# Freeze — EM Units Access No-Go

### Particle geometry cannot mint \((a,\alpha)\) · L1 for the length leftover · \(m_e\) ACCESS · slice-access for \(\alpha\) · Mach outside · F2 stands

**CPT Framework — absolute scale / coupling honesty**  
**Date:** 2026-08-20  
**Status:** **THEORY FREEZE** · `PASS_EM_UNITS_ACCESS_NOGO` · public chips **unchanged**  
**Design / results:** `a_geom_em_units_access_nogo_design.md` · `a_geom_em_units_access_nogo_results.md`  
**Census:** `scripts/investigations/s3_geometry/em_units_access_nogo_census.py` · `data/em_units_access_nogo_census.json`  
**Parents:** Lemma L1 · `paper_absolute_scale_wall.md` · Doctrine F2/F3/F16 · Willmore/α · 4D-ED forcing-out · G1 slice · M3 · AST freeze

Tags: **[E]** · **[E honesty]** · **[E process]** · **[Programme]** · **[M]** · **[O residual]** · **[F overclaim]**

---

## 0. One-line

> **Line in the sand: a 3-slice observer cannot mint the EM units from the particle’s own geometry. Absolute \(a\) is blocked by Lemma L1 (homogeneity of \(1/a\)); \(m_e=\hbar/(ac)\) is ACCESS of that leftover, not a second input. The value of \(\alpha\) is blocked as a *knot/bulk output* because it is a slice-relative coupling and the slice is not a reconstructible invariant of the knot — not because L1 forbids dimensionless numbers, not because a census of 137-formulae ran dry, and not because “higher-D is ineffable.” Mach is a cosmic relation among shadows, not a local mint. \(\alpha\) and \(a\) remain INPUT. No promote.**

---

## 1. Why freeze now

The programme already knew, as a *pattern*, that local geometry derives structure and withholds scale (`paper_one_wall.md`, `paper_absolute_scale_wall.md`). What was missing was a **bounded no-go** that:

- splits the length leftover \(a\) (dimensionful; \(m_e\) ACCESS) from \(\alpha\) (dimensionless),
- says *which* operators a 3-observer can reconstruct,
- forbids selling exhaustion, L1-misuse, or mysticism as the theorem.

That is this freeze. It protects Doctrine F2/F3/F16 as *consequences* of access, not as unfinished homework.

---

## 2. Reconstructible 3-algebra **[E process]**

A 3-observer of a confined photon reconstructs:

| Accessible | Tag |
|------------|-----|
| Rotation invariants \(\{N,a^2\}\) of the quaternionic knot | **[E]** (4D-ED forcing-out) |
| Integer windings / topology / \(\gcd\) | **[E]** (phase-closure lattice) |
| One \(U(1)\); 3D Weber \(\sim k^2\) | **[M/E]** face ED |
| Binding interpolation \(c=1-\lvert\cos\gamma\rvert\) **once the slice is given** | **[E]** geometry |
| Mass *ratios* to \(m_e\); \(3/(2\alpha)\) as a manifestation *signature* | **[E/M]** |
| \(\varepsilon/a\) as a 3D packaging *parameter* | **[M]** |

They do **not** independently reconstruct:

| Not independently accessible | Why |
|------------------------------|-----|
| Slice tilt \(\beta\) | \(\alpha=\tan\beta\) *is* the viewing angle; circular as an output |
| The slice itself | G1 — no knot arithmetic selects it |
| Which imaginary 2-plane is EM | Gauge; second EM slice forced out |
| Bulk radii / \(S^4,S^7\) moduli as meters | Dimensionless on unit spheres (I3, AS16) |
| An independent normal-bundle law for \(\varepsilon\) | \(\varepsilon^2/\alpha\) rhymes track the *hand* \(\varepsilon\) |
| The bulk modulus that would have set \(\alpha\) | 3-books freeze (M3): face \(\neq\) bulk imprint |

