Pith. sign in

REVIEW 2 major objections 5 minor 29 references

Universality for random matrices with an edge spectrum singularity

T0 review · 2 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read Edge-scaled random matrix averages with a Fisher–Hartwig singularity converge to a Fredholm determinant built from a Painlevé-XXXIV Riemann–Hilbert problem, with the full Gaussian asymptotics expressed through Painlevé-II sigma functions.

desk verdict Genuine Fredholm-level edge universality with a Fisher-Hartwig singularity for general V and complex β, but the Painlevé boundary data rest on an imported asymptotic lemma the authors admit is not precisely proved. read the letter →

arxiv 2411.18550 v2 pith:ZS2JJHUN submitted 2024-11-27 math-ph math.CAmath.MPnlin.SI

classification math-phmath.CAmath.MPnlin.SI MSC 47B3545B0530E2534E05
keywords randommatrixaveragesFisher-HartwigsingularitysoftedgePainlevé-XXXIVFredholmdeterminantsHankelRiemann-HilbertproblemscharacteristicpolynomialCLT
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper studies the large-n limit of invariant Hermitian random matrix ensembles when the eigenvalue average contains a Fisher–Hartwig-type singularity (a power-law modulus with a phase jump) placed exactly at the soft edge of the spectrum. It proves a universality theorem: for every one-cut regular real-analytic potential V, the edge-scaled average converges to a Fredholm determinant whose kernel is constructed from the Painlevé-XXXIV model Riemann–Hilbert problem. For the Gaussian case V(x)=$x^{2}$, it obtains the complete asymptotic expansion of this average, including the $n^{{1/3}}$ scaling and the logarithmic-in-n term, with the subleading term expressed through Painlevé-II $\sigma$ functions. The work extends earlier Gaussian, β=1 results of Forrester and Witte, and it establishes a central limit theorem for the logarithm of the absolute value of the edge-scaled characteristic polynomial.

What carries the argument

The central object is the Painlevé-XXXIV model Riemann–Hilbert problem (RHP B.1), a piecewise-constant jump problem whose solution Q(ζ;x,α,β) satisfies a Lax pair whose compatibility yields the Painlevé-XXXIV equation for q and the Jimbo–Miwa–Okamoto σ-Painlevé-II equation for σ=1/$4x^{2}$−a. This model problem replaces the Airy parametrix near the edge and provides the boundary data (17), (19) that enters the Painlevé formulas. The argument is carried by the Deift–Zhou nonlinear steepest descent analysis of the orthogonal-polynomial RHP, combined with the Its–Izergin–Korepin–Slavnov theory of integrable integral operators for the Fredholm determinant, and differential identities (36)–(38) for the Hankel determinants.

What would settle it

Compute numerically the left-hand side of (24) for, say, V(x)=$x^{2}$, a fixed α>−1, β≠1, and a few values of s, using Monte Carlo or high-precision quadrature for the Gaussian unitary ensemble with the Fisher–Hartwig factor, and compare the result to the right-hand side with the Painlevé-II $\sigma$ functions evaluated via their connection formulas; a mismatch beyond the claimed o(1) error would disprove Corollary 1.9.

Watch

Extended reading notes

Core claim

The central claim is that, for every one-cut regular real-analytic V and all α>−1, β∈C\((−∞,0), the edge-scaled average E_n[φ_n;λ_n,α,β;V] converges to a Fredholm determinant with kernel $A_s^{{αβ}}$ built from the Painlevé-XXXIV RHP (Theorem 1.5). For the quadratic potential, the full asymptotic is ln E_n = nα/2(1−ln2) + αs $n^{{1/3}}$ + ($α^{2}$/6)ln n + Ξ_α(s,β)+o(1), with Ξ_α expressed through Painlevé-II $\sigma$ functions (Corollary 1.9). The paper also derives ratio asymptotics for Hankel determinants with a single Fisher–Hartwig singularity for general one-cut potentials (Theorem 1.10), and a CLT for the edge-scaled log-characteristic polynomial (Corollary 1.11).

