REVIEW 1 major objections 5 minor
Extremal Functions and Widom Factors on Compact Riemann Surfaces
T0 review · 1 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash
Pith's one-line read This paper proves that on a compact Riemann surface of positive genus, the n-th root of the Chebyshev constant for any non-polar compact set converges to the set's logarithmic capacity, and that the refined extremal asymptotics are governed
desk verdict The root-law theorem is a genuine, carefully proved result, but the Szegő–Widom half has a concrete product-formula gap in §7.2 that needs repairing before the refined asymptotics 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 machinery is flat unitary line bundles over the punctured surface. The planar factor z−q becomes a meromorphic section B_q of such a bundle with |B_q|=e^{−G(p,q)}, G the bipolar Green kernel; a correction section uniform in the character cancels the moving character and restores single-valued functions for the upper bound. Refined asymptotics use a Cauchy kernel with an auxiliary divisor, whose unwanted poles are removed by a Mittag-Leffler step, and reflection of the Green function across the analytic boundary to a level curve where |Φ_E|=r<1, giving O(r^n) errors. Zeros are counted by dG=2∂g_E, which has 2g+p−1 interior zeros, the genus of the Schottky double.
What would settle it
On the square torus with E={|℘(z)|≤R}, the paper predicts W_{2m}=1 for even orders and W_3=√(R²+e²)/R>1 for order 3. Numerically or symbolically computing these Widom factors and finding W_3=1, or finding W_n→1 for all n, would directly contradict the claimed absence of a single-disc limiting theorem.
Extended reading notes
Core claim
On a compact Riemann surface of genus g>0, monic polynomials become admissible meromorphic functions with a single pole of order N at a marked point. The paper proves that the N-th root of their minimal sup-norm on any compact non-polar set E tends to the logarithmic capacity of E defined by the bipolar Green kernel. The planar factor z−q is replaced by a section B_q of a flat unitary line bundle with |B_q|=e^{−G(·,q)}, with a cohomological correction uniform in the character. For finite unions of real-analytic discs, M_{n,ρ}=cap(E)^n(μ_n+o(1)) and normalized extremals converge locally uniformly to the corresponding extremal section. Consequently Widom factors need not converge; even a singl
Load-bearing premise
The refined Szegő–Widom part rests on the assumption that the boundary curves of the discs and the weight are real-analytic; the proof reflects the Green function across the boundary to obtain a fixed collar with exponentially small errors, and if the boundary is merely smooth that geometric decay—and the quantitative O(r^n) control—collapses.
Editorial extensions
If this is right
- The root asymptotic t_N^{1/N}→cap(E) holds on every compact Riemann surface for any non-polar E avoiding the marked point, so logarithmic capacity is the correct exponential scale for extremal meromorphic approximation.
- A Bernstein–Walsh inequality holds on the surface: admissible functions satisfy |f(p)|≤‖f‖_E e^{N g_E(p)}, extending the planar domination argument to higher genus.
- For finite unions of disjoint analytic discs, the weighted Chebyshev constants satisfy M_{n,ρ}=cap(E)^n(μ_n+o(1)), and the normalized extremal functions converge locally uniformly to the unique extremal section of the flat bundle.
- Widom factors are bounded by the exponential of the sum of the equilibrium Green function over all 2g+p−1 interior critical points, the surface analogue of the Parreau–Widom sum.
- The set of limit points of the normalized constants is a finite union of closed intervals, and equals the full interval when the character orbit is dense; so even a single analytic disc in positive genus can have almost periodically oscillating Widom factors.
Reading between the lines
- A direct numerical test on a genus-one torus with E the inverse image of a disk under the Weierstrass function could check the predicted split: the paper's formulas give even-order Widom factors equal to 1 and a first odd-order factor strictly greater than 1, so observing convergence W_n→1 for all n would refute the claimed oscillation.
- The reliance on real-analytic boundaries suggests the refined asymptotics should persist under C^{2+} regularity, with geometric O(r^n) errors replaced by qualitative o(1); proving that would need a different boundary approximation mechanism.
