Pith. sign in

REVIEW 2 major objections 8 minor 9 references

Supersolution Construction and Optimal Hardy Inequality for Fractional Laplacians

T0 review · 2 major / 8 minor · reviewed 2026-07-31 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A new supersolution criterion produces optimal Hardy weights on general graphs, including fractional Laplacians, matching the continuous constant on the lattice.

desk verdict Clean sufficient criterion that drops local finiteness and gives optimal fractional Hardy weights on general graphs, with continuum asymptotics on Z^d; null-criticality step leans on companion preprints. read the letter →

arxiv 2607.24169 v1 pith:NR7PHG4T submitted 2026-07-27 math.AP math-phmath.MPmath.SP

Philipp Hake , Matthias Keller , Felix Pogorzelski This is my paper · ORCID
classification math.APmath-phmath.MPmath.SP MSC 35R0231C2047B3939A12
keywords Hardyinequalitysupersolutionconstructionnull-criticalweightsfractionalLaplacianGreenfunctiongraphsRieszkernelstochasticcompleteness
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

Hardy inequalities give quantitative control on functions by comparing energy to a weighted L2 norm. This paper supplies a new, weaker criterion that turns a positive superharmonic potential into an optimal Hardy weight via the supersolution construction. The criterion drops local finiteness and the older oscillation and properness assumptions, so it applies to non-locally finite graphs and in particular to fractional Laplacians. For ordinary Laplacians without killing, the Green function always yields a null-critical (hence optimal) weight. On the integer lattice the resulting weight for the fractional Laplacian has the same leading constant and decay as the continuum theory predicts.

What carries the argument

The supersolution construction: form w = L(u^{1/2})/u^{1/2} from a positive superharmonic potential u. The main technical lemma controls the energy of truncated logarithms of u, producing a null sequence that proves criticality and, under the Green identity, null-criticality.

What would settle it

Exhibit a transient graph with killing for which the Green function satisfies G(c/m) = 1, compute w = L(G_o^{1/2})/G_o^{1/2}, and check whether the ground state lies in L^2(w m); if it does, null-criticality fails exactly as the criterion predicts.

Watch

Extended reading notes

Core claim

On any transient connected graph, every strictly positive superharmonic potential u with Lu integrable produces the critical Hardy-type weight w = L(u^{1/2})/u^{1/2}; when an extra Green-potential identity holds (equivalently G(c/m) < 1), the weight is null-critical and therefore optimal near infinity. Specializing to the Green function and to the Riesz kernel of the fractional Laplacian yields optimal weights on general graphs, and on Z^d the weight asymptotics match the continuous constant c_{d,σ}|x|^{-2σ}.

Load-bearing premise

The weight is optimal only when the superharmonic function is an integrable potential and a Green-potential identity fails to hold; with a killing term that identity can hold and optimality is lost.

Editorial extensions

If this is right

  • Every transient graph without killing admits an optimal Hardy weight built directly from its Green function.
  • Fractional Laplacians on general stochastically complete graphs that satisfy a Nash or Sobolev inequality receive explicit null-critical Hardy weights from the Riesz kernel.
  • On Z^d the fractional Hardy weight has leading term identical to the continuum constant c_{d,σ}|x|^{-2σ}.
  • The same construction extends, via ground-state transform, to positive Schrödinger operators on graphs.

Reading between the lines

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

  • The removal of local-finiteness barriers suggests the same criterion could produce optimal weights for other non-local discrete operators whose Green kernels are known only through spectral calculus.
  • When the lattice error term can be sharpened (as the paper notes is possible for σ = 1), the discrete–continuous comparison becomes quantitative enough for sharp spectral-gap or eigenvalue bounds.
  • Graphs that are stochastically complete at infinity yet retain residual heat at infinity remain the natural test bed for whether null-criticality can survive mild killing.
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 / 8 minor

