Theorem on Critical Purity

Status: [T] Proven

The value $P_{\text{crit}} = 2/N$ is rigorously derived from several mathematically equivalent formulations of a single geometric principle (paths 1–4) and an independent autopoietic argument (path 5). The convergence of all approaches to a single value confirms the fundamentality of this threshold.

1. Theorem Statement

1.1 Main assertion

Theorem (Critical Purity):

For a holonomic system of dimension $N$, the critical purity:

$$P_{\text{crit}} = \frac{2}{N}$$

is the unique value satisfying the following equivalent conditions:

1. Geometric: $\|\Gamma - I_N/N\|_F^2 = \|I_N/N\|_F^2$
2. Informational: $D_{KL}(\Gamma \| I_N/N) = \frac{1}{2}$ nat (in linear approximation)
3. Structural: $|\mathbf{r}|^2 = 2\sigma^2$ (SNR = 1)
4. Spectral: $\lambda_{\max} = (1 + \sqrt{N-1})/N \approx 1/2$
5. Autopoietic: minimal breaking of $U(N)$ symmetry

1.2 For UHM (N = 7)

$$P_{\text{crit}} = \frac{2}{7} \approx 0.286$$

At this threshold:

• Structural deviation = scale of chaos
• Informational contribution = 1/2 nat
• Dominant mode ≈ 49% coherence
• $U(7)$ symmetry broken to distinguishability level

---

2. Necessary Definitions

2.1 Coherence matrix

The coherence matrix $\Gamma \in \mathcal{L}(\mathbb{C}^N)$ satisfies:

$$\Gamma^\dagger = \Gamma, \quad \Gamma \geq 0, \quad \mathrm{Tr}(\Gamma) = 1$$

2.2 Purity

$$P = \mathrm{Tr}(\Gamma^2) \in \left[\frac{1}{N}, 1\right]$$

State	Purity	Description
Pure	P = 1	Γ = |ψ⟩⟨ψ|
Maximally mixed	P = 1/N	Γ = I_N/N (chaos)

2.3 Frobenius norm

$$\|\Gamma\|_F^2 = \mathrm{Tr}(\Gamma^\dagger \Gamma) = \mathrm{Tr}(\Gamma^2) = P$$

---

3. Five Derivation Paths (four equivalent + one independent)

3.1 Path 1: Geometric (structural doubling principle)

Principle: A system is distinguishable from chaos if its deviation from chaos exceeds the scale of chaos.

Criterion:

$$\|\Gamma - I_N/N\|_F^2 > \|I_N/N\|_F^2$$

Left-hand side computation:

$$\begin{aligned} \|\Gamma - I_N/N\|_F^2 &= \mathrm{Tr}\left((\Gamma - I_N/N)^2\right) \\ &= \mathrm{Tr}(\Gamma^2) - \frac{2}{N}\mathrm{Tr}(\Gamma) + \mathrm{Tr}\left(\frac{I_N^2}{N^2}\right) \\ &= P - \frac{2}{N} + \frac{1}{N} \\ &= P - \frac{1}{N} \end{aligned}$$

Right-hand side computation:

$$\|I_N/N\|_F^2 = \mathrm{Tr}\left(\frac{I_N^2}{N^2}\right) = \frac{N}{N^2} = \frac{1}{N}$$

Threshold derivation:

$$P - \frac{1}{N} > \frac{1}{N} \quad \Rightarrow \quad \boxed{P > \frac{2}{N}}$$

Exact reformulation (majority of the Frobenius weight). The decomposition $\Gamma = I_N/N + \Delta$ is orthogonal in the Hilbert–Schmidt inner product ($\langle I_N/N, \Delta\rangle_F = \mathrm{Tr}(\Delta)/N = 0$), so Pythagoras gives the exact split of the state's total Frobenius weight into a noise part and a structural part:

$$P = \|\Gamma\|_F^2 = \underbrace{\|I_N/N\|_F^2}_{1/N\ \text{(noise)}} + \underbrace{\|\Delta\|_F^2}_{\text{structure}}, \qquad s := \frac{\|\Delta\|_F^2}{P}\ \text{(structural share)}.$$

The criterion $\|\Delta\|_F^2 > \|I_N/N\|_F^2$ is then exactly the statement that the structural component holds the strict majority of the state's Frobenius weight:

$$P > \frac{2}{N} \iff s > \frac{1}{2}.$$

Between exactly two orthogonal components, majority is the unique dominance threshold containing no free parameter — any other cut $s > c$ would introduce an arbitrary constant $c$. The apparent "factor 2" is therefore not a chosen constant: it is the arithmetic consequence of the majority criterion, $\tfrac1N + \tfrac1N = \tfrac2N$.

