REVIEW 1 major objections 2 minor
Eigenvalues and eigenfunctions of the fractional Laplacian on the interval
T0 review · 1 major / 2 minor · reviewed 2026-08-28 · deepseek-v4-flash
Pith's one-line read The fractional Laplacian on the interval has a three-term eigenvalue expansion and eigenfunctions bounded uniformly in the fractional order and the index.
desk verdict A substantial paper with a real algebraic error in the main three-term coefficient that needs fixing before it can be trusted. 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 argument is carried by the generalized half-line eigenfunction $F_\alpha(t)=\sin(t+\theta)-G_\alpha(t)$, where $G_\alpha$ is a nonnegative, completely monotone correction whose Laplace transform encodes the spectral information of the half-line operator. Gluing two reflected copies of $F_\alpha(\mu_n\,\cdot\,)$ across the interval with a smooth partition of unity produces a quasimode $\tilde\phi_n$ whose Rayleigh quotient already yields the two-term formula; the third term is extracted by expanding the residual $(A_I-\mu_n^\alpha)\tilde\phi_n$ and pairing it with the quasimode. For $0<\alpha<1$, where the cutoff error is the same size as the target term, the paper replaces the cut quasimode by an uncut combination $w_n$ and uses an exact quadratic-form identity to reach the sharper error. The uniform bound uses heat-semigroup estimates for orders bounded away from zero and, for small $\alpha$, a residual estimate of size $o(\alpha/n)$, an antiderivative bound for the eigenfunction, and a truncated quadratic-form test with $(\varphi_n-\Lambda)_+$.
What would settle it
For a fixed $\alpha\in(1,2)$, evaluate $I_\alpha(s)$ in (2.6) numerically near $s=0$ and check that $\pi I_\alpha(s)/s$ approaches $\pi\cot(\pi/\alpha)$; alternatively, compute $\lambda_n(\alpha)$ to high precision for a fixed $\alpha$ and check that $\lambda_n-\mu_n^\alpha-(-1)^n\kappa_\alpha\mu_n^{-2}$ follows the claimed power of $\mu_n$ with the claimed coefficient rather than a different exponent.
Extended reading notes
Core claim
The central result is a complete three-term asymptotic formula for the eigenvalues of the fractional Laplacian in the interval. Writing $\mu_n = n\pi/2 - (2-\alpha)\pi/8$ and $\kappa_\alpha = \alpha c_\alpha/2^{\alpha+1}$, the paper proves $$\lambda_n(\$\alpha$) = \mu_n^\$\alpha$ + (-1)^n \kappa_\$\alpha$ \$mu_n^{{-2}}$ + R_n(\$\alpha$),$$ where for $0<\alpha<1$ the remainder is $(A_\alpha + (-1)^n B_\alpha)\mu_n^{-2-\alpha} + o(\mu_n^{-2-\alpha})$; for $\alpha=1$ it is $(-1)^{n+1}(2\pi^2)^{-1}\mu_n^{-3}\log\mu_n + O(\mu_n^{-3})$; and for $1<\alpha<2$ it is $(-1)^n \eta_\alpha \mu_n^{-3} + o(\mu_n^{-3})$, with explicit constants $A_\alpha$, $B_\alpha$, and $\eta_\alpha$. The same arguments show that the normalized eigenfunctions satisfy $\|\varphi_n\|_{L^\infty(I)}\le C$ for an absolute constant $C$, uniformly in $n\ge 1$ and $0<\alpha<2$. Along the way the paper establishes that the $L^2$ distance between the true eigenfunction and its quasimode decays as $n^{-1-\alpha}$, giving a quantitative spectral approximation.
Load-bearing premise
The entire calculation leans on imported exact formulas for the half-line eigenfunction and its Laplace transform; if any of those formulas were inaccurate, the explicit constants in the three-term expansion would come out wrong.
Editorial extensions
If this is right
- The eigenvalue expansion is now known through three terms for every fixed $0<\alpha<2$, including the precise logarithmic term at $\alpha=1$ and the parity-dependent signs $(-1)^n$.
- The numerically conjectured $O_\alpha(n^{-2})$ remainder in the two-term formula is confirmed, with explicit constants controlling the next order in each regime.
- The uniform bound $\|\varphi_n\|_{L^\infty}\le C$ holds with a single constant for all orders $\alpha$ and all indices $n$, not only for $\alpha\ge 1/2$ as previously known.
- The eigenfunction approximation result $\|\varphi_n - \tilde\phi_n/\|\tilde\phi_n\|\|_{L^2} \lesssim n^{-1-\alpha}$ gives a constructive route to computing eigenfunctions and eigenvalues of the fractional Laplacian in one dimension.
