REVIEW 4 major objections 6 minor 21 references
Hopf's lemmas and boundary behaviour of solutions to the fractional Laplacian in Orlicz-Sobolev spaces
T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash
Pith's one-line read Positive supersolutions of the fractional a-Laplacian have positive boundary slope.
desk verdict A promising but currently unsupported extension of Hopf boundary lemmas to the fractional a-Laplacian; the key continuity proposition has a false estimate and the scaling step in Proposition 5.1 is off by an R^s. 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 load-bearing object is the scaled distance function λ d_B_R(x), whose fractional a-Laplacian can be made uniformly small by taking λ small (Proposition 4.1 and Corollary 4.2). Because the operator is not homogeneous, the power-scaling identity used for the fractional p-Laplacian is unavailable, so this continuity-at-zero property substitutes for it. Comparison principles for weak solutions then transfer the barrier estimate to u, giving u ≥ C d_B_R near the boundary and, in the sign-changing case, a positive lower bound on the difference quotient along admissible directions.
What would settle it
For A(t) = t^p with p ≤ 2 or p ≤ 1/(1−s), construct a bounded Lipschitz u for which (−Δ_a)^s(c_k u) does not tend to 0 uniformly as c_k → 0, or exhibit a positive weak supersolution of (−Δ_a)^s u ≥ c(x)a(u) with c ≤ 0 whose quotient u/δ_R has liminf 0 at an interior-ball boundary point; either observation would mark the true boundary of the theorem's range.
Extended reading notes
Core claim
The central discovery is that, despite the loss of homogeneity of the fractional a-Laplacian, one can still build boundary barriers from scaled distance functions. The key technical step is a continuity property: for a fixed bounded Lipschitz function u, the operator applied to c u tends to 0 uniformly as c → 0, under the two-sided power-type growth condition 0 < p−2 ≤ t a''(t)/a'(t) ≤ q−2 < ∞ together with p > max{1/(1−s), 2}. Using this property, the scaled distance λ d_B_R becomes a weak subsolution, and comparison with the solution of a ball problem yields a positive linear boundary quotient for supersolutions. For sign-changing potentials, the paper adapts an existing growth-condition strategy, replacing homogeneity with a condition on the growth of Φ(r) and obtaining a strictly positive one-sided directional derivative at the boundary.
Load-bearing premise
The two-sided power-type growth condition on the Young function, together with the requirement p > max{1/(1−s), 2}, is what makes the continuity and barrier arguments work; if that fails, for example for the fractional p-Laplacian with 1 < p ≤ 2 or small s, the proofs do not apply.
Editorial extensions
If this is right
- Theorem 1.1 gives a uniform linear lower bound u(x) ≥ Cε d_Ω(x) for torsion-type solutions in a boundary strip, a distance estimate suited to blow-up and symmetry arguments.
- Theorem 1.2 shows that positive supersolutions with a nonpositive potential cannot vanish faster than linearly at boundary points admitting an interior tangent ball.
- Theorem 1.3 yields a strictly positive one-sided directional derivative at the boundary inside a cone of directions even when the potential changes sign, and forces a positive boundary quotient whenever u/δ is continuous.
- The results extend Hopf-type boundary control to the Orlicz-Sobolev setting for operators that are not homogeneous and cannot be rescaled by a constant power.
Reading between the lines
- The continuity-at-zero property likely holds under weaker assumptions than p > max{1/(1−s), 2}; for power-like Young functions, homogeneity suggests uniform convergence to zero for all p > 1, so the range restriction may be an artifact of the proof rather than a true barrier.
- The linear boundary lower bound, rather than the d^s bound familiar from the fractional Laplacian, is a stronger conclusion for a fractional operator; checking it against explicit radial solutions in balls would be a cheap consistency test.
- The same barrier construction might yield a boundary Harnack-type principle or boundary regularity for eigenfunctions of the fractional a-Laplacian, since eigenfunctions are positive supersolutions and the sign-changing result already tolerates indefinite potentials.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes Hopf-type boundary estimates for the fractional a-Laplacian in Orlicz-Sobolev spaces. Three main theorems are claimed: a linear lower bound for solutions of a torsion-like problem (Theorem 1.1), a boundary lower bound for supersolutions with a constant-sign potential (Theorem 1.2), and a strict slope lower bound for sign-changing potentials under a growth condition in the class Z_{x0} (Theorem 1.3). The proofs are built on a continuity property of the operator at zero (Proposition 4.1) and on barriers constructed from distance functions, with comparison principles in Orlicz-Sobolev spaces. The paper also candidly states that earlier proofs in the literature could not be verified and proposes an alternative approach.