Load-bearing premise

The whole argument leans on the unique solvability and the large-x asymptotic expansions of the Painlevé-XXXIV model Riemann–Hilbert problem (summarized in Lemma B.3); if those asymptotics fail on the needed domain, the boundary conditions for the Painlevé formulas and the Hankel determinant expansions would lose their controlling input.

Editorial extensions

If this is right

  • If the paper's theorems hold, the edge-scaled generating functional for these ensembles is universal: the same Painlevé-XXXIV kernel appears for every one-cut regular V, not just the Gaussian weight.
  • The Gaussian expansion (24) gives the full leading-order large-n behavior of the average with a Fisher–Hartwig singularity at the edge, including all terms that diverge with n and the constant Ξ_α(s,β), which is new for β≠1.
  • The ratio result (26) makes precise how the edge-scaling of the singularity modifies the known bulk Fisher–Hartwig ratio asymptotics, with fractional powers of n and Painlevé functions entering only in the edge-sensitive terms.
  • The CLT (27) shows that the logarithm of the absolute value of the edge-scaled characteristic polynomial, after subtracting a deterministic n−1/3 shift, is asymptotically normal with variance (ln n)/3.
  • The Fredholm determinant formula (16) expresses the limiting distribution as the exponential of an integral of a difference of two Painlevé-II sigma solutions, a structure that appears to be new for Fredholm determinants in integrable systems.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The difference-of-two-Painlevé-transcendents structure in (16) likely extends to other edge-singularity models where a single Fredholm determinant encodes both sides of a jump; one might look for analogous formulas in thinned or conditioned edge processes.
  • The Painlevé-XXXIV model problem is the natural 'master kernel' for edge Fisher–Hartwig singularities; one could test numerically whether the Fredholm determinant on L^2(0,∞) with kernel A_s^{αβ} reproduces known Tracy–Widom-type distributions for special parameter choices beyond (α,β)=(0,1).
  • The CLT (27) is stated for the log-modulus of the characteristic polynomial; a similar argument with an additional phase factor might yield a joint CLT for the real and imaginary parts, connecting to the logarithmic-correlated-field picture at the edge.
  • Since the ratio asymptotics (26) hold for general one-cut V, one could try to push the same RHP approach to the bulk of the spectrum, where the model problem should degenerate to a Fourier-type kernel and the Painlevé-XXXIV data should drop out, recovering known bulk Fisher–Hartwig asymptotics.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. The paper studies the generating functional E_n[φ;λ,α,β;V] for Hermitian one-cut regular random matrix ensembles with a Fisher-Hartwig singularity placed at the soft edge. The main results are: (Theorem 1.5) a universal soft-edge limit of the Fredholm determinant factor F_n in terms of a kernel A_s^{αβ} built from the Painlevé-XXXIV model RHP B.1, together with Painlevé formulas (16) and (18) for the limiting hard-edge distribution F(s;α,β); (Theorem 1.7) for V=x^2, full large-n asymptotics of the Hankel determinant ratio including the n^{1/3} and log n terms, with the remainder expressed through Painlevé data; (Theorem 1.10) an extension of the Hankel ratio asymptotics to general one-cut regular V; and (Corollary 1.11) a CLT for the edge-scaled log-characteristic polynomial. The proof uses the Fokas-Its-Kitaev Riemann-Hilbert problem, nonlinear steepest descent with an outer parametrix, an Airy parametrix at the left endpoint, and a Painlevé-XXXIV parametrix at the right endpoint, followed by small-norm estimates and integration of differential identities.