- The same character-torus mechanism may govern other extremal problems on compact surfaces, such as L^2 or weighted extremal sections, with capacity fixing the exponential rate and the character orbit determining subexponential fluctuations.
- The explicit genus-one examples suggest that parity of the pole order can create an even/odd split in Widom factors, a phenomenon that may be worth probing for other symmetric sets of higher genus.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies Chebyshev extremal problems on a compact Riemann surface X of genus g>0, with a marked point P∞ at infinity. It defines admissible meromorphic functions with the only pole at P∞ and proves (Theorem 2.2) that the N-th root of the Chebyshev constant t_N(E,P∞) converges to the capacity cap(E) of a non-polar compact E, computed using the bipolar Green kernel. The proof follows the classical Fekete-point route, with a uniform character-correction lemma (Lemma 3.2) to handle the flat line bundles attached to the Fekete divisors, and a Bernstein–Walsh inequality (Theorem 4.1) for the lower bound. The second half develops the Szegő–Widom theory under Assumption 2.3 that E is a finite union of p analytic closed discs and ρ is a real-analytic weight. The paper proves a Widom-type H∞ duality (Theorem 2.4), a product formula for the extremal section (Theorem 2.5), a range theorem for the Widom constants (Theorem 2.7), and the main asymptotic Theorem 2.8: M_{n,ρ} = cap(E)^n(μ_n + o(1)) with locally uniform convergence of the normalized extremals to the Widom minimizer, and uniform convergence under a zero-separation hypothesis. It concludes that no Faber-type limit W_n→1 exists in positive genus even for a single analytic disc, and gives worked genus-1 examples via the Weierstrass ℘ function.
Significance. If the results are correct, this is a substantial contribution to constructive potential theory on Riemann surfaces. The root asymptotic (Theorem 2.2) is the natural surface analogue of Fekete–Szegő, and the Szegő–Widom asymptotics (Theorem 2.8) extend Widom's classical memoir to the setting of compact Riemann surfaces, where the flat-bundle character group is genuinely higher-dimensional. The paper is careful to credit prior work (Widom, Read, Bertola) and gives explicit, checkable examples in genus 1. The proofs are long and technical, but the key mechanisms—uniform character correction, the Cauchy-kernel projection, the Neumann-factor product formula, and the boundary compactness argument—are presented in detail. The paper is honest about its technical hypotheses (real-analytic boundaries and weights, and a conditional zero-separation assumption) and clearly states that the C^{2+} extension is not pursued.
major comments (1)
- [§7.2, proof of Theorem 2.5] The displayed definition of the quotient S is missing the inverse on F0. As printed, "S:= μ(ρ,χ) F0 Φ(·,D0)^{-1}" does not give the stated cancellation at D0 and would imply |S| = μ²ρ^{-1} on Γ, not |S| = ρ. The surrounding sentence ("the zero of Φ(·,D0)^{-1} cancels the pole of F0^{-1}") indicates the intended formula is S := μ(ρ,χ) F0^{-1} Φ(·,D0)^{-1}, for which the boundary computation |S| = ρ and the conclusion S = R are correct. The displayed equation must be fixed; with the intended definition, the proof of (2.9) is valid.
minor comments (5)
- [§10.3, equation (10.18)] After bounding |w(w²−e²)| ≤ R(R²+e²) on |w|=R, the text says this "gives the norm of y". The norm of y is the square root of that quantity, so the displayed t_3 is √(R(R²+e²)). The subsequent sentence should say it gives the square of the norm.
- [§4.1] Typo: "Frotman conditions" should be "Frostman conditions".
- [§7.1] The statement "A:= div0(F0η0)" and the later decomposition A = a_1+...+a_ĝ use boundary zeros with half-multiplicity; this convention is explained only briefly and should be made more prominent, since it is used again in (7.3) and (7.5).
