The Assumption Package of the Yang–Mills Construction: Formal Definitions and Derivations
Year: 2026
Type: preprint
Abstract: This document is the supplementary assumption package for the conditional mass-gap theorem for four-dimensional pure Yang–Mills theory. It formally states, defines and derives the complete set of external hypotheses (RG1–RG5, UV1, PH1–PH3, AN1, Y1–Y4, B1, C1) used in the main construction from Wilson lattice gauge theory. A volume-consistent revised condition RG4' is proved, removing a potential volume-divergence issue in the original statement. Several identities are formally cross-checked against the Lean 4 verification repository of the OpenAI Yang–Mills construction (2026-10-28). Crucially, the full collection of assumptions presented here is still insufficient to prove the mass gap. Explicit counterexample families satisfy all listed hypotheses with zero mass gap. Three further open mathematical ingredients are required to establish the mass-gap conclusion: a volume-uniform log-Sobolev bound, a connecting theorem from Euclidean to Hamiltonian quantum theory, and a nonperturbative Callan–Symanzik Ward identity for mass generation. The paper was compiled with the assistance of the Kimi AI assistant (Moonshot AI) from the Yang--Mills Shepherding Committee validation document, the OpenAI Lean submission, and the review-and-supplement packages; the mathematics was verified against the Lean sources, the independent repository yang-mills-group/YM, and the cited literature. The underlying formal verification uses Lean 4.32.0 and the Comparator tool (records of 2026-10-30, 11-03, 11-05 runs). AI assistance was used for text compilation, translation, and cross-checking; all theorem statements and proofs were checked against the cited certificates and sources. See the acknowledgements and Appendices A--E for the full audit trail.
Author: Zeng, Chunping
Language: English
Subfield: Nuclear and High Energy Physics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7236499852