Significance. If the results are correct, they constitute a substantial advance: they lift earlier Gaussian-only or β=1 results of Forrester-Witte and of Wu-Xu-Zhao to general one-cut regular potentials and to complex β∈C\(−∞,0), and they give the first full expansion of E_n including the n^{1/3} and log n corrections. The Riemann-Hilbert architecture is coherent and well matched to the problem: the explicit outer and local parametrices, the small-norm ratio problem, and the Fredholm determinant convergence argument are all present. The paper also gives concrete, falsifiable predictions, e.g. the α^2/6 log n term and the reduction to the Tracy-Widom distribution when (α,β)=(0,1). The main weakness is that the controlling boundary data for the Painlevé formulas are imported from an external asymptotic statement, Lemma B.3, which the text itself describes as not precisely covered by the cited references.

major comments (2)
  1. [Appendix B.2, Lemma B.3 and Corollary B.4] The large-x asymptotics of the model RHP B.1 are load-bearing for the central claims: equation (123) and Corollary B.4 determine the boundary constraints (17) and (19), which in turn fix the additive constants in the Painlevé representations (16) and (18) and enter the Hankel determinant expansions (21)–(26). The text explicitly states that references [16,27,29] do not precisely match the needs of RHP B.1 and that the results are only summarized. In particular, the exponentially small correction in (123) contains a prefactor (e^{iπα}−β) and an exponent depending on α and β uniformly for β∈C\(−∞,0), which is not visibly established in the cited papers. Since an error in this boundary data would propagate directly into F(s;α,β), η_α(s,β), and the full expansion (24), the authors must either prove Lemma B.3 in the needed normalization and domain, or identify a precise reference with matching hypotheses. As it stands, this is a new unproved assertion at the foundation of the main theorems.
  2. [Section 4, proof of (14)] The dominated convergence argument for the Fredholm series in Theorem 1.5 is written for β≥0 using the Hadamard bound (67), which relies on positive definiteness of the kernel. The passage to β∈C\(−∞,0) is then made by an application of Vitali's convergence theorem and the identity theorem. This requires a uniform domination of the series terms on compact subsets of β∈C\(−∞,0), not merely pointwise convergence of the kernel. The text does not spell out the needed local-uniform L^1 bounds for complex β; such bounds presumably follow from the same exponential decay estimates used for β≥0, but the argument should be made explicit, because the Hadamard inequality used for β≥0 is not available for indefinite kernels.
minor comments (5)
  1. [Title and abstract] The arXiv header contains the typo 'MA TRICES'; the word should be 'MATRICES'.
  2. [Section 1.5 and Corollary 1.11] There are minor typos: 'Assumptoin' in the statement of Corollary 1.11 and 'lenghty' in Section 1.5. These should be corrected in the final version.
  3. [Equations (16) and (18)] The Painlevé determinant formulas are stated for β∈C\(−∞,1), while the kernel convergence (14) is stated for β∈C\(−∞,0). The reason for the extra restriction (invertibility of I−A_s^{αβ} on L^2(0,∞)) is explained in Section 4, but it would help the reader to state this distinction explicitly in the theorem and to note that (14) itself remains valid for β∈C\(−∞,0).
  4. [Section 6, Proposition 6.11] In the proof of Proposition 6.11 the text says 'combining the above we obtain (113)', but the displayed result of that proposition is equation (114). Later, Corollary 6.12 refers to 'combining (94), (98) and (113)', where (114) appears to be intended. Please correct the cross-references.
  5. [Section 5, Corollary 5.10 and Theorem 1.7] The proof of (21) is carried out for α∈(−1,∞)\Z, while the theorem asserts all α>−1. The final extension by regularity is mentioned in the text before Section 3, but it would be helpful to state explicitly in the proof of Theorem 1.7 that the exceptional integer values follow by continuity in α, with the error term uniform on compact sets.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main asymptotics are derived from the model RHP and Painlevé Lax-pair identities, not from the statements they predict.