Significance. If the results are correct, this would be a valuable contribution to the nonlocal nonlinear PDE literature: the fractional a-Laplacian is a genuinely non-homogeneous operator, and boundary Hopf-type results for it are largely open. The strategy of proving a continuity property of the operator at zero and then using distance-function barriers is natural and, in parts, carefully executed. The assumptions on the Young function are clearly stated, and the paper explicitly identifies the range p>max{2,1/(1-s)} it needs. However, several load-bearing steps in the proofs are currently not justified: the key equicontinuity estimate in Proposition 4.1, the scaling in Proposition 5.1, the admissibility of the barrier in Theorem 1.2, and the use of the growth condition in Theorem 1.3. These gaps affect the central claims, although some appear repairable with modified arguments.
major comments (4)
- [Section 4, Eq. (4.3)] The displayed estimate (4.3) is not valid. From (4.1), the second factor is bounded by 2KL min(|x1-x2|,|z|)/|z|^s, not by 2KL|x1-x2||z|^{1-s}. When |z|<|x1-x2|, the true factor can be as large as C|z|^{1-s}, so the claimed bound fails by a power of |z|. A concrete counterexample is n=1, A(t)=t^3, s=1/2, u(x)=|x|, c_k=1, x1=-a, x2=a, z=a/2: the left side of (4.3) is of order a while the right side is of order a^2. Consequently the uniform equicontinuity of {f_k} is not established and the Arzelà-Ascoli step in Proposition 4.1 collapses. Since Corollary 4.2 and all three main theorems rely on Proposition 4.1, this is a load-bearing error. The statement may be repairable by splitting the integral into |z|<|x1-x2| and |z|>|x1-x2| and using the min bound, but the proof must be rewritten.
- [Section 5, Prop. 5.1] The scaling step in Proposition 5.1 is inconsistent with the definition of a_R in (2.8). With a_R(t)=a(t/R^s) and v(x)=u_R(Rx), the change of variables gives (-Delta_a)^s u_R(Rx)=R^{-s}(-Delta_{a_R})^s v(x), so v solves (-Delta_{a_R})^s v=R^s beta in B1, not beta. Therefore v is not the solution u1 of (5.2), and the uniqueness argument leading to (5.3) is invalid. Since Theorem 1.1 depends on Proposition 5.1, this is a load-bearing gap. A likely fix is to solve (5.2) with right-hand side R^s beta or to apply Corollary 4.2 directly to the functions d_{B_R}; either way the uniform-in-R constants require a separate argument.
- [Section 6, Theorem 1.2 proof] The barrier function introduced in the proof of Theorem 1.2 is not admissible for the comparison principle as written. Even after correcting the evident sign issue (the barrier should be of the form lambda d - chi_D u, not lambda d + chi_D u), the characteristic function chi_D makes the barrier discontinuous across partial D, and this discontinuity is not smoothed. Since D is only assumed to be a smooth domain strictly inside the region and u>0 on partial D, the barrier is not in W^{s,A}(B1 intersection Br) and certainly not in C(B1 intersection Br). The formula for h in (6.1) computes the contribution from the jump as if the function were discontinuous in y only, but the comparison principle Proposition 3.2 requires continuous admissible functions. A mollified cutoff or an approximation argument is needed before the comparison can be applied.
- [Section 7, Eq. (7.8)] The lower bound for h in (7.8) uses the quantity Phi(r) defined in (7.2) with inf_{B_{r/2}(x_r)}|u|, but alpha_r in (7.6) is defined using inf_{B_{r(1-rho)}(x_r)}u. Since B_{r/2}(x_r) is a proper subset of B_{r(1-rho)}(x_r), the latter infimum can be much smaller, so the inequality h <= -2c1 Phi(r) integral is not justified. The hypothesis limsup Phi(r)=+infinity does not imply the corresponding statement with the larger ball; for example, near a point x0 with u(x)=exp(-1/|x-x0|), Phi(r) tends to infinity while the larger-ball quantity tends to zero. This breaks the barrier estimate (7.10) and the proof of Theorem 1.3. The class Z_{x0} or the definition of alpha_r must be adjusted so that the growth condition controls exactly the infimum that appears in the barrier computation.
minor comments (6)
- [Section 6] The proof of Theorem 1.2 uses the same letter u for the given supersolution and for the newly constructed barrier (the line 'Define u(x) = lambda d(x) + chi_D(x)u(x)'). This is confusing and should be changed to a different symbol, e.g. v.
- [Section 4] In the proof of Proposition 4.1, the sentence 'For each n in N and x in Omega, we split fn(x)' should refer to k, not n.
- [Section 7] There are two consecutive steps labelled 'Step 3'; the second one should be 'Step 4'.
