A Martin kernel with isolated Harnack contact in an annulus
ORCID: 0009-0001-9320-500X · Zenodo record and DOI
Positive harmonic functions obey Harnack bounds: fixing their value at one point limits how large they can become elsewhere. This note gives an explicit boundary kernel on a smooth annulus that meets the resulting envelope only at the base point within a small neighborhood. It answers the equality-along-a-Green-line extension posed in Hayman and Lingham’s Problem 7.42.
Original manuscript abstract
For the annulus 1 < |z| < e, we exhibit a normalized minimal Martin kernel that is strictly below the envelope of normalized positive harmonic functions at every point of a punctured neighborhood of the base point. Four explicit positive harmonic competitors give a uniform first-order gap in every direction. The result disproves the extension to multiply connected domains of the equality-along-a-Green-line property in Hayman and Lingham's Problem 7.42, attributed to Arsove and Johnson. We give direct constructions of the Green and Martin kernels and a proof of minimality. The broader negative conclusion also has a simpler punctured-disc derivation from classical results; our theorem gives a smooth-annulus construction. OpenAI tools were used extensively in developing and independently checking the argument. This note has not undergone human peer review or formal proof-assistant verification.
Read the paper
Open the original PDF · Download the PDF · BibTeX
The manuscript and Zenodo record identify version 1.0.0 and the declared publication date 12 October 2026. The record was created on 13 October 2026 UTC; the declared date is not a separate certification of when it first became publicly visible.
The annulus and the isolated contact
Fix the annulus D = {z ∈ ℂ : 1 < |z| < e}, the base point a = e1/2, and the inner-boundary point ξ = −1. For positive harmonic functions normalized at a, define the one-sided Harnack envelope
Ha(z) = sup{h(z) : h > 0 harmonic on D, h(a) ≤ 1}.
A Harnack chain makes this supremum finite at every interior point, and Ha(a) = 1. Let Kξ be the minimal Martin kernel at ξ, normalized by Kξ(a) = 1. The theorem proves that there is a neighborhood U of a in D such that
Kξ(z) < Ha(z) for every z ∈ U ∖ {a}.
The normalized kernel therefore cannot agree with the envelope along any nontrivial curve issuing from a, including a Green line. Equality at a is automatic. The theorem does not exclude isolated contact farther away, and it does not assert the same local behavior for every kernel or every annulus.
Why four harmonic competitors are enough
Use logarithmic coordinates z = ex+iy, with 0 < x < 1 and y taken modulo 2π. The annulus becomes a periodic cylinder, with base A = (1/2, 0). Periodizing the classical strip Poisson density constructs the kernel explicitly:
p(x, t) = sin(πx) / [2(cosh(πt) − cos(πx))],
P(x, y) = ∑k∈ℤ p(x, y − π + 2πk),
K(x, y) = P(x, y) / P(A).
The paper constructs the Green function directly and identifies its boundary Martin limit, including tangential approaches. Odd harmonic reflection near the boundary pole and the maximum principle establish minimality.
For the local gap, set b = 1/[2 cosh(1/2)]. Four explicit competitors are
hx,±(x, y) = 1 ± (x − 1/2),
hy,±(x, y) = 1 ± b sin(y) cosh(x − 1/2).
Each is positive and harmonic throughout the cylinder and equals 1 at A. Their gradients point in the two coordinate directions. The kernel gradient is ∇K(A) = (−c, 0), with 0 < c < 8/83 < 1/10, while b > 1/4. These exact bounds give, for every nonzero direction d,
maxj (∇hj(A) − ∇K(A)) · d > |d|/10.
A common quadratic Taylor remainder for this finite family turns the uniform first-order gap into strict inequality at every sufficiently near point other than A. Since the envelope dominates each competitor, it also dominates their maximum. The argument does not assume that the envelope is smooth, and the all-direction estimate is analytic rather than a claim extrapolated from sampled directions.
The question, classical context and priority limits
The target is the extension of equality along a Green line issuing from the base point in Hayman and Lingham’s 2018 Problem 7.42, attributed to Arsove and Johnson. In a disc centered at the base, a normalized Poisson kernel equals the envelope along the radius toward its pole. Equality at only the base point would hold by normalization and is not the extension addressed here.
The smooth annulus has two nondegenerate boundary components. Strip kernels, conformal coordinates, harmonic reflection and the maximum principle are classical tools. Herron, Bear and Smith, and Shiga provide related Harnack-metric results. Symmetric metric identities do not by themselves decide where each prescribed kernel touches this one-sided envelope.
The broader negative conclusion also has a simpler punctured-disc derivation from classical positive-harmonic representations. Köhn’s 1966 classification and the representation discussed by Bear and Smith supply that context. This deduction does not establish earlier recognition of the Green-line consequence and does not yield the smooth-annulus theorem.
No first historical resolution or certification of current worldwide openness is claimed. The 2018 problem-list update reports what its editors knew then. A bounded source search and its full-text gaps do not establish global priority.
Original files and citation
The two original files are provided in the Zenodo record’s order:
- annular-martin-kernel-note.pdf (64,878 bytes) · Zenodo original PDF
- annular-martin-kernel-verification.zip (190,305 bytes) · Zenodo original ZIP
Alec Kriebel. A Martin kernel with isolated Harnack contact in an annulus. Version 1.0.0, 12 October 2026. doi:10.5281/zenodo.23332261. Download BibTeX.
The paper and authored supplementary materials use CC BY 4.0. Cited third-party works retain their own rights.
Status and collaboration
OpenAI tools were used extensively in solving, drafting and checking the argument, including adversarial verification. This is an unrefereed preprint, with no completed human peer review or formal proof-assistant certification. The analytic proof carries the theorem; integrity checks and numerical diagnostics do not replace it.
Independent mathematical verification, substantive developments of the potential-theoretic picture and work toward a journal submission would be useful contributions. Collaboration and any authorship would reflect actual scholarly contributions.
Related literature
- W. K. Hayman and E. F. Lingham, Research Problems in Function Theory (New Edition), arXiv:1809.07200v2 (2018), Problem 7.42 and Update 7.42, pp. 173–174. arXiv version 2.
- D. A. Herron, The Harnack and other conformally invariant metrics, Kodai Mathematical Journal 10 (1987), 9–19. Publisher DOI.
- H. S. Bear and W. Smith, A tale of two conformally invariant metrics, Journal of Mathematical Analysis and Applications 318 (2006), 498–506. Publisher DOI. The paper’s Proposition 2.6 locator refers to the author manuscript dated 23 May 2005.
- H. Shiga, Conformal invariants defined by harmonic functions on Riemann surfaces, Journal of the Mathematical Society of Japan 68 (2016), 441–458. Publisher DOI.
- J. Köhn, Die Harnacksche Metrik in der Theorie der harmonischen Funktionen, Mathematische Zeitschrift 91 (1966), 50–64, especially p. 63. Publisher DOI.