full rationale

The central derivations are self-contained chains of RHP and Painlevé analysis rather than predictions that reduce to their own inputs. Formula (16) is obtained in Section 4 from the undressing transformation and the Its–Izergin–Korepin–Slavnov framework: equation (75) gives d/ds ln F(s;α,β) = a(s,α,β−1) − a(s,α,β) = σ(s,α,β−1) − σ(s,α,β), and integrating with F(∞;α,β) = 1 yields (16); this is a genuine reduction, not a definitional identity. Likewise (18) follows from Lemma B.2 and the same boundary data, and (21) is derived by integrating the differential identities (36), (37), with the function η_α(s,β) given explicitly by (92); equation (22) is a consequence of that construction rather than a fitted parameter renamed as a prediction. Theorem 1.10 follows from the θ-deformation identity (38) and the explicit contour integral evaluations (105)–(113), with no free parameter fitted to the target expansion. The model problem B.1 and its large-x asymptotics are external inputs taken from [15,16,27,29], none of which are self-citations of the present authors; the text explicitly states that 'None of those three references precisely match our needs for RHP B.1' and summarizes Lemma B.3. This is a correctness or rigor caveat about the provenance of the external boundary data, not circularity: the paper's new statements do not presuppose the target theorems, and the derivation does not rely on a self-citation chain, an ansatz smuggled in via citation, or a renamed known result. The consistency checks against Forrester–Witte, Tracy–Widom, and [2, Theorem 2] further confirm that the new content is independently derived.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The central derivation is mostly self-contained once the Painlevé-XXXIV model problem and its asymptotics are granted; these are taken from prior work, and no new physical entities or fitted parameters are introduced.

assumptions (4)
  • domain assumption Assumption 1.3: V is real analytic, one-cut regular, with support [-sqrt(2), sqrt(2)], density strictly positive in the interior and vanishing like a square root at the endpoints, and the Euler-Lagrange inequality is strict.
    Defines the class of potentials; used throughout for the g-function, Szegő function, parametrices, and estimate (104).
  • domain assumption Unique solvability of the Painlevé-XXXIV model RHP B.1 and the large-x asymptotics in Lemma B.3.
    Imported from [15,16,27,29]; supplies the boundary conditions (17) and (19) and the Painlevé data a, b, Q_k used in all main theorems.
  • domain assumption Orthogonal polynomial RHPs are solvable for large n and the Hankel determinants D_n are nonvanishing in the required parameter domain.
    Underpins the factorization (8) and the differential identities (36)-(38); asserted via Theorem 3.17 and cited general theory.
  • standard math Deift-Zhou nonlinear steepest descent theory and the Its-Izergin-Korepin-Slavnov integrable operator theory.
    Used in Theorem 3.17, the small-norm estimates, and the derivation of (16).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Universality for random matrices with an edge spectrum singularity." pith.science (2026). https://pith.science/paper/ZS2JJHUN

@misc{pith2026241118550,
  author       = {Pith},
  title        = {Pith review of: Universality for random matrices with an edge spectrum singularity},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZS2JJHUN}},
  note         = {Machine review of arXiv:2411.18550}
}
abstract

We study invariant random matrix ensembles \begin{equation*} \mathbb{P}_n(d M)=Z_n^{-1}\exp(-n\,tr(V(M)))\,d M \end{equation*} defined on complex Hermitian matrices $M$ of size $n\times n$, where $V$ is real analytic such that the underlying density of states is one-cut regular. Considering the average \begin{equation*} E_n[\phi;\lambda,\alpha,\beta]:=\mathbb{E}_n\bigg(\prod_{\ell=1}^n\big(1-\phi(\lambda_{\ell}(M))\big)\omega_{\alpha\beta}(\lambda_{\ell}(M)-\lambda)\bigg),\ \ \ \ \ \omega_{\alpha\beta}(x):=|x|^{\alpha}\begin{cases}1,&x<0\\ \beta,&x\geq 0\end{cases}, \end{equation*} taken with respect to the above law and where $\phi$ is a suitable test function, we evaluate its large-$n$ asymptotic assuming that $\lambda$ lies within the soft edge boundary layer, and $(\alpha,\beta)\in\mathbb{R}\times\mathbb{C}$ satisfy $\alpha>-1,\beta\notin(-\infty,0)$. Our results are obtained by using Riemann-Hilbert problems for orthogonal polynomials and integrable operators and they extend previous results of Forrester and Witte \cite{FW} that were obtained by an application of Okamoto's $\tau$-function theory. A key role throughout is played by distinguished solutions to the Painlev\'e-XXXIV equation.