**This list is the theorem’s domain.** Claims about “all mathematics” are **K-INEFFABLE**.

---

## 3. Theorem N-ME — local electron mass **[E]**

\[
\boxed{
\begin{aligned}
&\text{Any finite family of energies homogeneous of degree }{-1}\text{ in }a\\
&\text{is invariant under }a\mapsto\lambda a.\text{ A balance among them cancels }a\\
&\text{and constrains only dimensionless combinations (Lemma L1).}\\
&\text{Absolute }m_e=\hbar/(ac)\text{ requires a second length }L_*\\
&\text{not built from }a\text{ or }\lambda_C(m_e).
\end{aligned}
}
\]

**Access corollary.** Every candidate \(L_*\) that is 3-reconstructible is either

1. circular Compton / Bohr / classical radius,
2. cosmic Mach (\(R\) or \(H_0\); soft; **consumes \(\alpha\)**), or
3. a bulk radius the 3-slice does not independently measure.

A 3-observer cannot mint \(a\) from particle data (hence cannot mint \(m_e=\hbar/(ac)\)). They can adopt \(a\) as the geometric unit, read \(m_e\), or relate it to cosmology.

Census algebra: \(C_{\mathrm{sum}}=n+\alpha\) independent of \(\lambda\) **[E]**.

---

## 4. Theorem N-α — fine-structure constant, access form **[E honesty / programme]**

L1 does **not** close \(\alpha\). Homogeneity of \(1/a\) leaves dimensionless couplings on the table (L1(ii)). The no-go is reconstruction, not degree.

\[
\boxed{
\begin{aligned}
&\alpha\text{ is a slice-relative coupling (projection / leak / dipole of the 3-embedding).}\\
&\text{It is not a function of the reconstructible 3-algebra of §2.}\\
&\text{It names the embedding; it is not an output of the knot.}
\end{aligned}
}
\]

Same category as binding: gcd does not tell Hopf from Free; the slice does. Promoting a knot invariant, a unit-sphere volume, or a fitted tube thickness to \(\alpha\) is that error again.

**Three-bin rule** (every *local* \(\alpha\) factory):

| Bin | Meaning | Verdict |
|-----|---------|---------|
| **B1** | Reparametrisation of the slice (\(\tan\beta\); leak; \((k\varepsilon)^2\) with \(\varepsilon\) from \(\alpha\); resonance \(E_{\mathrm{rot}}=\alpha E_{\mathrm{base}}\)) | Circular as a mint |
| **B2** | Unselected bulk invariant (Willmore \(\tan=1\); \(\pi/24\); Haar volumes) | Structure, not a selector of \(\beta\) |
| **B3** | Combinatorial package (lab five-term; demoted \(\sqrt{2\pi^4(\pi^4-1)}\)) | **[M]** or **[D]**; not free theorem |

Known factories are all binned (`a_geom_em_units_access_nogo_results.md`). That ledger is **not** the theorem (**K-CENSUS**).

**What \(\alpha\) *is*, downstairs.** Single integer power; 3D-face reading; load-bearing in the lepton ladder, Mach M13, and the spin-sub-band scale \(j=\sqrt\alpha\,l\). Appearance in many books is a manifestation signature (M3-5), not a derivation.

---

## 5. What stays outside the no-go

### Mach is not swallowed **[E process]**

Mixed-mass Mach / M13 is a **cosmic** \(L_*\). Hubble is 3-accessible in principle. The closure

\[
G=\frac{3\alpha}{8\sqrt{6}}\frac{\hbar^2 H_0}{c\,m_e^2 m_p}
\]

is a percent-class **relation among** \(\{G,H_0,\alpha,m_e,m_p\}\), convention-conditional **[M/P]**. It **consumes** \(\alpha\); it does not emit \(\alpha\); it is not a local particle mint of \(m_e\). Residual bricks (\(\sqrt{6}\), coefficient book, \(R\)-ontology) remain the AST programme. **K-MACH-SWALLOW** held.

### Resonance is a use, not a mint

Richard’s rung \(E_{\mathrm{rot}}(j)=\alpha E_{\mathrm{base}}\) is a 3D-face balance. It correctly puts \(\alpha\) in the occupied spin sub-band. **K-RESONANCE** held.

### Structure continues

Lattice, Willmore stage, spin–statistics, binding-as-slice, ratios, eversion, visibility-as-occupancy — none of these required a local 511 keV or 137 machine. Closing the mint **focuses** that programme (AS5).

