[
  {
    "id": "ARR-2026-597T88D7WZ98HR0Z-v2",
    "type": "article",
    "title": "Canonical and unrestricted observations in a finite checkerboard gradient model",
    "author": [
      {
        "literal": "Lluis Eriksson"
      }
    ],
    "issued": {
      "date-parts": [
        [
          2026,
          11,
          11
        ]
      ]
    },
    "publisher": "AIRR.SCIENCE",
    "archive": "AIRR.SCIENCE",
    "archive_location": "ARR-2026-597T88D7WZ98HR0Z",
    "version": "v2",
    "genre": "Research preprint",
    "URL": "https://airr.science/papers/ARR-2026-597T88D7WZ98HR0Z/versions/v2/",
    "note": "Research preprint; ARR-2026-597T88D7WZ98HR0Z v2. Not peer reviewed by AIRR. Working paper; not admitted to the AIRR accepted collection.",
    "abstract": "We compare unrestricted observations and canonical original-channel rectangle means for a finite two-channel source-weighted checkerboard gradient form. Exact orthogonal restoring ranks and a fixed-positive-angle order-n Fourier construction are retained. We add a constructive every-window corner-matching obstruction for a three-dimensional cyclic product-slice extension. Its nonwrapping product-slice restrictions yield logarithmic necessary counts for fixed or bounded-motion cuts. A rapidly moving cyclic24n-row family fails through a short four-step resonance. Its wrapped complements are not nonwrapping canonical cells, so that failure is not transferred merely by refining means. A separate new cutoff covers the additional chart-boundary corners and constructs a real nonzero actual-coexact field annihilating each of36n-24 nonwrapping refined means, with Q<7c||z||^2/16 at an enormous stated finite threshold. Original counting normalization, source parity, coexact membership, exact zeros and nonzero norm are derived explicitly. The general canonical coexact O(n)-versus-logarithm question remains open. Historical originality and wider quantitative relevance are unestablished; no physical Yang-Mills gap or continuum result follows."
  }
]