Figures

Figures reproduced from arXiv: 2411.18550 by the authors.

Figure 1
Figure 1. The oriented jump contour ΣS, shown in red, for S(z) in the complex z-plane. (2) The non-tangential limiting values S±(z), z ∈ ΣS \ {√ 2} from either side of ΣS satisfy the relation S+(z) = S−(z)GS(z) with GS(z) = GS(z; s, α, β, n, θ) given by GS(z) =  1 |z − √ 2| αe nη(z) 0 1  , z < − √ 2 − snθ; GS(z) = L(z), z ∈ ∂L+ ∪ ∂L−. For the remaining parts of ΣS on the real line, if √ 2 − snθ < √ 2, then GS(z) =  0 |z − … view at source ↗
Figure 2
Figure 2. The oriented jump contour ΣR, shown in red, for R(z) in the complex z-plane. (2) The limiting values on ΣR satisfy R+(z) = R−(z)GR(z) with GR(z) = GR(z; s, α, β, n, θ) given by GR(z) = I + (√ 2 − z) −α e nξ(z)P(z)  0 0 1 0 P(z) −1 , z ∈ ∂Ω+ ∪ ∂Ω−, and GR(z) = I + |z − √ 2| α e nη(z)P(z)  0 1 0 0 P(z) −1 , z ∈ R \ [− √ 2 − snθ − ϵ, √ 2 + ϵ], on the parts of ΣR disjoint from the circles ∂Dϵ(− √ 2 − snθ) ∪ ∂Dϵ( √ 2… view at source ↗
Figure 3
Figure 3. The oriented jump contour ΣQ, shown in red, for the model function Q(ζ) in the complex ζ-plane. (2) The limiting values Q±(ζ) on ΣQ \{0} satisfy Q+(ζ) = Q−(ζ)GQ(ζ) with GQ(ζ) = GQ(ζ; α, β) given by GQ(ζ) =  1 β 0 1 , ζ ∈ Σ1; GQ(ζ) =  0 1 −1 0 , ζ ∈ Σ3 GQ(ζ) =  1 0 e iπα 1  , ζ ∈ Σ2; GQ(ζ) =  1 0 e −iπα 1  , ζ ∈ Σ4 (3) Near ζ = 0, ζ 7→ Q(ζ) is weakly singular in that, with some ζ 7→ Qb(ζ) = Qb(ζ; x, α, β) ana… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 23 canonical work pages

  1. [15]

    A. R. Its, A. B. J. Kuijlaars, and J. ¨Ostensson,Critical edge behavior in unitary random matrix ensembles and the thirty-fourth Painlev´ e transcendent, Int. Math. Res. Not. IMRN9(2008), Art. ID rnn017, 67, DOI 10.1093/imrn/rnn017

  2. [27]

    Wu, S.-X

    X.-B. Wu, S.-X. Xu, and Y.-Q. Zhao,Gaussian unitary ensemble with boundary spectrum singularity andσ-form of the Painlev´ e II equation, Stud. Appl. Math.140(2018), no. 2, 221–251, DOI 10.1111/sapm.12197. 46

  3. [1]

    Berestycki, C

    N. Berestycki, C. Webb, and M. D. Wong,Random Hermitian matrices and Gaussian multiplicative chaos, Probab. Theory Related Fields172(2018), no. 1-2, 103–189, DOI 10.1007/s00440-017-0806-9

