Transseries completion of the Yang–Lee string equation: full Stokes data, the oscillatory negative-x region, and the other low-dimensional string equations

The Yang–Lee edge singularity coupled to two-dimensional gravity is governed by the Brézin–Marinari–Parisi (BMP) string equation, the second member of the Painlevé-I hierarchy. Codesido (2017) showed that for positive $x$ the Borel–Padé resummation of the divergent asymptotic series for the pole-free solution reproduces the exact numerical solution with no nonperturbative corrections, and posed three open directions: (i) test the other low-dimensional (multicritical/minimal-model) string equations to decide whether their Borel sums alone reproduce the pole-free solutions or genuine transseries sectors are required; (ii) apply the transseries/Borel–Écalle machinery in cases with instanton actions of positive real part; (iii) reconstruct the oscillatory $x<0$ region of the Yang–Lee solution, where complex exponentially small contributions interfere, by transseries resummation and match it to the exact solution. Later work computed the resonant instanton-action structure of the Yang–Lee equation and built local transseries reconstructions including negative-tension instanton sectors, and completed the analogous program for the first member, Painlevé I; but the complete nonlinear Stokes data and the global resurgent reconstruction of the Yang–Lee equation — in particular of its oscillatory region — and the corresponding per-model determinations for the remaining string equations remain open.

Background

Minimal models coupled to two-dimensional gravity (also called non-critical or minimal strings) are the simplest exactly solvable string theories: they arise as double-scaling limits of hermitian one-matrix models, in which the size $N$ of the matrix is taken to infinity while the matrix potential is tuned to its $k$-th multicritical point. The $k$-th multicritical point corresponds to the $(2,2k-1)$ minimal conformal field theory coupled to gravity, so $k=2$ is pure gravity and $k=3$ is the Lee–Yang edge singularity (the $(2,5)$ minimal model; the Lee–Yang edge singularity is the critical endpoint of the distribution of zeros of the partition function in the complex temperature or fugacity plane). In the double-scaling limit the genus expansion of the free energy becomes an asymptotic series in the string coupling $g_s \sim 1/N$, and the 'specific heat' $u(z)$ (the double-scaled two-point function of the lowest-dimension operator, with $F''(z) = -u(z)/2$ for the free energy) satisfies a nonlinear ordinary differential equation — the string equation. For $k=2$ this is Painlevé I, $u^2 - u''/6 = z$; for $k=3$ it is the Brézin–Marinari–Parisi (BMP) or Yang–Lee equation. In the conventions of the source paper it reads $f^3 + \frac{1}{10}f^{(4)} + f f'' + \frac{1}{2}(f')^2 - x = 0,$ with the physical boundary condition $f(x) \to \mathrm{sign}(x)\,|x|^{1/3}$ as $|x| \to \infty$ and the requirement that the solution be free of poles on the real axis (poles would be zeros of the partition function). The same equation appears in the mathematics literature as $P_I^2$, the second member of the Painlevé-I hierarchy; its real pole-free solution is also known as the Gurevich–Pitaevskii solution of the KdV equation, and it is oscillatory in the negative region, where the oscillations are governed by elliptic (Boutroux-type) asymptotics.

Perturbatively, the physical solution admits a formal series $f(x) \simeq x^{1/3}\sum_k \mu_k x^{-7k/3}$ whose coefficients grow factorially, $|\mu_k| \sim (2k)!$, so the series is divergent and must be resummed. The Borel transform divides the coefficients by $(\beta k)!$ (here $\beta = 2$), producing a series with finite radius of convergence which is analytically continued — in practice by a Padé approximant, a rational function matching the series, giving 'Borel–Padé' resummation — and then integrated along a ray; the result is well defined when the Borel plane has no singularities on the integration path. When exponentially small 'instanton' sectors $\sim e^{-A/x}$ are needed on top of the perturbative series, one obtains a transseries: a formal expansion in both the coupling and the instanton sectors, each with its own asymptotic expansion. Crossing Stokes lines in the coupling's phase plane, the transseries parameters jump by Stokes automorphisms; the associated Stokes constants and connection formulae constitute the Stokes data, which is what makes a transseries a globally defined solution. When two instanton actions sum to a third (here the actions come in symmetric $\pm$ pairs), the transseries is called resonant, which complicates the alien (Écalle) calculus; for the Yang–Lee equation the transseries turns out to be a resonant four-parameter object with instanton actions $A_{1,2} = (6/7)\sqrt{5 \pm i\sqrt{5}}$ (which have positive real part) and their negative-tension (anti-instanton) partners, the latter corresponding to anti-eigenvalues or negative-tension D-branes in the matrix-model description.