Summary. The paper develops a new criterion for optimal (null-critical) Hardy weights on general graphs via the supersolution construction w = L(u^{1/2})/u^{1/2}, replacing the properness/bounded-oscillation hypotheses of [KPP18b] by the integrability condition Lu ∈ ℓ¹(X,m) together with u ∈ D₀ ∪ P. The abstract engine is Theorem 8 (criticality, and null-criticality under ∑ mLu ≠ ∑ cu), proved internally via a truncation family λ_T, the ground state transform, and the estimate Q_v(λ_T(u)) ≤ 4 log T ‖Lu‖₁ (Lemmas 7–14, Proposition 14). Theorem 1 and Theorem 2 specialize to superharmonic potentials and to the Green function; Theorem 3 applies this to the fractional Laplacian L_σ on arbitrary (non-locally-finite) stochastically complete graphs, where the Green function is the Riesz kernel; Theorem 4 shows that on Z^d the resulting weight has leading term c_{d,σ}|x|^{-2σ} with the constant expected from the continuum. Section 6 extends the result to subcritical Schrödinger operators via the ground state transform (Theorem 19).

Significance. If the results hold, this is a solid advance in discrete criticality theory. It removes the local-finiteness restriction of earlier supersolution constructions and thereby yields the first optimal Hardy weights for fractional Laplacians on general graphs, with only standard structural hypotheses (transience, stochastic completeness). The criterion is essentially parameter-free: once the graph and the potential u are fixed, w is defined from L and u, not fitted, and the sufficient conditions (Lu ∈ ℓ¹, ∑ mLu ≠ ∑ cu) are checkable and are characterized in terms of stochastic completeness at infinity. On Z^d the weight matches the expected continuum constant c_{d,σ} with controlled error (Theorem 4), a falsifiable asymptotic statement. The internal proof of Theorem 8 is self-contained and carefully written, with explicit citation of the Green formula and Fatou/dominated-convergence passages. The authors are also candid about the non-sharpness of the criterion (remark after Theorem 1, the N₀ example after Theorem 2), which strengthens rather than weakens the paper.

major comments (2)
  1. [§4, proof of Theorem 8] The final step of the proof of Theorem 8 (last paragraph of §4) imports the decisive implication for null-criticality from the companion preprint [HKP26c, Propositions 21 and 22]: that positive criticality of w forces u^{1/2} ∈ D₀ and ∑_X mLu = ∑_X cu. The hypotheses of those propositions are not restated here, and there is a visible setting mismatch to check: Theorem 8 is formulated for Hardy *type* weights w : X → R that may take negative values, whereas [HKP26c] is framed around the fractional Laplacian and (in the cited criticality theory) non-negative weights. If Propositions 21–22 assume w ≥ 0, or impose restrictions on the class D₀ ∪ P or on m, then Theorem 8 as stated would need corresponding restrictions. Since null-criticality is the central conclusion of the paper, I ask the authors to quote the two propositions verbatim (or in a lemma) and verify explicitly that their hypothe
  2. [§4, proofs of Theorem 1 and Theorem 2] The deduction of Theorem 1 from Theorem 8 rests on [HKPS26, Theorem 1]: for every superharmonic u ∈ P, G(c/m) < 1 if and only if ∑ mLu ≠ ∑ cu. Theorem 2 similarly routes through [HKPS26, Theorem 1/Theorem 26]. HKPS26 is a concurrent preprint by (a superset of) the same authors, and the equivalence is stated here only by citation. Given that Theorems 1–2 are the headline results, the manuscript should state the precise hypotheses of the cited theorems (e.g., any measure-finiteness, local-finiteness, or regularity assumptions) and confirm they apply to every transient connected graph over (X,m) and every superharmonic u ∈ P. This is related to Major Comment 1 and can presumably be handled in the same added remark.
minor comments (8)
  1. [§4, proof of Corollary 9] The citation reads '[KL W21, Theorem 6.26x]' — the trailing 'x' appears to be a stray character (presumably a part label such as 6.26 (a)/(b)).
  2. [§4, proof of Lemma 13] 'Propostion 6' is misspelled (should be 'Proposition 6').
  3. [§2 Theorem 3 vs. §5 Theorem 15] Theorem 15(b) proves transience of b_σ for all σ ∈ (0,1] under a Sobolev inequality of dimension d > 2, but Theorem 3 states the conclusion only for σ ∈ (0,1). Presumably this is because L_σ is defined only for σ ∈ (0,1) in this paper, but the discrepancy should be flagged in one phrase to avoid confusion.
  4. [§4, Theorem 8] Theorem 8 is stated for a 'connected graph' with no transience assumption, while the hypothesis u ∈ P only makes sense when the Green function exists (transient case), and Lemma 7(b) invokes the Green operator. For recurrent graphs the potential class is empty and the statement degenerates (e.g., constant u gives w = c). A one-line comment on the standing assumptions under which Theorem 8 is non-vacuous would help the reader.
  5. [§4, proof of Proposition 14] In the first display of the proof, the left-hand side m( w̃−w)(vφ_n)^2(x) has x free while the right-hand side sums over X; it should be stated explicitly that the inequality holds for every fixed x ∈ X (using m > 0), which is what the subsequent T → ∞ argument uses.
  6. [§2, Theorem 4] For σ = 1, d ≥ 3 the error term q = 3 is attributed to [KPP18b, Theorem 7.2]; please confirm that this reference indeed yields O(|x|^{-3}) (rather than only the leading term), since the announced improvement to q = 4 is deferred to [HKP26a].
  7. [§4, Lemma 10] The definitions of λ^±_T use the indicator 1_{[0,t]}(s) with s ranging over [1/T,1] resp. [1,T]; for t ≤ 0 this is empty and the notation is slightly compressed. A half-sentence clarifying λ^±_T(t) = 0 for t ≤ 0 (used later, e.g., λ_T(u(x)) for u(x) < 1/T in Proposition 14) would smooth the reading.
  8. [References / throughout] The citation pattern leans heavily on concurrent preprints of the authors ([HKP26a/b/c], [HKPS26], [Hak25]). This is natural for a research program, but for archival value the key imported statements (Green formula for potentials, Propositions 21–22 of HKP26c, Theorem 1 of HKPS26, the Riesz-kernel asymptotics of HKP26b) should be restated with hypotheses in the final version so the paper remains readable independently of the companions.