  4. [2]

    Bogatskiy, T

    A. Bogatskiy, T. Claeys, and A. Its,Hankel determinant and orthogonal polynomials for a Gaussian weight with a discon- tinuity at the edge, Comm. Math. Phys.347(2016), no. 1, 127–162, DOI 10.1007/s00220-016-2691-y

  5. [3]

    Bohigas and M

    O. Bohigas and M. P. Pato,Randomly incomplete spectra and intermediate statistics, Phys. Rev. E74(2006), 036212, DOI 10.1103/PhysRevE.74.036212

  6. [4]

    Charlier,Asymptotics of Hankel determinants with a one-cut regular potential and Fisher-Hartwig singularities, Int

    C. Charlier,Asymptotics of Hankel determinants with a one-cut regular potential and Fisher-Hartwig singularities, Int. Math. Res. Not. IMRN24(2019), 7515–7576, DOI 10.1093/imrn/rny009

  7. [5]

    Christophe Charlier, Benjamin Fahs, Christian Webb, and Mo Dick Wong,Asymptotics of Hankel determinants with a multi-cut regular potential and Fisher-Hartwig singularities, Mem. Amer. Math. Soc.310(2025), no. 1567, DOI 10.1090/memo/1567

  8. [6]

    P. A. Deift,Orthogonal polynomials and random matrices: a Riemann-Hilbert approach, Courant Lecture Notes in Math- ematics, vol. 3, New York University, Courant Institute of Mathematical Sciences, New York; American Mathematical Society, Providence, RI, 1999

Show all 29 references
  1. [7]

    Deift, T

    P. Deift, T. Kriecherbauer, K. T.-R. McLaughlin, S. Venakides, and X. Zhou,Uniform asymptotics for polynomials orthogo- nal with respect to varying exponential weights and applications to universality questions in random matrix theory, Comm. Pure Appl. Math.52(1999), no. 11, 1...

  2. [8]

    Deift and X

    P. Deift and X. Zhou,A steepest descent method for oscillatory Riemann-Hilbert problems. Asymptotics for the MKdV equation, Ann. of Math. (2)137(1993), no. 2, 295–368, DOI 10.2307/2946540

  3. [9]

    Y. V. Fyodorov, G. A. Hiary, and J. P. Keating,Freezing Transition, Characteristic Polynomials of Random Matrices, and the Riemann Zeta Function, Phys. Rev. Lett.108(2012), 170601, DOI 10.1103/PhysRevLett.108.170601

  4. [10]

    A. S. Fokas, A. R. Its, and A. V. Kitaev,The isomonodromy approach to matrix models in2D quantum gravity, Comm. Math. Phys.147(1992), no. 2, 395–430

  5. [11]

    L. D. Faddeev and L. A. Takhtajan,Hamiltonian methods in the theory of solitons, English, Classics in Mathematics, Springer, Berlin, 2007. Translated from the 1986 Russian original by Alexey G. Reyman

  6. [12]

    P. J. Forrester and N. S. Witte,Application of theτ-function theory of Painlev´ e equations to random matrices: PIV, PII and the GUE, Comm. Math. Phys.219(2001), no. 2, 357–398, DOI 10.1007/s002200100422

  7. [13]

    Gambier,Sur les ´ equations diff´ erentielles du second ordre et du premier degr´ e dont l’int´ egrale g´ en´ erale est a points critiques fixes, Acta Math.33(1910), no

    B. Gambier,Sur les ´ equations diff´ erentielles du second ordre et du premier degr´ e dont l’int´ egrale g´ en´ erale est a points critiques fixes, Acta Math.33(1910), no. 1, 1–55, DOI 10.1007/BF02393211

  8. [14]

    A. R. Its, A. G. Izergin, V. E. Korepin, and N. A. Slavnov,Differential equations for quantum correlation functions, Proceedings of the Conference on Yang-Baxter Equations, Conformal Invariance and Integrability in Statistical Mechanics and Field Theory, 1990, pp. 1003–1037, D...