Codesido (arXiv:1712.02752) verified numerically that, for $x>0$, the Borel–Padé resummation of the perturbative series alone matches the high-accuracy numerical solution of the BMP equation to all computed digits — no nonperturbative corrections are needed there, because the Borel singularities lie off the positive integration path. His conclusions state the open directions addressed below. Subsequent literature has substantially advanced the surrounding program: Aniceto–Schiappa–Vonk developed the resurgent-transseries analysis of Painlevé I as the string equation of pure 2D gravity and showed that sectors beyond the standard D-brane instantons are required; Gregori–Schiappa analyzed the multicritical and minimal-string string equations nonperturbatively and found their transseries solutions to be resonant; Baldino–Schiappa–Schwick–Vega computed the complete resurgent Stokes data for Painlevé I (and II), and van Spaendonck–Vonk derived complete two-parameter connection formulae for the Painlevé I transseries via exact WKB analysis — so for the first member of the hierarchy the global program is complete. On the mathematical side, Grava–Kapaev–Klein proved existence of the tritronquée (asymptotically pole-free in large sectors) solutions of $P_I^2$ and Dai–Long obtained full asymptotics of its real pole-free solution, but by isomonodromy and elliptic methods rather than by transseries resummation. Most recently, Kager–Rodrigues–Schiappa–Schwick–Tamarin constructed exact transseries solutions for the Yang–Lee equation (among four canonical examples) including the negative-tension sectors, and verified them locally against numerical solutions — while explicitly deferring the global reconstruction, which requires the complete nonlinear Stokes data, to unpublished companion work.

Problem Statement

Complete the nonperturbative, resurgent-transseries solution theory of the low-dimensional string equations, taking the Yang–Lee (Brézin–Marinari–Parisi) string equation as the prototype, along the directions left open by the source paper — its second direction, concerning models whose instanton actions have positive real part, being realized inside the Yang–Lee equation itself and therefore addressed within item 1:

1. (Resurgent reconstruction of the Yang–Lee equation, including the oscillatory negative-$x$ region.) For the BMP equation $f^3 + \frac{1}{10}f^{(4)} + f f'' + \frac{1}{2}(f')^2 - x = 0$ (up to the harmless normalization conventions used across the literature), reconstruct the physical pole-free solution — which oscillates for $x<0$ due to Stokes phenomena and the interference of complex exponentially small contributions — by transseries/Borel–Écalle resummation, extract the complete nonlinear Stokes data (Stokes constants, Stokes automorphisms, and connection formulae) of the resonant transseries, and match the reconstruction against the exact numerical solution. The relevant transseries is the resonant four-parameter transseries of the Yang–Lee equation, whose instanton actions $A_{1,2} = (6/7)\sqrt{5 \pm i\sqrt{5}}$ have positive real part and whose negative-tension (anti-instanton) partners are known to be required for sharp matches; this complete nonlinear Stokes data is not yet available. The reconstruction should reproduce the exact solution in the oscillatory negative-$x$ region — extending the agreement already established for $x>0$ — to the precision attainable by direct numerical integration.

2. (Determination for the other low-dimensional string equations.) For the other string equations of the multicritical hierarchy (equivalently the $(2,2k-1)$ minimal models coupled to 2D gravity, for $k \geq 4$; the higher members built from the Gel'fand–Dikii KdV potentials — the differential polynomials $R_\ell[u]$ generated by the large-$\lambda$ expansion of the resolvent of the KdV Schrödinger operator, which determine the string equation of each multicritical point — as surveyed in Di Francesco–Ginsparg–Zinn-Justin), determine model by model, separately on the $x>0$ branch and on the oscillatory $x<0$ branch, whether the Borel sum of the perturbative series alone reproduces the pole-free (tritronquée-type) solution — as it does for the Yang–Lee equation at $x>0$ — or whether genuine transseries corrections are required. A general criterion settling the whole family counts as a resolution of this item; settling a single order does not. Both items are mandatory: the Yang–Lee reconstruction, its complete nonlinear Stokes data, and its numerical match (item 1), together with the family-wide determination (item 2), are joint requirements, and neither item alone resolves the task.