Circularity Check

2 steps flagged · score 2.0 of 10

Core criticality proof is internal and parameter-free; null-criticality leans on companion preprints by the same authors but is not forced by definition or fit.

  1. self citation load bearing [Proof of Theorem 8, final paragraph (§4)]
    "Moreover, [HKP26c, Proposition 21 and Proposition 22] states that u^{1/2} ∈ D_0 and ∑_X Lu m = ∑_X cu are necessary conditions for w to be positive critical. Hence, the inequality ∑_X Lu m ≠ ∑_X cu implies null-criticality of w."

    Criticality of w is proved internally, but the upgrade from critical to null-critical (hence optimal near infinity) is not re-proved: it rests entirely on two propositions from a concurrent preprint by the same authors. If those propositions carry hidden restrictions, the null-criticality claim here inherits them without independent verification in this text.

  2. self citation load bearing [Proof of Theorem 1 (§4); also Remark after Theorem 2]
    "By [HKPS26, Theorem 1], the condition for null-criticality G(c/m)(x)<1 is equivalent to ∑_X m Lu ≠ ∑_X c u for every superharmonic u ∈ P. Thus, the result follows immediately from Theorem 8..."

    Theorem 1’s stated null-criticality criterion G(c/m)<1 is identified with the sum condition of Theorem 8 only via another concurrent self-citation (HKPS26). The equivalence is load-bearing for the Green-function and fractional-Laplacian corollaries, though again it is an external lemma rather than a definitional tautology.

