Nonperturbative Exponential Effects and Finite-N Sectors

Agreement to every algebraic order in $1/N$ does not determine the exact finite-$N$ theory. Contributions proportional to $e^{-aN}$, $e^{-bN^2}$, or an entropy-scale exponential are smaller than every power of $1/N$ at fixed kinematics, yet can encode branes, competing gravitational saddles, topology change, finite-rank identities, or spectral discreteness. Their scaling and interpretation depend on the theory and observable; there is no universal nonperturbative completion supplied by genus counting alone.

Required background. From Genus Counting to a Holographic String Regime fixes the perturbative handle expansion, and Corrections, Nonuniform Limits, and Failure Modes explains how exponentially small terms become relevant in scaled domains.

Helpful background. Euclidean Tunneling Saddles and Boundary Conditions develops saddle and contour dependence, while EFT Truncation Errors and Breakdown Diagnostics distinguishes asymptotic error estimates from exact remainders.

Beyond the perturbative genus series

For a normalized observable with a closed-string organization, write

$$\mathcal A_{\mathrm{pert}}(N) \sim \sum_{g=0}^{\infty}\frac{a_g}{N^{2g}}.$$

The symbol $\sim$ denotes an asymptotic expansion, not necessarily a convergent sum. Even if every coefficient $a_g$ is known, the exact observable may instead have a transseries-like asymptotic structure

$$\mathcal A(N) \sim \sum_{g\ge0}\frac{a_g}{N^{2g}} +\sigma_1e^{-A_1N} N^{\beta_1}\sum_{k\ge0}\frac{b_k}{N^{k/r_1}} +\sigma_2e^{-A_2N^2} N^{\beta_2}\sum_{k\ge0}\frac{c_k}{N^{k/r_2}} +\cdots .$$

Here $A_i$ are positive saddle actions in the displayed regime, $N^{\beta_i}$ are fluctuation prefactors, $r_i$ allow fractional-power expansions, and $\sigma_i$ specify how the corresponding sectors enter. All of these are dynamical data. The first sum is the ordinary genus expansion. Each exponential labels a sector that no additional power of $1/N$ can generate; the factor multiplying it is that sector’s own fluctuation series.

The genus coefficients can display large-order growth that diagnoses candidate exponential scales, but large-order analysis alone need not choose an integration contour, boundary condition, or the values of the sector parameters. An equality would require a specified resummation and nonperturbative definition; the asymptotic symbol does not presume either Aniceto, Başar, and Schiappa 2019, §§ 2–5.

Same power series, different exact answers

The ambiguity can be demonstrated without assuming a divergent series. Define, for $N>1$,

$$F_{\pm}(N) = \frac{1}{1+N^{-2}} \mathbin{\pm}e^{-N}.$$

Both functions have exactly the same all-orders algebraic asymptotic expansion,

$$F_{\pm}(N) \sim 1-\frac1{N^2}+\frac1{N^4}-\frac1{N^6}+\cdots ,$$

because for every integer $K>0$,

$$\lim_{N\to\infty}N^K \bigl[F_+(N)-F_-(N)\bigr]=0.$$

Nevertheless, $F_+-F_-=2e^{-N}$. An observable rescaled by $e^N$, or evaluated in a sector whose leading answer is itself $e^{-N}$, distinguishes the two completions at order one. Thus equality of every $1/N$ coefficient establishes perturbative equivalence, not uniqueness of the exact answer.

In an actual path integral, not every arbitrary exponential addition is allowed. Unitarity, symmetries, factorization, analyticity, boundary conditions, charge quantization, and a specified microscopic definition constrain the sectors. The example exposes the logical gap that those additional inputs must close.

Brane and gravitational action scales

In familiar matrix-like holographic examples, $g_s\sim N^{-1}$. A Euclidean D$p$-brane wrapped on a cycle of volume $V_{p+1}$ has, schematically,

$$I_{D p}\sim \frac{V_{p+1}}{g_s\ell_s^{p+1}}\sim O(N)$$

when the geometric ratios are held fixed, and hence can contribute as $e^{-O(N)}$. A purely gravitational saddle instead has the action scale

$$I_{\mathrm{grav}}\sim \frac{L^{d-1}}{G_{d+1}}\sim O(N^2)$$

in standard adjoint examples. If this saddle exceeds the dominant saddle by $\Delta I_{\mathrm{grav}}=O(N^2)$, its relative weight is $e^{-\Delta I_{\mathrm{grav}}}=e^{-O(N^2)}$. These are scaling guides, not universal identities: wrapped volumes, fluxes, charges, couplings, zero modes, action differences, and the normalization of the observable alter both exponent and prefactor. Polchinski’s identification of D-branes as carriers of Ramond–Ramond charge supplies the string-theory basis for the inverse-$g_s$ tension scale Polchinski 1995, pp. 4724–4726.