Interpretation: one dominance principle for both thresholds

This is the same plurality-dominance principle that fixes the reflection threshold: for $R_{\mathrm{th}}$ the self-model must dominate each of $K = 3$ competing hypotheses (plurality $\Rightarrow 1/3$); for $P_{\mathrm{crit}}$ the structural component must dominate the single competing component, noise ($K = 2$ orthogonal parts $\Rightarrow$ majority $1/2$ of the weight $\Rightarrow P > 2/N$). One principle, two thresholds.

---

3.2 Path 2: Information-theoretic

Principle: A system carries sufficient information for distinguishability if its divergence from chaos exceeds an information quantum.

Kullback–Leibler divergence:

$$D_{KL}(\Gamma \| I_N/N) = \mathrm{Tr}(\Gamma \log \Gamma) - \mathrm{Tr}\left(\Gamma \log \frac{I_N}{N}\right)$$

Using $\log(I_N/N) = -\log(N) \cdot I_N$:

$$D_{KL}(\Gamma \| I_N/N) = -S_{vN}(\Gamma) + \log(N)$$

where $S_{vN}(\Gamma) = -\mathrm{Tr}(\Gamma \log \Gamma)$ is the von Neumann entropy.

Expansion for states close to $I_N/N$:

For $\Gamma = I_N/N + \delta\Gamma$ with small $\delta\Gamma$:

$$D_{KL}(\Gamma \| I_N/N) \approx \frac{N}{2} \cdot \mathrm{Tr}(\delta\Gamma^2) = \frac{N}{2} \cdot \left(P - \frac{1}{N}\right)$$

Minimum distinguishability:

The distinguishability threshold in the quadratic approximation = $\frac{1}{2}$ nat.

$$\frac{N}{2} \cdot \left(P - \frac{1}{N}\right) \geq \frac{1}{2}$$ $$P - \frac{1}{N} \geq \frac{1}{N} \quad \Rightarrow \quad \boxed{P \geq \frac{2}{N}}$$

Scope of applicability

Path 2 uses the quadratic approximation D_KL(Γ ‖ I/N) ≈ (N/2)(P − 1/N), valid when P − 1/N ≪ 1. The threshold D_KL = 1/2 nat is a convention (analogous to p-value 0.05 in statistics). Path 2 is a supporting argument, consistent with P_crit = 2/N, not an independent rigorous derivation — and the audit below makes precise how much weight it can bear.

How far Path 2 actually is from exact (measured 2026-08-06)

Two facts, both machine-verified:

1. The threshold sits exactly where the approximation fails. At $P = 2/N$ we have $P - 1/N = 1/7 \approx 0.143$, which is not $\ll 1$.
2. $D_{\mathrm{KL}}(\Gamma\,\|\,I/N)$ is not a function of purity alone. Over all states with $P = 2/7$ the exact divergence spans $[0.359,\,0.622]$ nat — $0.344$ for one dominant mode plus six equal, $0.437$ for two dominant plus five — against the approximation's $0.500$. So the correspondence "$D_{\mathrm{KL}} = 1/2 \iff P = 2/N$" holds for some spectra and not others; it is not an equivalence.

By contrast Path 1 is exact and state-independent: $\|\Gamma - I/N\|_F^2 = P - 1/N$ identically (verified to $10^{-16}$), and $\|I/N\|_F^2 = 1/N$, so the majority criterion is $P > 2/N$ with no approximation anywhere. Path 3 is likewise exact — the Weingarten constant $1/(N(N+1)) = 1/56$ multiplies both sides and cancels. Read the "convergence of five paths" accordingly: two are proofs, the rest are corroboration.

:::

Interpretation for engineers

Information threshold: The system must carry at least 1/2 nat of information beyond maximum entropy. This is a fundamental distinguishability limit in information theory.