---

## 6. Kill board (frozen)

| ID | Forbidden |
|----|-----------|
| **K-CENSUS** | Exhaustion of 137-formulae sold as a theorem |
| **K-L1α** | Lemma L1 sold as a no-go for \(\alpha\) |
| **K-INEFFABLE** | “Higher-D therefore we cannot know” without §2 |
| **K-MACH-SWALLOW** | Mach hidden inside the particle no-go |
| **K-RESONANCE** | \(j=\sqrt\alpha\,l\) as a derivation of \(\alpha\) |
| **K-LABSUM** | Lab geometric sum as free \(\alpha\) theorem |
| **K-EPS** | \(\varepsilon^2/\alpha\) rhyme with hand \(\varepsilon\) as derivation |
| **K-CHIP** | Flip F2 / F3 / F16 from this freeze |
| **K-PROMOTE** | Hub promote without human gate |

---

## 7. Re-open rules (narrow)

| Target | Re-open only if |
|--------|-----------------|
| Local \(m_e\) | A **new** non-circular \(L_*\) not built from \(a\) or \(\lambda_C(m_e)\), surviving L1(iii). Rebrands of the 2026 census fail by protocol (P-AS2). |
| \(\alpha\) from the knot | A **new** 3-reconstructible invariant **not** in §2, independently fixed (no CODATA \(\alpha\), no fitted \(\varepsilon\), no term shopping), that selects \(\beta\) **blind**. |
| \(\alpha\) as eigenvalue | Two **independent** geometric conditions (confinement \(\varepsilon\) vs leak \(f(\varepsilon)\)), neither fed CODATA \(\alpha\). Still does not mint \(m_e\) without \(L_*\). |

Default after this freeze: **do not spend WPs hunting 137 or 511 keV from the knot.**

---

## 8. Doctrine contact

| Chip | After this freeze |
|------|-------------------|
| F2 \(\alpha\) derived | **HELD** — this freeze is why |
| F3 / F16 local \(m_e\) | **HELD** |
| Lab geometric \(\alpha\) sum | **[M]** packaging only |
| Public \(\alpha\), \(a\) | **INPUT** (3-slice units) · \(m_e\) ACCESS |

No chip flip. No promote.

---

## 9. Findings register

| ID | Finding | Tag |
|----|---------|-----|
| **AU1** | Reconstructible 3-algebra of §2 is the domain of the no-go | **[E process]** |
| **AU2** | N-ME: L1 + access corollary; local \(m_e\) mint closed | **[E]** |
| **AU3** | N-α: \(\alpha\) is slice-relative; not a knot/bulk output | **[E honesty]** |
| **AU4** | L1 is not an \(\alpha\) no-go | **[E]** |
| **AU5** | Census of factories is a B1–B3 ledger, not the theorem | **[E honesty]** |
| **AU6** | Mach stays outside; consumes \(\alpha\); does not mint units | **[E process / M]** |
| **AU7** | Resonance / leak / dipole *read* \(\alpha\); they do not mint it | **[E/M]** |
| **AU8** | Public F2/F3/F16 unchanged; no promote | **[E process]** |

---

## 10. One-line for agents

> **\(a\) blocked by L1; \(m_e=\hbar/(ac)\) ACCESS; \(\alpha\) blocked as slice-access on the present 3-algebra (not L1, not census, not mysticism); \(\alpha\) and \(a\) INPUT downstairs; Mach outside; F2 stands.**

N-α is a freeze of the **domain**, plus a factory ledger. It is not a completeness theorem of 3-geometry and not a proof that no dimensionless identity can equal \(\alpha\). Stop hunting 137/511 from the knot.

**Looked into (16 Sep, A-INPUT-RETAG):** leftover length named \(a\) INPUT; \(m_e\) ACCESS; XOR; two leftovers unchanged (`a_geom_p1_a_input_retag.md`).

**Review draft (20 Aug 2026):** `paper_access_completeness_3slice.md` **v0.3** — T-ME / T-α **proved**; C-α uses the reading (A3 scar visible); A1–A5 applied. Not submission-ready. No chip flip. No promote.

**INV catalogue (20 Aug 2026):** `a_geom_pi_catalogue_freeze.md` — positive list of what a 3-slice *should* predict (INV / SHADOW). Complements T-ME/T-α; does not mint leftovers.