Four distinct finite-N mechanisms

It is useful to separate mechanisms that are often grouped under “nonperturbative.”

Brane sectors. Extended objects have tensions nonanalytic in $g_s$. Boundary operators whose quantum numbers grow with $N$, such as determinant or subdeterminant operators, can couple to these sectors at leading order even though no fixed operator length captures their large-$N$ scaling.

Additional spacetime saddles. A Euclidean path integral may admit saddles with different topology or action. Whether they contribute depends on the contour and boundary conditions. A subdominant saddle can be exponentially suppressed in one observable and essential in another.

Finite-rank identities. At finite $N$, traces of sufficiently high powers of an $N\times N$ matrix are related by the Cayley–Hamilton theorem. The large-$N$ Fock-space picture can therefore overcount states when operator length or charge reaches the relevant finite-rank threshold, which is $O(N)$ in the half-BPS examples. Scaling as $J\sim N$ makes this crossover possible but does not by itself prove that the threshold has been crossed. This is exact finite-$N$ structure, not a small loop correction.

Level discreteness. Work inside one fixed symmetry sector. Let $D_{\mathrm{distinct}}(E,W)$ count distinct energy values in a narrow band of width $W$. If $D_{\mathrm{distinct}}(E,W)\sim e^{S_{\mathrm{spec}}(E)}$, then

$$\rho_{\mathrm{distinct}}(E) \sim \frac{e^{S_{\mathrm{spec}}(E)}}{W}, \qquad t_H(E)=2\pi\rho_{\mathrm{distinct}}(E)$$

in $\hbar=1$ units. The spectral entropy $S_{\mathrm{spec}}$ counts distinct levels, so identifying it with the thermodynamic entropy additionally requires the absence of parametrically large degeneracies. In chaotic spectra and in the random-matrix model below, suitably averaged spectral form factors develop late-time ramp and plateau behavior on the corresponding scales; one fixed realization retains noisy fluctuations Altland and Sonner 2021, §§ 2.1–2.6.1. Ordinary correlators can also show late-time noise or Poincaré recurrences, but those phenomena are observable-dependent and need not occur at the same time Maldacena 2003, §§ 2–3. A smooth saddle expansion can miss all of them.

These mechanisms can have similar exponential sizes without being physically equivalent. An $e^{-N^2}$ correction to a partition function from a competing saddle and an entropy-suppressed term in a declared late-time normalization—for example, an $e^{-S_{\mathrm{spec}}}$ plateau after dividing a spectral form factor by the square of the number of levels in the window—must not be identified solely from scaling.

A D-brane-sensitive boundary observable

Consider the standard top-down example of $\mathcal N=4$ $SU(N)$ super-Yang–Mills theory and type-IIB strings on $AdS_5\times S^5$. Fixed-length, normalized single-trace correlators organize into a genus expansion. By contrast, the following unnormalized, supersymmetry-protected half-BPS operator contains $N$ elementary fields:

$$\mathcal O_{\det}=\det Z =\frac{1}{N!} \epsilon_{i_1\cdots i_N}\epsilon^{j_1\cdots j_N} Z^{i_1}{}_{j_1}\cdots Z^{i_N}{}_{j_N}$$

It creates a state with charge of order $N$; its normalization must be supplied when it is inserted in a correlator. Determinants and subdeterminants furnish boundary descriptions of giant-graviton D3-branes, and their finite-rank cutoff reproduces the stringy exclusion principle Balasubramanian et al. 2002, §§ 2–4. This operator–brane dictionary builds on the giant-graviton construction of McGreevy, Susskind, and Toumbas 2000, § 3.2.

A recent protected calculation supplies a complementary finite-$N$ check. For a class of purely adjoint $U(N)$ partition functions and superconformal indices, the $m$th correction to the infinite-$N$ answer is of order $e^{-mN}$ and can be reproduced by eigenvalue instantons in the unitary matrix integral Chen, Mahajan, and Tang 2026, pp. 1–3. This evidence is restricted to those partition functions, indices, and matrix-integral assumptions; it is not a universal statement about unprotected observables.