In practice: At $P = 2/N$ the structural information is one binary distinction — "structure exists" versus "no structure". Note that this is a qualitative reading, not a numerical one: the convention $D_{\mathrm{KL}} = 1/2$ nat equals $0.5/\ln 2 \approx 0.72$ bits, not $1$ bit. Demanding exactly one bit would set the threshold at $\ln 2$ nat and give $P = (1 + 2\ln 2)/N \approx 0.341$, not $2/N \approx 0.286$.

---

3.3 Path 3: Helstrom / Haar single-shot detection

Principle: A Haar-random single-shot measurement on $\Gamma$ produces a statistically detectable deviation from the noise reference $I/N$ iff $P > 2/N$.

Setup. Let $\Pi = |\psi\rangle\langle\psi|$ with $|\psi\rangle$ Haar-uniform on the unit sphere of $\mathbb{C}^N$. For a self-adjoint $A$, the $\Pi$-induced observable is $\mathrm{Tr}(A\Pi)$.

First-moment (Haar invariance). $\mathbb E_\Pi[\Pi] = I/N$ (unitary invariance), hence $\mathbb E_\Pi[\mathrm{Tr}(A\Pi)] = \mathrm{Tr}(A)/N$.

Second-moment (Weingarten). The standard $U(N)$-Weingarten formula gives $\mathbb E_\Pi[\Pi\otimes\Pi] = \frac{1}{N(N+1)}(I + \mathrm{SWAP}).$ For $N=7$: $\mathbb E_\Pi[\Pi\otimes\Pi] = (I+\mathrm{SWAP})/56$. Hence $\mathbb E_\Pi[\mathrm{Tr}(A\Pi)^2] = \mathrm{Tr}((A\otimes A)\cdot\mathbb E[\Pi\otimes\Pi]) = \frac{1}{N(N+1)}(\mathrm{Tr}(A)^2 + \|A\|_F^2).$

Variance formula. $\mathrm{Var}_\Pi(\mathrm{Tr}(A\Pi)) = \mathbb E[\mathrm{Tr}(A\Pi)^2] - \mathbb E[\mathrm{Tr}(A\Pi)]^2 = \frac{\|A\|_F^2}{N(N+1)} - \frac{\mathrm{Tr}(A)^2}{N^2(N+1)}$

(using $\tfrac{1}{N(N+1)} - \tfrac{1}{N^2} = -\tfrac{1}{N^2(N+1)}$; verified numerically against Haar sampling).

Applied to $A = \Delta = \Gamma - I/N$ (traceless, $\mathrm{Tr}(\Delta)=0$): $\mathrm{Var}_\Pi(\mathrm{Tr}(\Delta\Pi)) = \frac{\|\Delta\|_F^2}{N(N+1)} = \frac{P - 1/N}{N(N+1)}.$

Detection threshold. The observer's expected single-shot quadratic detection signal (above the zero-signal noise baseline) exceeds the reference scale $\|I/N\|_F^2/(N(N+1)) = 1/(N^2(N+1))$ iff $\|\Delta\|_F^2 > \|I/N\|_F^2 \iff P > 2/N.$

$\boxed{P > \frac{2}{N}}$

Interpretation: operational realization of the majority criterion

The Weingarten computation of the variance is exact [T], and the Haar constant $1/(N(N+1)) = 1/56$ (for $N=7$) multiplies both sides of the comparison and drops out — so the threshold is basis-independent and realizable by a physical Haar-random single-shot probe. What Path 3 adds is thus operational content: the Frobenius-majority criterion of Path 1 is not merely geometric bookkeeping but the detectability condition of an implementable random-basis measurement. The comparison scale itself ($\|I/N\|_F^2$, i.e. the majority criterion) is the same dominance principle as in Path 1 — Path 3 realizes it operationally rather than forcing the constant independently.

---

3.4 Path 4: Spectral condition (characterization, not an independent derivation)

Principle: For an identity to exist, the system must have a dominant mode.