- [Section 7] In the final argument of Theorem 1.3, the step from u >= w to u(x_k) >= psi_r(x_k) uses implicitly that u_-(x_k)=0 near x0; this follows from u>0 in B_R cap Omega but should be stated explicitly.
- [Section 3, Prop. 3.5] The proof of the strong maximum principle is terse: the viscosity argument only rules out nonnegative test functions, and the conclusion u identically zero requires an additional argument about the set of zeros, e.g. using the nonlocal nature of the operator or a known strong maximum principle. A fuller proof or a precise citation would help.
- [Section 1] The paper should state more prominently that the standing assumption p>max{2,1/(1-s)} excludes the fractional Laplacian itself (A(t)=t^2) and the fractional p-Laplacian for 1<p<=2; this is not an error but a significant limitation of the scope.
Circularity Check
No significant circularity: the Hopf-type conclusions are derived from the stated growth conditions and independent published lemmas, not assumed by construction.
full rationale
The paper's derivation chain is not circular. The main Hopf-type lower bounds (Theorems 1.1, 1.2, and 1.3) are obtained by building explicit barriers and applying comparison principles; they do not assume the desired boundary behavior. In Theorem 1.2, the comparison function u = λd + χ_D u uses the unknown supersolution u only as an interior correction away from the boundary point, while the linear lower bound at te_n comes from the scaled distance term λd; this is a standard comparison construction, not an assumption of the conclusion. In Theorem 1.3, the class Z_x0 is a genuine extra non-flatness condition rather than a disguised form of (1.7): under the standing assumption p > 1/(1−s), a linear lower bound would make Φ(r) ~ r^{p−1−ps} tend to 0, so the Z_x0 condition limsup Φ(r)=∞ is stronger than, not equivalent to, the conclusion. The key continuity result Proposition 4.1 and its Corollary 4.2 are proved from the growth assumption (2.5) and the power bounds (2.6), not imported as an assumption. Cited results with overlapping authorship ([19], [11], [9], [20], [5]) are published, parameter-free lemmas and theorems that do not state the target Hopf lemma; they are used as independent supporting tools. The paper itself explicitly flags that it cannot verify prior arguments in [6, Lemma 4.1] and [21], but it then supplies its own alternative proof, so that limitation does not create circularity. Any suspected error in estimate (4.3) would be a correctness issue, not a circularity issue, because the estimate is derived rather than assumed.
Assumptions & free parameters
assumptions (5)
- domain assumption Young function satisfies 0 < p-2 <= t a''(t)/a'(t) <= q-2 < infinity, as in equation (2.5).
- domain assumption p > max{1/(1-s), 2} as imposed before Proposition 4.1.
- domain assumption Interior ball condition at the boundary point, equivalently C^{1,1} boundary via [18].
- domain assumption Z_x0 growth condition in Theorem 1.3: limsup_{r to 0} (inf_{B_{r/2}(x_r)} |u|)^(p-1) / r^(ps) = infinity, equations (7.1) and (7.2).
- standard math Imported comparison, regularity, scaling, and decomposition results from [9] and [11], including [11, Theorem 4.9], [11, Lemmas C.1 and C.5], and [9, Lemma 3.10].
Cite this review
Pith. "Pith review of Hopf's lemmas and boundary behaviour of solutions to the fractional Laplacian in Orlicz-Sobolev spaces." pith.science (2026). https://pith.science/paper/CEWJIPWC
@misc{pith2026241113498,
author = {Pith},
title = {Pith review of: Hopf's lemmas and boundary behaviour of solutions to the fractional Laplacian in Orlicz-Sobolev spaces},
year = {2026},
howpublished = {\url{https://pith.science/paper/CEWJIPWC}},
note = {Machine review of arXiv:2411.13498}
}
abstract
In this article we study different extensions of the celebrated Hopf's boundary lemma within the context of a family of nonlocal, nonlinear and nonstandard growth operators. More precisely, we examine the behavior of solutions of the fractional $a-$Laplacian operator near the boundary of a domain satisfying the interior ball condition. Our approach addresses problems involving both constant-sign and sign-changing potentials.