  9. [16]

    7, 1523–1558, DOI 10.1088/0951-7715/22/7/002

    ,Asymptotics for a special solution of the thirty fourth Painlev´ e equation, Nonlinearity22(2009), no. 7, 1523–1558, DOI 10.1088/0951-7715/22/7/002

  10. [17]

    Its and I

    A. Its and I. Krasovsky,Hankel determinant and orthogonal polynomials for the Gaussian weight with a jump, Integrable systems and random matrices, 2008, pp. 215–247, DOI 10.1090/conm/458/08938

  11. [18]

    I. V. Krasovsky,Correlations of the characteristic polynomials in the Gaussian unitary ensemble or a singular Hankel determinant, Duke Math. J.139(2007), no. 3, 581–619, DOI 10.1215/S0012-7094-07-13936-X

  12. [19]

    A. B. J. Kuijlaars and K. T-R McLaughlin,Generic behavior of the density of states in random matrix theory and equilibrium problems in the presence of real analytic external fields, Comm. Pure Appl. Math.53(2000), no. 6, 736–785, DOI 10.1002/(SICI)1097-0312(200006)53:6¡736::AI...

  13. [20]

    Okamoto,Studies on the Painlev´ e equations

    K. Okamoto,Studies on the Painlev´ e equations. III. Second and fourth Painlev´ e equations,PII andP IV, Math. Ann.275 (1986), no. 2, 221–255, DOI 10.1007/BF01458459

  14. [21]

    F. W. J. Olver, D. W. Lozier, R. F. Boisvert, and C. W. Clark (eds.),NIST handbook of mathematical functions, U.S. Department of Commerce, National Institute of Standards and Technology, Washington, DC; Cambridge University Press, Cambridge, 2010

  15. [22]

    Pastur and M

    L. Pastur and M. Shcherbina,Eigenvalue distribution of large random matrices, Mathematical Surveys and Monographs, vol. 171, American Mathematical Society, Providence, RI, 2011

  16. [23]

    E. B. Saff and V. Totik,Logarithmic potentials with external fields, Grundlehren der mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], vol. 316, Springer-Verlag, Berlin, 1997. Appendix B by Thomas Bloom

  17. [24]

    C. A. Tracy and H. Widom,Level-spacing distributions and the Airy kernel, Comm. Math. Phys.159(1994), no. 1, 151–174

  18. [25]

    Vanlessen,Strong asymptotics of Laguerre-type orthogonal polynomials and applications in random matrix theory, Constr

    M. Vanlessen,Strong asymptotics of Laguerre-type orthogonal polynomials and applications in random matrix theory, Constr. Approx.25(2007), no. 2, 125–175, DOI 10.1007/s00365-005-0611-z

  19. [26]

    Widom,The strong Szeg˝ o limit theorem for circular arcs, Indiana Univ

    H. Widom,The strong Szeg˝ o limit theorem for circular arcs, Indiana Univ. Math. J.21(1971/72), 277–283, DOI 10.1512/iumj.1971.21.21022

  20. [28]

    Xu and Y.-Q

    S.-X. Xu and Y.-Q. Zhao,Painlev´ e XXXIV asymptotics of orthogonal polynomials for the Gaussian weight with a jump at the edge, Stud. Appl. Math.127(2011), no. 1, 67–105, DOI 10.1111/j.1467-9590.2010.00512.x

  21. [29]

    M. L. Yattselev,Strong asymptotics of Hermite-Pad´ e approximants for Angelesco systems, Canad. J. Math.68(2016), no. 5, 1159–1200, DOI 10.4153/CJM-2015-043-3. School of Mathematics, University of Bristol, Fry Building, Woodland Road, Bristol, BS8 1UG, United King- dom Email a...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.