Spectrum of $\Gamma$:

$$\mathrm{Spectrum}(\Gamma) = \{\lambda_1, \lambda_2, \ldots, \lambda_N\}, \quad \lambda_1 \geq \lambda_2 \geq \cdots \geq \lambda_N \geq 0$$

With constraints:

$$\sum_i \lambda_i = 1, \quad \sum_i \lambda_i^2 = P$$

Optimization problem:

Find the maximum $\lambda_1$ for a given $P$:

$$\max \lambda_1 \quad \text{subject to} \quad \sum_i \lambda_i = 1, \quad \sum_i \lambda_i^2 = P, \quad \lambda_i \geq 0$$

Solution (Lagrange method):

By symmetry, the optimum is reached when $\lambda_2 = \lambda_3 = \cdots = \lambda_N = \lambda$:

$$\lambda_1 + (N-1)\lambda = 1 \quad \Rightarrow \quad \lambda = \frac{1 - \lambda_1}{N-1}$$ $$\lambda_1^2 + (N-1)\lambda^2 = P$$

Substituting:

$$\lambda_1^2 + \frac{(1 - \lambda_1)^2}{N-1} = P$$

Solving the quadratic equation:

$$\lambda_1 = \frac{1 + \sqrt{(N-1)(NP - 1)}}{N}$$

At $P = 2/N$:

$$\lambda_{\max} = \frac{1 + \sqrt{(N-1)(2 - 1)}}{N} = \frac{1 + \sqrt{N-1}}{N}$$

For $N = 7$:

$$\lambda_{\max} = \frac{1 + \sqrt{6}}{7} \approx 0.493 \approx \frac{1}{2}$$

Interpretation

~50% dominance threshold: At $P = 2/N$ the dominant mode captures approximately half of the coherence. This is the 1:1 threshold — structure is barely distinguishable from the uniform distribution.

Spectral structure at $P = 2/7$:

• $\lambda_1 \approx 0.493$ (49.3% of coherence)
• $\lambda_2 = \cdots = \lambda_7 \approx 0.085$ (8.5% each)

---

3.5 Path 5: Symmetry breaking ($U(N)$ stabilizer)

Principle: Sufficient structure is required for self-modeling $\varphi(\Gamma)$: chaos $I/N$ has maximal symmetry and admits no preferred direction; structure exists only when the symmetry is broken non-trivially.

Stabilizer group. For $\Gamma \in \mathcal D(\mathbb C^N)$: $\mathrm{Stab}(\Gamma) = \{U \in U(N) : U\Gamma U^\dagger = \Gamma\}.$

Lemma (Schur's lemma applied to $I/N$). $\mathrm{Stab}(\Gamma) = U(N)$ iff $\Gamma = I/N$.

Proof. $I/N$ is scalar, hence commutes with every $U$. Conversely, if $\Gamma$ commutes with every $U \in U(N)$, then $\Gamma$ lies in the commutant of the standard $U(N)$-action on $\mathbb C^N$; since this action is irreducible, Schur gives $\Gamma \in \mathbb C \cdot I$; trace-1 forces $\Gamma = I/N$. $\square$

Stabilizer dimension bound. If $\Gamma$ has $k$ distinct eigenvalues with multiplicities $m_1,\ldots,m_k$ (with $\sum m_i = N$), then $\mathrm{Stab}(\Gamma) = U(m_1)\times\cdots\times U(m_k)$, real Lie dimension $\sum m_i^2$. For $k=1$ this is $N^2$ (the $I/N$ case). For $k\ge 2$ the maximum is attained at the most unequal split $(1, N-1)$, giving $1 + (N-1)^2 = N^2 - 2N + 2 < N^2$. For $N = 7$: max non-constant stabilizer dimension is $37 < 49$.

Strengthened symmetry-breaking criterion. Mere inequality $\mathrm{Stab}(\Gamma)\subsetneq U(N)$ is equivalent to $\|\Delta\|_F > 0$, which is satisfied for any $\Gamma \ne I/N$ (arbitrarily small breaking). The strengthened criterion requires that the traceless component dominate the scalar reference: $\|\Gamma - I_N/N\|_F \ge \|I_N/N\|_F.$ By Path 1 this is equivalent to: $\boxed{P \ge \frac{2}{N}}.$

Dependence on Path 1 — clarified

The strengthened criterion $\|\Delta\|_F \ge \|I/N\|_F$ coincides with Path 1 at the algebraic level. Path 5's independent content is the representation-theoretic statement that $I/N$ is the unique $U(N)$-symmetric density matrix (Schur's lemma on the irreducible fundamental representation), making $I/N$ the canonical "maximally symmetric" reference. This is what justifies the choice of reference used in Path 1 — without it, the critical purity would depend on an arbitrary reference state.

