal strata volumes of quadratic differentials. p. 1 Invent. Math., 226 (2021), 897–1010.

[4] V. Delecroix, E. p. 2 Goujard, P. p. 2 Zograf, A. Zorich, Uniform lower bound for intersection numbers of psi-classes. p. 9 Symmetry Integrability Geom. p. 9 Methods Appl., 16 (2020), 86–98.

[7] L.A. Dickey, Soliton Equations and Hamiltonian Systems, 2nd edition. World Scientific, 2003.

[9] B. Dubrovin, D. p. 2 Yang, D. Zagier, Classical Hurwitz numbers and related combinatorics. p. 3 Moscow Mathematical Journal, 17 (2017), 601–633.

[14] J. Guo, P. Norbury, D. p. 2 Yang, D. Zagier, Combinatorics and large genus asymptotics of the Br´ezin–Gross–Witten numbers. p. 2 arXiv:2412. p. 020388.

[15] J. Guo, D. Yang, On the large genus asymptotics of psi-class intersection numbers. p. 0 Math. Ann., 388 (2024), 61–97.

[17] N. Joshi, A.V. Kitaev, On Boutroux’s tritronqu´ee solutions of the first Painlevé equation. p. 0 Stud. Appl. Math., 107 (2001), 253–291.

[18] A. A. Kapaev, Quasi-linear Stokes phenomenon for the Painlevé first equation. p. 1 J. Phys. A, 37 (2004), 11149–11167.

[20] M. Kontsevich, Intersection theory on the moduli space of curves and the matrix Airy function. p. 0 Comm. Math. Phys., 147 (1992), 1–23.

[28] E. Witten, Two-dimensional gravity and intersection theory on moduli space. p. 0 Surveys Diff. Geom., 1 (1991), 243–310.

[29] J. Zhou, Emergent geometry and mirror symmetry of a point. p. 9 arXiv:1507. p. 001679.

[30] P. Zograf, An explicit formula for Witten’s 2-correlators. p. 7 J. Math. Sci., 240 (2019), 535–538.

[44] C(d1, 3g − 1 − d1) = 1/54g (2g − 1)! p. 7 g d X j−1−d=−1 ηg, d, d ≥ −1. p. 7

[47] C(d) = Xσ∈Sn σ(n)=n (−1)S − σ +1 X J∈(Z+ 1/2) n, five or more words copied word for word>. p. 2

[50] M(e):= max 0, min 1≤i≤n,ei, (50). p. 8

[74] The numbers θX,n have the upper bound θX,n ≤ f(X, n). p. 9

[75] 1/π ≤ f(X, n) ≤ 1 for any n ≥ 1, X ≥ 1.

[77] C(d1, d2) ≤ C(0, 3g − 1).

[84] where u = ∂ 2 x (log τ), uk = ∂ k t0 (u). p. 0

[85] C(d) = C(06, d) + 1/6 X I⊔J=1,…,n (X(03, dI − 1)! (X(03, dJ − 1)! (X(d) + 1)! C(03, dI) C(03, dJ) + 2/3 X I⊔J=1,…,n (X(02, dI − 1)! (X(04, dJ - 1)! (X(d) + 1)! C(02, dI) C(04, dJ) + 1/3 X I⊔J⊔K=1,…,n (X(02, dI − 1)! (X(02, dJ - 1)! (X(02, dK) − 1)! (X(d) + 1)! C(02, dI) C(02, dJ) C(02, dK), five or more words copied word for word>. p. 2

[87] By applying (86) and (87), we obtain (82). p. 4

[94] cg ∼ A · 50g(g − 1)! squared / ((5g − 5)! (5g − 3) (3g − 3)!), g → ∞.

[96] From Theorem 1 and Lemma 4 we know that as Y → ∞, θ(α)Y = 1/π + O(Y−1), α = 0, 1, 2, with an absolute O-constant. p. 4

[98] Given sufficiently small epsilon > 0, when α ≤ epsilonY we know from θ(2)Y = 1/π + O(Y−1) that the maximum in the right-hand side of (98) equals 1/π +O(Y−1) with an absolute O-constant. p. 7

[99] and hence is O(Y−1) with an absolute O-constant. p. 4

[104] C(0k, 2/3g−3+k) ∼ 1/π e − k 2/30g (1 + o(1)) + O(g). p. 5

[105] Y k j=1 1 + 2 + j − k / 3(5g − 5 + 2k - j) ≤ exp k(5 − k)/6(5g−5+k) → 0, g → ∞.

[106] C(d) = 1/π X k Ck(d') X j k, as dn → ∞.

[107] where d ≥ 0 and aw,l,k are certain numbers defined via the following generating function (X(w) − 1)! p. 0 C(w) (X − X(w) - 1)! X! (X − X(w))l = X k a w,l,k / X k.

[109] where d = (d′), where pr(d') denotes the multiplicity of r in d'.

[119] where the left-hand side is understood as a power series in X−1. p. 2

[120] for all d′.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Automated proof verification for polynomiality conjectures: The system can check if coefficients in large genus asymptotic expansions of psi-class intersection numbers satisfy the polynomiality conjecture by comparing them to predicted structures derived from Theorem 3 and equation (114).

  2. Derivation of Painlevé I solution asymptotics: The AI can use the explicit formula for Witten’s intersection numbers in terms of C(d) and Theorem A to derive asymptotic behavior for formal solutions of the Painlevé I equation, specifically calculating the leading term A = 1/(2 pi 2 r 3).

  3. Uniform large genus asymptotic prediction: The system can predict the asymptotic behavior of normalized Witten’s intersection numbers C(d) for any fixed genus g and varying partitions d, using Theorem 1 and Corollary 1 to establish bounds based on the relation C(3g - 2) <= C(d) <= C(23g - 3).

  4. Multiplicity tracking in asymptotic expansions: The AI can analyze the structure of c˜k functions defined in (113) to determine the multiplicities p0, p1, etc., of the arguments, allowing it to verify that these multiplicities are bounded by degree estimates deg c˜k <= 3k - 1.

  5. Identification of universal polynomials: The system can identify the universal polynomials bck(p2, p3,...) using the relation (109) and (120), which are used to represent Cbk(d') for all d', enabling rapid calculation of intersection numbers based on the structure of their arguments.

Sources

Related papers