A clean boundary test compares two measurements:

1. a four-point function of fixed-dimension single-trace operators, computed through several genera; and
2. a normalized two-point function or transition amplitude involving $\mathcal O_{\det}$ or a subdeterminant with charge $J\sim N$.

The first tests the coefficients $a_g$ in a fixed-charge sector. The second changes the scaling of the external state and directly probes finite-rank and brane physics. Perfect agreement of the perturbative genus series in measurement 1 does not predict measurement 2, because the limit $N\to\infty$ with $J$ fixed differs from the limit with $J/N$ fixed. The determinant test is therefore a charge-scaled brane-state test, not by itself an exponentially suppressed virtual-brane amplitude.

A genuinely exponential D-brane sector

In the same dual pair, a D$(-1)$-brane is dual to a Yang–Mills instanton. With the standard D3-brane conventions

$$\tau=\frac{\theta}{2\pi}+\frac{4\pi i}{g_{\mathrm{YM}}^2} =C_0+\frac{i}{g_s}, \qquad \lambda=g_{\mathrm{YM}}^2N,$$

the one-instanton factor is

$$e^{2\pi i\tau} =e^{-8\pi^2/g_{\mathrm{YM}}^2+i\theta} =e^{-8\pi^2N/\lambda+i\theta}.$$

At fixed ’t Hooft coupling $\lambda$, the one-instanton sector carries the semiclassical weight $e^{-8\pi^2N/\lambda}$, up to determinant, collective-coordinate, zero-mode, and operator-normalization prefactors. A zero-mode-saturating observable such as the sixteen-fermion supercurrent correlator detects this sector. Its exact $SU(N)$ dependence and large-$N$ AdS/CFT match were computed by Dorey et al. 1998, pp. 145–151, extending the $SU(2)$ calculation of Bianchi et al. 1998, §§ 3–5. Generic correlators need not receive this contribution: the instanton zero modes and selection rules must be saturated. This is a direct comparison between algebraic genus data and a D-brane-scale exponential.

Orders of limits and observable dependence

Three scaled limits recur:

$$\begin{gathered} N\to\infty\ \text{with }J\text{ fixed},\\ N\to\infty\ \text{with }J/N\text{ fixed},\\ N\to\infty\ \text{with }t/t_H\text{ fixed}. \end{gathered}$$

The first supports ordinary few-particle genus counting. The second retains brane states and can retain finite-rank constraints when the fixed ratio reaches the relevant threshold. The third resolves mean distinct-level spacings in observables such as a suitably energy-filtered spectral form factor; it is not a universal recurrence limit for every correlator. None can be substituted for another. Likewise, an exponential invisible in a vacuum correlator may be leading in a charged sector or in the difference between two nearly degenerate saddles.

Jackiw–Teitelboim gravity: fixed theories and ensembles

A fixed theory has one Hamiltonian and one spectrum; a matrix ensemble averages observables over many Hamiltonians. In Jackiw–Teitelboim gravity, the double-scaled matrix integral reproduces the topological expansion and supplies a nonunique ensemble completion. Its exact spectral predictions are ensemble averages, while any one matrix realization has a particular, noisy discrete spectrum Saad, Shenker, and Stanford 2019, §§ 1, 3, and 5. For $n$ boundaries, a genus-$g$ term carries $e^{S_0(2-2g-n)}$. Define the model-specific topological coupling $g_{\mathrm{top}}\equiv e^{-S_0}$, distinct from the ten-dimensional string coupling above; each added handle costs $g_{\mathrm{top}}^2$ Mertens and Turiaci 2023, §§ 4.3–4.4 and 5.2.

For a chosen completion with a positive effective barrier, a forbidden-region one-eigenvalue saddle carries $\exp[-c/g_{\mathrm{top}}]$. Allowed-region determinant sectors instead carry oscillatory phases $\exp[\pm i c(E)/g_{\mathrm{top}}]$. In the connected two-level correlator, their cosine term combines with the perturbative cylinder term, while the full correlator also contains the delta-function contact term from counting the same eigenvalue twice. Together these produce the local sine-kernel and contact structure whose Fourier transform gives the ensemble-averaged ramp and plateau Saad, Shenker, and Stanford 2019, §§ 5.3 and A.2.