---

3.6 Path 6: Octonionic norm [I]

Interpretation [I]

In the octonionic interpretation, purity $P = \mathrm{Tr}(\Gamma^2)$ is connected to the norm on $\mathrm{Im}(\mathbb{O})$. The normativity $|xy| = |x||y|$ ensures a multiplicative metric. The threshold $P_{\text{crit}} = 2/7$ can be interpreted as the minimum norm of a vector in $\mathrm{Im}(\mathbb{O})$ ≅ ℝ⁷ at which its projection onto structural directions (Fano triplets) exceeds the noise projection.

Status: [T], bridge [T] (closed, T15). Compatible with the other five paths. See structural derivation.

---

4. Convergence of All Paths

4.1 Results table

Path	Principle	Main tool	Result
1. Geometric	Frobenius structural dominance	HS Pythagoras	P > 2/N ✓
2. Informational	Relative entropy 2nd-order	Operator Taylor of $\log$	P = 2/N at D=1/2 nat ✓
3. Single-shot detection	Haar-averaged observable variance	Weingarten 2nd moment ($1/(N(N+1))$)	P > 2/N ✓
4. Spectral	Dominant eigenvalue optimum	Lagrange multipliers	λ_max = (1+√(N−1))/N at P = 2/N ✓
5. Symmetry breaking	Stabilizer dimension	Schur's lemma + Cauchy-Schwarz	P > 2/N ✓

4.2 Uniqueness theorem

Theorem: The value $P_{\text{crit}} = 2/N$ is the unique one at which all five criteria coincide.

Proof: Uniqueness follows from the algebraic equivalence of conditions 1–4 (all express the same geometric requirement in different terms). The autopoietic criterion (5) yields the same threshold from an independent symmetry-breaking requirement. All five formulations lead to $P - 1/N = 1/N$. ∎

Logical structure of the five paths

The derivation has one load-bearing core and four structural supports, each with a precisely delimited role:

Core (the derivation proper). $P_{\mathrm{crit}} = 2/N$ follows from two ingredients, both parameter-free:

Ingredient	Content	Status
Reference canonicity (Path 5)	$I/N$ is the unique $U(N)$-invariant density matrix (Schur's lemma on the irreducible fundamental representation) — the only canonical "chaos" reference	[T]
Majority criterion (Path 1)	With the orthogonal split $P = \|I/N\|_F^2 + \|\Delta\|_F^2$, distinguishability = the structural component holds the strict majority of the Frobenius weight, $s > 1/2$ — the unique dominance threshold between two orthogonal components containing no free constant	[T] equivalence; the majority principle is the same plurality-dominance rule that fixes $R_{\mathrm{th}} = 1/3$

Together: $s > \tfrac12 \iff P > 2/N$, with no adjustable parameter anywhere.

Supports (delimited roles, per their own sections):

Path	Role	Relation to the core
3 (Haar detection)	Operational realization [T]: exact Weingarten variance; the Haar constant drops out of the comparison, so the majority criterion is basis-independent and physically measurable by a random single-shot probe	Realizes the Path-1 criterion operationally; does not force the constant independently
4 (Spectral)	Characterization [T] (per §3.4: not an independent derivation): at $P = 2/N$ the extremal spectrum has $\lambda_{\max} = (1+\sqrt{N-1})/N \approx 1/2$	Describes what the threshold state looks like spectrally
2 (KL entropy)	Interpretive confirmation: reduces to Path 1 in the quadratic approximation; the $D_{KL} = 1/2$ nat cut is a convention	Information-theoretic reading of the same inequality

The theorem-level content is therefore: given the canonical reference (Schur) and the parameter-free majority criterion, $P_{\mathrm{crit}} = 2/N$ is forced [T] — with operational measurability (Path 3), spectral shape (Path 4), and informational meaning (Path 2) established as corollaries, not as independent forcings.