full rationale

The paper constructs Hardy-type weights by the supersolution formula w = L(u^{1/2})/u^{1/2} and proves criticality from first principles: Lipschitz cutoffs λ_T, the main technical lemma (Lemma 7), energy comparisons (Lemmas 11–13), and Proposition 14 give Q_v(λ_T(u)) ≤ C log T and force any larger weight to coincide with w. No parameter is fitted to data and then relabeled a prediction; the continuum constant c_{d,σ} appears only as an asymptotic comparison with an independently stated expansion. Null-criticality (the step that upgrades critical to optimal near infinity) does invoke two concurrent self-citations—[HKP26c, Props. 21–22] for the necessary conditions of positive criticality, and [HKPS26, Thm. 1] equating G(c/m)<1 with the unequal-sum criterion. Those citations are load-bearing for the optimality claim but are ordinary mathematical lemmas, not self-definitional loops or uniqueness theorems that forbid alternatives by fiat. The fractional-Laplacian and Z^d applications inherit the same structure plus external heat-kernel/Riesz asymptotics. Overall this is normal concurrent-work dependence, not circularity of the derivation chain. Score 2.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The argument rests on the standard Dirichlet-form calculus of graphs over discrete measure spaces (connectedness, transience, Green operator, ground-state transform) plus the new sufficient conditions Lu∈ℓ¹ and u∈P∪D_0. No numerical free parameters appear. Invented entities are limited to the usual functional-analytic objects (Hardy-type weights, Agmon ground states) already standard in the literature; the paper does not postulate new physical mediators or dimensions.

assumptions (6)
  • domain assumption Connected graphs over discrete measure spaces admit a regular Dirichlet form Q with associated self-adjoint Laplacian L and, when transient, a positive Green function G.
    Taken as background from KLW21; used throughout Sections 2–4 to define potentials and criticality.
  • domain assumption Criticality of a Hardy-type weight is equivalent to existence of a unique (up to scalars) positive Agmon ground state Lv=wv.
    Cited from KPP20 Thm 5.3 / HKP26c Thm 14; used to convert the supersolution construction into criticality.
  • domain assumption Null-criticality is equivalent to the ground state failing to lie in ℓ²(|w|m), and implies optimality near infinity.
    Cited from KPP20 and HKP26c Prop 15; converts the integral test ∑m Lu≠∑cu into the optimality claim.
  • domain assumption For stochastically complete b the fractional Laplacian L^σ is realized by a graph (b_σ,c_σ) with c_σ=0, and its Green function equals the discrete Riesz kernel k_σ.
    Taken from HKP26c Thms 24–25; needed for Theorems 3–4.
  • domain assumption G(c/m)<1 iff ∑m Lu≠∑cu for every superharmonic potential u (and related stochastic-completeness characterizations).
    Cited from concurrent HKPS26; bridges the abstract null-criticality test to the Green-function condition in Theorems 1–2.
  • ad hoc to paper Lipschitz maps fixing 0 send D_0 into itself, and bounded monotone Lipschitz maps preserve the Green identity for potentials with integrable Laplacian (Lemma 7).
    Main new technical lemma; proved in Section 3 from Fatou, dominated convergence and the classical Green formula.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Supersolution Construction and Optimal Hardy Inequality for Fractional Laplacians." pith.science (2026). https://pith.science/paper/NR7PHG4T

@misc{pith2026260724169,
  author       = {Pith},
  title        = {Pith review of: Supersolution Construction and Optimal Hardy Inequality for Fractional Laplacians},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NR7PHG4T}},
  note         = {Machine review of arXiv:2607.24169}
}
read the original abstract

We give a new criterion to show optimality of Hardy weights for general operators on graphs via the supersolution construction. For Laplacians on graphs without killing terms this always gives rise to an optimal Hardy weight via the Green's function without any further assumptions. Furthermore, in contrast to earlier results, our result is not restricted to locally finite graphs. This allows us in particular to obtain optimal Hardy weights for the fractional Laplacian on general graphs. For the fractional Laplacian on the Euclidean lattice, we then obtain an optimal Hardy weight with the constant and asymptotics as it is expected from the continuous setting.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

9 extracted references

  1. [17]

    Optimal Poincar´ e-Hardy-type inequali- ties on manifolds and graphs

    [FR26] Florian Fischer and Christian Rose. “Optimal Poincar´ e-Hardy-type inequali- ties on manifolds and graphs”. In:Indag. Math.37 (2026), pp. 903–933. [Fit00] Patrick J. Fitzsimmons. “Hardy’s inequality for Dirichlet forms”. In:J. Math. Anal. Appl.250.2 (2000), pp. 548–560. REFERENCES 23 [Gup23] Shubham Gupta. “Hardy and Rellich inequality on lattices”...