Reference graph
Works this paper leans on
-
[21]
A. Sen, A note on Hopf ’s lemma and strong minimum principle for nonlo cal equations with non- standard growth . In Forum Mathematicum (Vol. 35, No. 6, pp. 1549-1561). De Gruy ter. (2023). 2 22 PABLO OCHOA AND ARIEL SALORT P. Ochoa. Universidad Nacional de Cuyo, F ac. de Ingenier ´ıa. CONICET. Universidad J. A. Maza, Parque Gral. San Mart ´ın 5500, Mendoz...
work page 2023
-
[1]
Boundary behavior of solutions to fractional $p$-Laplacian equation
A. Ataei, Boundary behavior of solutions to fractional p-Laplacian e quation. arXiv preprint arXiv:2304.03624. (2023). 2
work page Pith review arXiv 2023
-
[2]
Bertoin, L´ evy processes, vol
J. Bertoin, L´ evy processes, vol. 121 of Cambridge Tracts in Mathematics, Cambridge University Press, Cambridge, 1996
work page 1996
-
[3]
L. Caffarelli, Non-local Diffusions, Drifts and Games , Springer Berlin Heidelberg, Berlin, Heidel- berg, 2012, pp. 37–52. 1
work page 2012
-
[4]
R. Cont and P. Tankov, Financial modelling with jump processes , Chapman & Hall/CRC Financial Mathematics Series, Chapman & Hall/CRC, Boca Raton, FL, 2004. 1
work page 2004
-
[5]
L. de Borb´ on, L. Del Pezzo and P.Ochoa, Weak and viscosity solutions for non-homogeneous frac- tional equations in Orlicz spaces. Advances in Differential Equations 27 11: 735-780. 1
-
[6]
L. Del Pezzo, and A. Quaas, A Hopf ’s lemma and a strong minimum principle for the fractio nal p-Laplacian. Journal of Differential Equations, 263(1), 765-778. (2017). 9
work page 2017
-
[7]
S. Dipierro, N. Soave and E. Valdinoci, A fractional Hopf Lemma for sign-changing solutions . Communications in Partial Differential Equations, 49(3), 217-241. 1, 2
Show all 21 references
-
[8]
Fall and S
M. Fall and S. Jarohs, Overdetermined problems with fractional Laplacian . ESAIM: Control, Opti- misation and Calculus of Variations, 21(4), 924-938. (2015). 3, 16, 19 1
2015
-
[9]
Fern´ andez Bonder, M
J. Fern´ andez Bonder, M. P´ erez-Llanos, and A. Salort,A H¨ older infinity Laplacian obtained as limit of Orlicz fractional Laplacians . Rev Mat Complut 35, 447–483 (2022). 8, 9
2022
-
[10]
Fern´ andez Bonder and A
J. Fern´ andez Bonder and A. Salort, Fractional order Orlicz-Sobolev spaces . J. Functional Analysis 277, 333–367 (2019)
2019
-
[11]
Fern´ andez Bonder, A
J. Fern´ andez Bonder, A. Salort and H. Vivas, Global H¨ older regularity for eigenfunctions of the fractional g-Laplacian, Journal of Mathematical Analysis and Applications, 526(1), 1273 32, 2023. 8, 13, 15
2023
-
[12]
Fern´ andez Bonder and A
J. Fern´ andez Bonder and A. Salort, Fractional order Orlicz-Sobolev spaces Journal of Functional Analysis 277 (2), 333-367 2, 7
-
[13]
Gilboa and S
G. Gilboa and S. Osher, Nonlocal operators with applications to image processing, Multiscale Model. Simul., 7 (2008), pp. 1005–1028. 1
2008
-
[14]
Gilbarg and N
D. Gilbarg and N. Trudinger, Elliptic partial differential equations of second order . 2nd Ed., Springer Verlag, Berlin, 1983. 1
1983
-
[15]
Greco and R
A. Greco and R. Servadei, Hopf ’s lemma and constrained radial symmetry for the fracti onal Lapla- cian. Mathematical Research Letters, 23(3), 863-885. (2016)
2016
-
[16]
Krasnoselskii y Y
M. Krasnoselskii y Y. Rutickii, Covex functions and Orlicz spaces . Noordhoff, Groningen, The Netherlands, 1961. 1
1961
-
[17]
Hiroaki, T
A. Hiroaki, T. Kilpelainen, N. Shanmugalingam and X. Zhong. Boundary Harnack principle for p-harmonic functions in smooth Euclidean domains . Potential Analysis 26 (3), 281-301. (2007). 5 20
2007
-
[18]
Lewicka and Y
M. Lewicka and Y. Peres, Which domains have two-sided supporting unit spheres at eve ry boundary point, Expositiones Mathematicae, 38(4), 548-558. 7
-
[19]
Molina, A
S. Molina, A. Salort and H. Vivas, Maximum principles, Liouville theorem and symmetry result s for the fractional g-Laplacian, Nonlinear Analysis, 212, 112465, 2021. 6
2021
-
[20]
Ochoa and A
P. Ochoa and A. Salort, Hopf ’s lemmas and boundary point results for the fractional p-Laplacian. Discrete and Continuous Dynamical Systems. doi: 10.3934/dcds.20 24109. 1, 2, 3, 16
Reviewed August 12, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.