Public note · 30 October 2026 · Nanjie Ma ·
ORCID 0009-0002-4415-1209
About / Work subpage. English only. Not a homepage hero. Not a breakthrough notice. Not cosmogenesis payment.
In an earlier theoretical sketch of Discrete Fragment Cosmology (DFC), the Algebraic Closure Emergence Hypothesis (ACEH), and the Quantum Narrative Matrix (QNM), AI-assisted drafting and concurrent checking wrote unsettled claims in the voice of settled results. Some checks in the course of the work were AI-assisted; the author is responsible for the overclaim.
I withdraw treating that chain as a completed, forward-generated origin of the universe. I withdraw, as established results, the identification of a parent N = 22 with a unique remove-1 landing Neff = 21 / U(21) that generates eighteen cosmological parameters, and the claim of an established cosmic spin-handedness or macroscopic rotation bias.
What remains, as a working hypothesis and not as an established result, is staged. The covering kernel is a written covering-growth object with a history-independent characterization: a unique minimum, Hasse in-degree 1 or 2, and Hasse out-degree at most 2. Covering births are not spacetime events. The honest evolution class from that kernel is covering to a family of complete objects, not a forced selection of one observable cosmos. An interpolating sequential type from that family to a later complete object is not exhibited.
The idea is this. If the universe has an early-seed stage, that stage should be a checkable growth object, and it should not preload spacetime, an action, or simplices onto the seed. The covering kernel is written to that standard. Its advantages are a rigid local degree constraint, a characterization that does not depend on growth history, and the refusal to identify covering births with spacetime events. The author’s working judgment is that, among written early-seed candidates, the covering kernel has the highest potential and is the more likely early seed. That is not an established origin of the universe, and it does not mean that a later complete object has already grown from the kernel.
ACEH / π, e, i remains a programme calibration and encoding spine, not the covering growth law, and is not derived from the kernel. The QNM eighteen-parameter table is aimed at first-principles derivation; at present it remains a frozen spectral readout dictionary, not a completed first-principles result. Named 22 → 21 may be read as a calibration attachment of that dictionary, not as a unique generative stair and not as an established exporter of the eighteen parameters.