    2026

  2. [37]

    An improved discrete p-Hardy inequality

    [FKP23] Florian Fischer, Matthias Keller, and Felix Pogorzelski. “An improved discrete p-Hardy inequality”. In:Integral Equations Operator Theory95.4 (2023), Pa- per No. 24,

    2023

  3. [41]

    Heat kernel and essential spectrum of infinite graphs

    2009, pp. 241–263. [Woj09] Rados law K. Wojciechowski. “Heat kernel and essential spectrum of infinite graphs”. In:Indiana Univ. Math. J.58.3 (2009), pp. 1419–1441. 24 REFERENCES [Woj11] Radoslaw Krzysztof Wojciechowski. “Stochastically incomplete manifolds and graphs”. In:Random walks, boundaries and spectra. Vol

    2009

  4. [47]

    Upper bounds for symmetric Markov transition functions

    Springer INdAM Ser. Springer, Cham, [2021]©2021, pp. 1–22. [CKS87] E. A. Carlen, S. Kusuoka, and D. W. Stroock. “Upper bounds for symmetric Markov transition functions”. In:Ann. Inst. H. Poincar´ e Probab. Statist.23.2 (1987), pp. 245–287. [CR18] ´Oscar Ciaurri and Luz Roncal. “Hardy’s inequality for the fractional powers of a discrete Laplacian”. In:J. A...

    2021

  5. [64]

    Progr. Probab. Birkh¨ auser/Springer Basel AG, Basel, 2011, pp. 163–179. P. Hake, Institut f¨ur Mathematik, Universit¨at Leipzig, 04109 Leipzig, Ger- many Email address:philipp.hake@math.uni-leipzig.de M. Keller, Israel Institute of Advanced Studies, Jerusalem, Israel; Insti- tut f¨ur Mathematik, Universit¨at Potsdam 14476 Potsdam, Germany Email address:m...

    2011

  6. [223]

    Bound- ary representation of Dirichlet forms on discrete spaces

    [KLSS19] Matthias Keller, Daniel Lenz, Marcel Schmidt, and Michael Schwarz. “Bound- ary representation of Dirichlet forms on discrete spaces”. In:J. Math. Pures Appl. (9)126 (2019), pp. 109–143. [KL W13] Matthias Keller, Daniel Lenz, and Rados law K. Wojciechowski. “Volume growth, spectrum and stochastic completeness of infinite graphs”. In:Math. Z.274.3-...

    2019

  7. [358]

    Optimal Hardy inequality for fractional Laplacians on the integers

    Grundlehren der mathematischen Wis- senschaften [Fundamental Principles of Mathematical Sciences]. Springer, Cham, 2021, pp. xv+668. [KN23] Matthias Keller and Marius Nietschmann. “Optimal Hardy inequality for fractional Laplacians on the integers”. In:Ann. Henri Poincar´ e24.8 (2023), pp. 2729–2741. [KPP18a] Matthias Keller, Yehuda Pinchover, and Felix P...

    2021

  8. [1990]

    Optimal Hardy weight for second-order elliptic operator: an answer to a problem of Agmon

    [DFP14] Baptiste Devyver, Martin Fraas, and Yehuda Pinchover. “Optimal Hardy weight for second-order elliptic operator: an answer to a problem of Agmon”. In:J. Funct. Anal.266 (2014), pp. 4422–4489. [Fis24] Florian Fischer. “On the optimality and decay ofp-Hardy weights on graphs”. In:Calc. Var. Partial Differential Equations63.7 (2024), Paper No. 162,

    2014

Show all 9 references
  1. [2025]

    Optimal Hardy In- equalities for Random Walks onZ 2

    [HKP26a] Philipp Hake, Matthias Keller, and Felix Pogorzelski. “Optimal Hardy In- equalities for Random Walks onZ 2”. In:preprint(2026). [HKP26b] Philipp Hake, Matthias Keller, and Felix Pogorzelski. “Optimal Hardy In- equality for Fractional Laplacians on the Lattice”. In:arX...

    2026

Pith tools

Reviewed July 31, 2026 · model on record in the stance chip above.