Reading between the lines
- A natural testable extension is to replace the fractional Laplacian by a general Bernstein function $\psi(-\Delta)$ on an interval; the same half-line representation should yield a three-term formula whose constants come from the tail of the corresponding correction term, provided an analogue of the Laplace-transform identity holds.
- In higher dimensions the paper's methods do not directly apply; one could test numerically whether the uniform $L^\infty$ bound persists for balls, where symmetry might allow a similar one-dimensional reduction, and whether it fails for domains with corners.
- The explicit value $Q_\alpha=\alpha^2/(2K_\alpha)$ obtained in the small-order analysis could be used to compute next-order heat-trace or partition-function corrections for the interval, connecting the eigenvalue asymptotics to the short-time heat-kernel expansion.
- The proof suggests the crude bound $\exp(100)$ is far from sharp; a numerical survey of $\|\varphi_n\|_{L^\infty}$ over $\alpha$ and $n$ could locate the actual supremum and reveal whether any non-uniform effect appears as $\alpha\to 0$.
Formalized claims in Lean
-
Claim #1: The central result is a complete three-term asymptotic formula for the eigenvalues of the fractional Laplacian in the interval. Writing $\mu_n = n\pi/2 - (2-\alpha)\pi/8$ and $\kappa_\alpha = \alpha c_\alpha/2^{\alpha+1}$, the paper proves $$\lambda_n(\$\alpha$) = \mu_n^\$\alpha$ + (-1)^n \kappa_\$\alpha$ \$mu_n^{{-2}}$ + R_n(\$\alpha$),$$ where for $0<\alpha<1$ the remainder is $(A_\alpha + (-1)^
/-- @claim 1 The central result is a complete three-term asymptotic formula for the eigenvalues of the fractional Laplacian in the interval. Writing $\mu_n = n\pi/2 - (2-\alpha)\pi/8$ and $\kappa_\alpha = \alpha c_\alpha/2^{\alpha+1}$, the paper proves $$\lambda_n(\$\alpha$) = \mu_n^\$\alpha$ + (-1)^n \kappa_\$\alpha$ \$mu_n^{{-2}}$ + R_n(\$\alpha$),$$ where for $0<\alpha<1$ the remainder is $(A_\alpha + (-1)^ -/ def central_claim : Prop :=
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies the fractional Laplacian (−Δ)^{α/2} with zero exterior condition on the interval I=(−1,1), 0<α<2. Theorem 1 claims a three-term asymptotic expansion for the n-th eigenvalue λ_n(α), with explicit constants κ_α, A_α, B_α, η_α, improving the two-term formula of Kwaśnicki and confirming the numerically predicted O_α(n^{−2}) remainder. Theorem 2 claims an absolute bound on the L^∞ norms of the normalized eigenfunctions, uniform in both α and n. The proofs construct quasimodes from half-line eigenfunctions imported from Kwaśnicki’s work, derive residual estimates in W^{1,1}, and then extract the third term in the eigenvalue expansion; the eigenfunction bound is obtained via heat-semigroup estimates, an antiderivative bound, and a truncated quadratic-form argument for small α.
Significance. If the main results are correct after the coefficient issue described below is fixed, this is a substantial contribution: it settles the conjectural third-order eigenvalue asymptotics for the fractional Laplacian in an interval and the Kwaśnicki conjecture on uniform L^∞ eigenfunction bounds. The paper is commendably explicit: all constants are given in closed form, the proofs are structured through verifiable propositions, and the argument does not fit parameters to the interval eigenvalues. The proof of Theorem 2 is especially strong, as it is uniform in both n and α. The use of the already-proved two-term formula to identify the auxiliary integral Q_α in (5.28) is a consistency check and is not circular, because the two-term formula is established before that point.
major comments (1)
- [Section 5, Proposition 5.4 (final paragraph) and equation (1.7)] The final simplification in Proposition 5.4 is inconsistent with the definition of B_alpha in (1.7). Starting from the coefficient in (5.29), with x=πα/2, the coefficient of −ε μ_n^{−2−α} before the final identification is α c_α Γ(1+2α)/(Γ(1+α) 2^{1+2α}) (2 cos x − sec x). Using c_α/Γ(1+α)=sin x/π and sin x(2 cos x − sec x)=cos(2x) tan x, this equals α Γ(1+2α) cos(πα) tan(πα/2)/(π 2^{1+2α}). The paper instead writes this expression as equal to B_alpha, which has tan(πα/2) in the denominator. Consequently Theorem 1(i) and equation (1.9) are off by a factor tan^2(πα/2) for every 0<α<1 except α=1/2, where the coefficient vanishes. The proof itself supplies the numerator version, so the statement should be corrected by changing (1.7) to the numerator form (or, if the intended coefficient is the denominator form, by recomputing the preceding algebra in (5.29)). This is a load-bearing error in the headline three-term formula and must be fixed before acceptance.
minor comments (2)
- [Throughout; especially (2.3), (2.13), (5.20), and the definitions of K_α and M_α] The notation √α/2 is ambiguous in the typeset text. In context it is consistently used to mean √(α/2), but this is not clear without parentheses. Please write √(α/2) explicitly in all displayed formulas.