- [Remark 2.12] The remark that the H∞ duality is not new and that the novelty lies in the connection to the scalar Chebyshev problem is welcome and should be retained in the published version.
- [References] Some references have inconsistent spellings (e.g., "Achyeser" for Akhiezer). A final proofreading pass would help.
Circularity Check
No material circularity; one non-load-bearing self-citation only.
full rationale
The central derivation is not circular. The capacity cap(E) is defined independently via the bipolar Green kernel (Definition 2.1), while the Chebyshev constants t_N(E,P∞) and Widom factors W_N are defined through the admissible-function extremal problem (2.4). Theorem 2.2 is then proved by a Fekete-divisor argument with a Riemann–Roch correction (Section 4), not by presupposing the equality. The Szegő–Widom part explicitly builds on external benchmarks: Theorem 2.4 is identified as the rank-one case of Widom's duality [55], the product formula follows Widom's memoir [54], and Remark 2.12 explicitly disclaims novelty in the underlying H∞ theory. These are independent supports, not outputs of the paper. The only self-citation, [17] in §2.2, is a literature pointer and is not load-bearing. I also flag, per the reviewing rule, two explicit limitations that are not circularity: Remark 2.9 says real-analyticity is primarily technical and the exponential estimates are not proved under weaker regularity, and Theorem 2.8's uniform version depends on an unproved zero-separation hypothesis. The §7.2 proof of Theorem 2.5 appears to contain a boundary-modulus discrepancy (|S| = μ²/ρ, not ρ), but that is a correctness gap rather than a circular reduction, so it does not raise the circularity score.
Assumptions & free parameters
assumptions (9)
- standard math Riemann–Roch theorem and Serre duality for line bundles on compact Riemann surfaces
- standard math Fekete–Szegő theory and Frostman's theorem for the bipolar Green kernel
- standard math Widom's rank-one H∞ bundle duality theorem
- domain assumption Existence and residue properties of Bertola's Cauchy kernel C_D(q,p)
- standard math Behnke–Stein period theorem and Forster's prescribed-factor-of-automorphy theorem
- standard math Schwarz reflection for real-analytic boundary
- standard math Poincaré–Hopf index theorem
- domain assumption Assumption 2.3: E is a finite union of real-analytic closed discs and ρ is real-analytic on Γ
- domain assumption Zero-separation: the divisors D_{ρ,χ_n} stay away from Γ
Cite this review
Pith. "Pith review of Extremal Functions and Widom Factors on Compact Riemann Surfaces." pith.science (2026). https://pith.science/paper/FQ23ZOBE
@misc{pith2026260721260,
author = {Pith},
title = {Pith review of: Extremal Functions and Widom Factors on Compact Riemann Surfaces},
year = {2026},
howpublished = {\url{https://pith.science/paper/FQ23ZOBE}},
note = {Machine review of arXiv:2607.21260}
}
abstract
We study the Chebyshev extremal problem on a compact Riemann surface $X$ of genus $g>0$. As an analog to monic polynomials, we consider admissible meromorphic functions having no poles away from a marked point $P_{\infty}$ (with adequate normalisation). For a nonpolar compact set $E\subset X\setminus\{P_{\infty}\}$, we show that the $n$-th root of the Chebyshev constant $t_n(E)$ converges to the capacity of $E$, and we establish the corresponding Bernstein-Walsh inequality. In the second part of the paper, using a Cauchy kernel adapted to Riemann surfaces, we study the refined Szeg\H{o}--Widom asymptotics for the extremals as well as the Widom factors $$ W_n(E)=\frac{t_n(E)}{\operatorname{cap}(E)^n},$$ assuming that $E$ is a finite union of $p$ closed discs with analytic boundaries. The geometry of the Schottky double, which has genus $2g+p-1$, enters explicitly into these asymptotics. We find that there is no analogue of Faber-type asymptotics even for one single boundary curve. We conclude by constructing explicit examples in genus $1$ using the Weierstrass-$\wp$ function.
Reviewed August 1, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.