With the ensemble-mean density $\bar\rho(E)=e^{S_0}\rho_0(E)+\cdots$, the instanton weight and the Heisenberg time $t_H(E)=2\pi\bar\rho(E)$ are different kinds of quantities. The scale $e^{S_0}$ is a parameter of this matrix model and should not automatically be identified with a boundary gauge-group rank $N$. Moreover, a separate topological model shows that generic gravitational-ensemble elements can fail boundary factorization and positivity tests, so an ensemble interpretation requires its own Hilbert-space checks Marolf 2025, §§ 2–4. The JT example neither establishes that a generic higher-dimensional holographic CFT is an ensemble nor fixes the interpretation of its exponentially small sectors.

Common pitfalls

Calling every brane probe an exponential correction. A determinant inserted as an external operator creates a charge-scaled brane state. An exponentially suppressed virtual-brane contribution is a different observable and requires its own action, zero-mode analysis, and selection rules.

Identifying a mechanism from its exponential size. Branes, competing spacetime saddles, spectral discreteness, and finite-rank identities can produce numerically similar scales. The charges, contour, observable, and independent semiclassical data decide the interpretation.

Replacing a fixed theory by an ensemble without saying so. A smooth ensemble-averaged plateau does not equal the noisy spectral form factor of one Hamiltonian. State the averaging, smoothing, or energy-window prescription before making a late-time claim.

Counting degeneracies as new level spacings. Random-matrix comparison requires a fixed symmetry sector and distinct energy values. Thermodynamic entropy can replace $S_{\mathrm{spec}}$ only after controlling parametrically large degeneracies.

Exercises

1. Prove exponential suppression

Show that $e^{-aN}=o(N^{-K})$ for every fixed $a>0$ and integer $K>0$.

Solution — exponential suppression

Consider $N^Ke^{-aN}$. Its logarithm is $K\log N-aN$, which tends to $-\infty$. Hence the product tends to zero, proving that the exponential is smaller than every fixed inverse power.

2. Locate the finite-rank crossover

Why can Cayley–Hamilton relations matter at leading order along some paths with $J\sim N$ while remaining invisible for every fixed operator length? Why is $J\sim N$ not sufficient by itself?

Solution — finite-rank crossover

For an $N\times N$ matrix, independent trace relations first constrain sufficiently high powers and products. Any operator of fixed length remains below those thresholds once $N$ is large. If the length grows proportionally to $N$, it can reach an $O(N)$ trace-identity threshold, so finite-rank relations may alter leading state counting. The proportionality constant and sector matter: a path remaining below every relevant threshold need not activate such a relation.

3. Compare instanton scaling limits

Rewrite the Yang–Mills one-instanton factor in terms of $\lambda=g_{\mathrm{YM}}^2N$. How does its large-$N$ scaling differ between fixed $\lambda$ and fixed $g_{\mathrm{YM}}^2$?

Solution — instanton scaling

The magnitude is

$$\left\lvert e^{2\pi i\tau}\right\rvert =e^{-8\pi^2/g_{\mathrm{YM}}^2} =e^{-8\pi^2N/\lambda}.$$

At fixed $\lambda$ it is exponentially small as $e^{-O(N)}$. At fixed $g_{\mathrm{YM}}^2$ it is independent of $N$, while $\lambda$ grows with $N$. The phrase “nonperturbative in $1/N$” therefore presupposes the ’t Hooft scaling limit.

4. Decide which exponential is resolved

Suppose a normalized observable has a controlled relative remainder $O(N^{-12})$ in a fixed kinematic domain. Compare corrections $e^{-aN}$ and $e^{-bN^2}$ at asymptotically large $N$. Does this remainder bound determine either correction?

Solution — resolving exponential sectors

Both exponentials eventually lie below this algebraic error band, with $e^{-bN^2}$ becoming smaller more rapidly. The bound therefore resolves neither correction and does not determine its action or sector weight. Resolving one requires a sector-sensitive observable or an independent microscopic or semiclassical definition.

5. Separate handles from eigenvalue sectors

At fixed boundary number $n$, what factor relates the genus-$g+1$ and genus-$g$ JT weights? For fixed $c>0$, can a power series in that factor generate $\exp[-c/g_{\mathrm{top}}]$?

Solution — topological and exponential sectors

The ratio is

$$\frac{e^{S_0(2-2(g+1)-n)}}{e^{S_0(2-2g-n)}} =e^{-2S_0}=g_{\mathrm{top}}^2.$$

Every perturbative handle contribution is therefore algebraic in $g_{\mathrm{top}}^2$. The function $\exp[-c/g_{\mathrm{top}}]$ is smaller than every such power as $g_{\mathrm{top}}\to0^+$, so it belongs to a separate nonperturbative sector.