- [Section 6, proof of Lemma 6.4] The symbol P is used both for the pointwise quantity P_k = sup_x |∫_{−1}^x φ_k(t)dt| and for the scalar maximum P = max{...}. This is locally clear but could be confusing; renaming the scalar, for example P_max or M, would improve readability.
Circularity Check
No circularity: the third-term constants are derived from half-line eigenfunction formulas and explicit residual computations, not fitted to interval eigenvalues; the only consistency check reuses an independently proved two-term formula.
full rationale
The derivation chain is self-contained in the relevant sense. The half-line representation F_alpha(t)=sin(t+theta)-G_alpha(t), the Laplace transform identity (2.13), and the tail expansions (2.7)-(2.11) are imported from [23, Theorem 1.1, Example 6.1, Lemma 4.21, Lemma 4.27], which is an independent prior work of Kwaśnicki; the present paper does not define the interval eigenvalues in terms of those formulas. The constants kappa_alpha, A_alpha, B_alpha, and eta_alpha are computed from the half-line expansions and from explicit residual pairings such as (3.20), (4.13)-(4.14), and (5.29), rather than fitted to the eigenvalues being expanded. The identification Q_alpha = alpha^2/(2K_alpha) in Proposition 5.4 uses the independently proved two-term formula (3.19) as a consistency check on the coefficient of mu_n^{-2}; this is not circular because Theorem 3 was already established in Section 3 before the small-alpha refined quasimode w_n is introduced. No fitted parameter is renamed as a prediction. The two-term formula is proved before it is used in Section 5, so the later comparison is not a self-definitional loop. The skepticism about Theorem 1(i) concerns an internal algebraic inconsistency: the displayed simplification in the last paragraph of Proposition 5.4 yields tan(pi alpha/2) in the numerator while (1.7) defines B_alpha with tan(pi alpha/2) in the denominator. That is a correctness concern, not a circularity, because it does not reduce the claimed prediction to an input. The paper also settles, rather than assumes, the numerical conjectures of [20] and [24]. Self-citations [17] and [18] are background and are not load-bearing. Therefore the circularity score is 0.
Assumptions & free parameters
assumptions (5)
- domain assumption Half-line eigenfunction F_alpha(t)=sin(t+theta)-G_alpha(t) satisfies the eigenvalue equation (2.2) and the Laplace transform identity (2.13).
- domain assumption The two-sided eigenvalue bounds ((k-1)pi/2)^alpha <= lambda_k <= (k pi/2)^alpha.
- domain assumption The heat-kernel bound ||e^{-tA_I}f||_{L^infinity} <= h_alpha t^{-1/(2 alpha)} ||f||_{L^2}.
- standard math The principal value identity p.v. integral_0^infinity t^{p-1}/(1-t) dt = pi cot(pi p).
- standard math The quadratic form identity E_{alpha,J}(u,v) = (1/2pi) integral |xi|^alpha hat u hat v d xi and the equivalence of definitions of the fractional Laplacian.
Cite this review
Pith. "Pith review of Eigenvalues and eigenfunctions of the fractional Laplacian on the interval." pith.science (2026). https://pith.science/paper/NCHIC3HA
@misc{pith2026260823457,
author = {Pith},
title = {Pith review of: Eigenvalues and eigenfunctions of the fractional Laplacian on the interval},
year = {2026},
howpublished = {\url{https://pith.science/paper/NCHIC3HA}},
note = {Machine review of arXiv:2608.23457}
}
abstract
We prove a three-term asymptotic formula for the eigenvalues of the fractional Laplacian on the bounded interval $(-1,1)$. This improves the eigenvalue asymptotics of Kulczycki--Kwa\'snicki--Ma{\l}ecki--St\'os and Kwa\'snicki, and confirms the conjectural $O_\alpha(n^{-2})$ remainder suggested by the numerical simulations of Kaleta--Kwa\'snicki--Ma{\l}ecki. Moreover, we prove that the normalized eigenfunctions are bounded uniformly in the eigenvalue index $n$ and the fractional order $\alpha$. This settles the conjecture proposed by Kwa\'snicki through numerical experiments. Furthermore, we prove that the $n$-th eigenfunction has exactly $n-1$ zeros in the interval $(-1,1)$ and every zero is simple, and hence there are exactly $n$ nodal domains. A key ingredient in the proof is an explicit representation of the eigenfunction.
Reviewed August 28, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.