Research Article · Preprint Mathematics
A finite-order obstruction to total positivity of an arithmetic theta trace kernel
- Primary subject
- Mathematics
- Subject subfields
- Classical and Real Analysis · Number Theory and Special Functions
- Paper type
- Research Article
- Manuscript stage
- Preprint
- Manuscript
- Archived
- Edition
- 1 · Permanent link
Overview
Abstract
We exhibit an explicit negative sixth-order Toeplitz minor of the infinite arithmetic kernel G(x) = ∑ₙ≥₁ exp(−n²e^(−x)). At the increasing nodes xᵢ = −log(5/2) + i/100 and yⱼ = j/100, 0 ≤ i,j ≤ 5, its determinant lies strictly between −978 × 10^(−59) and −977 × 10^(−59). A portable certificate uses only integer and rational arithmetic and bounds the complete omitted arithmetic tail; an independently constructed ball-arithmetic enclosure verifies the same full-series minor. Positive exponential gauges and translation transfer this counterexample to every kernel C e^(−3x/4) ∑ₘ≠₀ exp(−αm²e^(−x)), C,α > 0, disproving the total-positivity assertion for the explicit trace kernel in arXiv:2602.01248v1. We also give an elementary two-summand counterexample to positive-mixture closure, classify finite positive common-tilt mixtures of Gumbel kernels, and recover an infinite-order obstruction from the Gamma–zeta transform. Separate twelfth-order ordered and derivative certificates supply additional reproducible checks. The arithmetic trace kernel is distinct from the even Riemann density; no conclusion about the Riemann hypothesis follows.
Keywords total positivity · Pólya frequency functions · arithmetic theta trace kernels · Toeplitz minors · exact rational certificates · Gamma–zeta transform
Scope of the work
Full infinite arithmetic trace fails PF6 for every positive C and α in the literal arXiv:2602.01248v1 kernel. Exact whole-tail rational ordered-minor interval and independent directed-ball reconstruction. No smallest failing order/global PF5, original physical-theta sign, off-critical-line zeta zero or RH consequence. Unseen revisions and conditional companion results are outside this correction.
Reproduction notes
Ten-member package includes editable source, exact standard-library sixth-order and additional twelfth-order ordered verifiers, independent sixth-order Arb reconstruction, and a separate twelfth-order derivative verifier. All four pass normally and with Python -O. Optional python-flint 0.9.0; README.txt gives commands and MANIFEST.sha256 covers all other public members.
The manuscript
Read the paper
The complete manuscript, as published in this edition.
If the reader is unavailable on your device, open or download the PDF directly.
Author disclosures
OpenAI Codex assisted with literature discovery, mathematical development and checking, reproducibility code and manuscript preparation. The author is responsible for the content. Internal AI-assisted checks and Archive posting do not constitute external peer review or formal proof verification.
References and related publications
- Watson, primitive Markovian cycles I, arXiv:2602.01248v1 (literal version corrected)
- Watson, primitive Markovian cycles II, arXiv:2602.06080v1
- Belton, Guillot, Khare and Putinar, Preservers of totally positive kernels and Pólya frequency functions
- Michałowski, certified failure for the different de Bruijn–Newman kernel, arXiv:2602.20313v1
A stable scholarly record
Cite this edition
Scott Schubr. (2026). A finite-order obstruction to total positivity of an arithmetic theta trace kernel (Version 1). Luminary Research Archive. https://research.luminaryoddities.com/finite-order-obstruction-total-positivity-arithmetic-theta-trace-kernel/v/1
BibTeX
@techreport{luminary_finite_order_obstruction_total_positivity_arithmetic_theta_trace_kernel_v1,
title = {A finite-order obstruction to total positivity of an arithmetic theta trace kernel},
author = {Scott Schubr},
year = {2026},
institution = {Luminary Research Archive},
type = {Research Article},
note = {Version 1},
url = {https://research.luminaryoddities.com/finite-order-obstruction-total-positivity-arithmetic-theta-trace-kernel/v/1}
}Loading discussion…