Limits of the claim

Evidence cutoff. The literature check for this page extends through 28 August 2026. Top-down dualities, protected brane operators and indices, matrix models, and semiclassical saddles give explicit examples of exponential and finite-$N$ physics. They show that perturbative genus data can be incomplete, but do not provide a universal list of sectors, determine a unique contour from the asymptotic series, or prove a common completion for all holographic theories.

For detailed brane origins, see D-Branes, Open-Closed Duality, and Gauge Sectors; for completion tests, see Nonperturbative Definition Proposals: Objects, Evidence, and Falsifiers; and for topology and ensemble questions, see Non-Unique JT Matrix-Integral Completion: Fixed-Theory, Disorder-Average, and Ensemble Distinctions and Fixed-Theory Factorization and Nonperturbative Completion Tests. Mutable saddle, ensemble, and finite-$N$ assessments continue in Holographic Reconstruction and Gravitational Path Integrals and the broader Holography and Quantum Gravity research field guide. The next article, Necessary, Sufficient, and Heuristic Bulk Criteria, synthesizes what these limitations allow one to claim.

The chapter overview contains the structure diagram and validity and failure diagram. They are embedded there once so that their shared chapter-level context is not repeated on every article.

For the chapter-wide comparison of assumptions, counterevidence, falsifiers, and claim ceilings, see the claim-domain table.

References

• Altland, Alexander, and Julian Sonner. “Late Time Physics of Holographic Quantum Chaos.” SciPost Physics 11, 034 (2021). doi:10.21468/SciPostPhys.11.2.034. Open PDF.
• Aniceto, Inês; Gökçe Başar; and Ricardo Schiappa. “A Primer on Resurgent Transseries and Their Asymptotics.” Physics Reports 809, 1–135 (2019). doi:10.1016/j.physrep.2019.02.003. Open PDF.
• Balasubramanian, Vijay; Berkooz, Micha; Naqvi, Asad; and Strassler, Matthew J. “Giant Gravitons in Conformal Field Theory.” Journal of High Energy Physics 2002, 034 (2002). doi:10.1088/1126-6708/2002/04/034. Open PDF.
• Bianchi, Massimo; Michael B. Green; Stefano Kovacs; and Giancarlo Rossi. “Instantons in Supersymmetric Yang–Mills and D-Instantons in IIB Superstring Theory.” Journal of High Energy Physics 1998, 013 (1998). doi:10.1088/1126-6708/1998/08/013. Open PDF.
• Chen, Yiming; Raghu Mahajan; and Haifeng Tang. “Giant Graviton Expansion from Eigenvalue Instantons.” Journal of High Energy Physics 2026, 147 (2026). doi:10.1007/JHEP06(2026)147. Open PDF.
• Dorey, Nick; Valentin V. Khoze; Michael P. Mattis; and Stefan Vandoren. “Yang–Mills Instantons in the Large-N Limit and the AdS/CFT Correspondence.” Physics Letters B 442, 145–151 (1998). doi:10.1016/S0370-2693(98)01233-7. Open PDF.
• Maldacena, Juan. “Eternal Black Holes in Anti-de Sitter.” Journal of High Energy Physics 2003, 021 (2003). doi:10.1088/1126-6708/2003/04/021. Open PDF.
• Marolf, Donald. “On the Nature of Ensembles from Gravitational Path Integrals.” Journal of Physics A: Mathematical and Theoretical 58, 035204 (2025). doi:10.1088/1751-8121/ad9cd4. Open PDF.
• McGreevy, John; Susskind, Leonard; and Toumbas, Nicolaos. “Invasion of the Giant Gravitons from Anti-de Sitter Space.” Journal of High Energy Physics 2000, 008 (2000). doi:10.1088/1126-6708/2000/06/008.
• Mertens, Thomas G., and Gustavo J. Turiaci. “Solvable Models of Quantum Black Holes: A Review on Jackiw–Teitelboim Gravity.” Living Reviews in Relativity 26, 4 (2023). doi:10.1007/s41114-023-00046-1. Open PDF.
• Polchinski, Joseph. “Dirichlet-Branes and Ramond–Ramond Charges.” Physical Review Letters 75, 4724–4727 (1995). doi:10.1103/PhysRevLett.75.4724.
• Saad, Phil; Shenker, Stephen H.; and Stanford, Douglas. “JT Gravity as a Matrix Integral.” arXiv:1903.11115 [hep-th] (2019). arXiv:1903.11115. Open PDF.