The common objective of both parts is a resurgent-transseries nonperturbative definition of these exactly solvable string theories, in which the divergent genus expansion, its instanton sectors, and the Stokes phenomena tying them together reconstruct the exact solutions of the string equations.

References

1. S. Codesido, "On the resummation of the Lee-Yang edge singularity coupled to gravity", arXiv:1712.02752 [hep-th] (2017), https://arxiv.org/abs/1712.02752
2. E. Brézin, E. Marinari, G. Parisi, "A non-perturbative ambiguity free solution of a string model", Phys. Lett. B 242 (1990) 35–38, DOI: 10.1016/0370-2693(90)91590-8, https://doi.org/10.1016/0370-2693(90)91590-8
3. P. Gregori, R. Schiappa, "From Minimal Strings towards Jackiw–Teitelboim Gravity: On their Resurgence, Resonance, and Black Holes", Class. Quantum Grav. 41 (2024) 115001, DOI: 10.1088/1361-6382/ad32df, arXiv:2108.11409 [hep-th], https://arxiv.org/abs/2108.11409
4. J. Kager, J. Rodrigues, R. Schiappa, M. Schwick, N. Tamarin, "Exact Solutions to Matrix Models and String Theories: The Local Construction", arXiv:2602.15101 [hep-th] (2026), https://arxiv.org/abs/2602.15101
5. I. Aniceto, G. Başar, R. Schiappa, "A Primer on Resurgent Transseries and Their Asymptotics", Phys. Rep. 809 (2019) 1–135, DOI: 10.1016/j.physrep.2019.02.003, arXiv:1802.10441 [hep-th], https://arxiv.org/abs/1802.10441
6. I. Aniceto, R. Schiappa, M. Vonk, "The Resurgence of Instantons in String Theory", Commun. Number Theor. Phys. 6 (2012) 339–496, DOI: 10.4310/cntp.2012.v6.n2.a3, arXiv:1106.5922 [hep-th], https://arxiv.org/abs/1106.5922
7. S. Baldino, R. Schiappa, M. Schwick, R. Vega, "Resurgent Stokes Data for Painlevé Equations and Two-Dimensional Quantum (Super) Gravity", Commun. Number Theor. Phys. 17 (2023) 385–552, DOI: 10.4310/cntp.2023.v17.n2.a5, arXiv:2203.13726 [hep-th], https://arxiv.org/abs/2203.13726
8. A. van Spaendonck, M. Vonk, "Painlevé I and exact WKB: Stokes phenomenon for two-parameter transseries", J. Phys. A 55 (2022) 454003, DOI: 10.1088/1751-8121/ac9e29, arXiv:2204.09062 [hep-th], https://arxiv.org/abs/2204.09062
9. T. Grava, A. Kapaev, C. Klein, "On the Tritronquée Solutions of $P_I^2$", Constr. Approx. 41 (2015) 425–466, DOI: 10.1007/s00365-015-9285-3, arXiv:1306.6161 [math-ph], https://arxiv.org/abs/1306.6161
10. D. Dai, W. Long, "Asymptotics and Total Integrals of the $P_I^2$ Tritronquée Solution and Its Hamiltonian", SIAM J. Math. Anal. 56 (2024) 5350–5371, DOI: 10.1137/23m1592304, https://doi.org/10.1137/23m1592304
11. P. Di Francesco, P. Ginsparg, J. Zinn-Justin, "2D Gravity and Random Matrices", Phys. Rep. 254 (1995) 1–133, DOI: 10.1016/0370-1573(94)00084-G, arXiv:hep-th/9306153, https://arxiv.org/abs/hep-th/9306153
12. R. Couso-Santamaría, M. Mariño, R. Schiappa, "Resurgence Matches Quantization", J. Phys. A 50 (2017) 145402, DOI: 10.1088/1751-8121/aa5e01, arXiv:1610.06782 [hep-th], https://arxiv.org/abs/1610.06782