---

5. Spectral Characterization

5.1 Optimal spectrum at the boundary

Theorem (Spectrum at $P = P_{\text{crit}}$):

At $P = 2/N$, the optimal spectrum (maximizing $\lambda_{\max}$) has the form:

$$\begin{aligned} \lambda_1 &= \frac{1 + \sqrt{N-1}}{N} \\ \lambda_2 = \lambda_3 = \cdots = \lambda_N &= \frac{N - 1 - \sqrt{N-1}}{N(N-1)} \end{aligned}$$

5.2 Numerical values

N	P_crit = 2/N	λ_max at P_crit
2	1.000	1.000
3	0.667	0.789
4	0.500	0.683
5	0.400	0.618
6	0.333	0.573
7	0.286	0.493
8	0.250	0.457

5.3 Verification for N = 7

$$\lambda_1 = \frac{1 + \sqrt{6}}{7} \approx 0.493$$ $$\lambda_2 = \cdots = \lambda_7 = \frac{6 - \sqrt{6}}{42} \approx 0.085$$

Verification:

$$\lambda_1 + 6\lambda_2 = 0.493 + 6 \times 0.085 = 1.000 \quad \checkmark$$ $$\lambda_1^2 + 6\lambda_2^2 = 0.243 + 6 \times 0.0072 = 0.286 = \frac{2}{7} \quad \checkmark$$

---

6. Hierarchy of Purity Thresholds

6.1 Full hierarchy

$$P_{\text{crit}}^{\text{regen}} < P_{\text{crit}}^{\text{geom}} < P_{\text{safe}} < P_{\text{target}}$$

Threshold	Formula	Value (N=7)	Purpose
P_crit^regen	γ/(κ_rate · Coh_E^min)	≈ 0.033	Dynamical (κ > γ)
P_crit^geom	2/N	≈ 0.286	Structural (main)
P_safe	P_crit^geom + margin	0.30	Operational (with margin)
P_target	—	0.50	Recommended

6.2 Interpretation

• $P_{\text{crit}}^{\text{regen}} \approx 0.033$: Minimum for regeneration to exceed dissipation
• $P_{\text{crit}}^{\text{geom}} = 2/7 \approx 0.286$: Minimum for structural distinguishability from chaos (main threshold)
• $P_{\text{safe}} = 0.30$: Operational threshold with 5% margin
• $P_{\text{target}} = 0.50$: Recommended operating point

Important

A system with $P_{\text{crit}}^{\text{regen}} < P < P_{\text{crit}}^{\text{geom}}$ can regenerate, but has no structural identity — it is indistinguishable from noise.

---

7. Practical Applications

7.1 For AI systems engineers

Viability criterion:

/// Viability check: P > P_crit = 2/N (T-39a [T]).
public pure fn is_viable<const N: Int>(gamma: &StaticMatrix) -> Bool
    where requires N >= 2
{
    let p = (gamma.matmul(&gamma)).trace().real();
    p > 2.0 / (N as Float)
}

Structural deviation computation:

/// Structural deviation ‖Γ − I/N‖_F² = P − 1/N.
///
/// **Interpretation**:
/// - deviation < 1/N: indistinguishable from noise
/// - deviation = 1/N: viability boundary
/// - deviation > 1/N: structured system
public pure fn structural_deviation<const N: Int>(gamma: &StaticMatrix) -> Float
    where requires N >= 2
{
    let p = (gamma.matmul(&gamma)).trace().real();
    p - 1.0 / (N as Float)
}

Dominance threshold:

/// Dominant eigenvalue threshold λ_max at P = P_crit = 2/N.
///
/// For N = 7: returns ≈ 0.493.
public pure fn dominant_eigenvalue_threshold(n: Int { self >= 2 }) -> Float {
    (1.0 + ((n - 1) as Float).sqrt()) / (n as Float)
}

7.2 For consciousness researchers

Connection with interiority levels:

Level	Condition	Interpretation
L0 (Interiority)	ρ_E ≠ 0	Inner state exists
L1 (Phenomenal geometry)	rank(ρ_E) > 1	Structure of qualities
Viability	P > 2/7	Distinguishability from chaos
L2 (Cognitive qualia)	R ≥ 1/3, Φ ≥ 1, D_diff ≥ 2*	Reflexive access

*$D_{\text{diff}}$ requires tensor structure; in the minimal 7D formalism $C_{\min} = \Phi \times R$ is used — see dimension-e.md.

Key conclusion: $P > 2/N$ is a necessary condition for L1 and L2. Without structural distinguishability, phenomenology is impossible.

7.3 For physicists

Analogies with phase transitions:

UHM	Statistical physics	Meaning
P = 2/N	Critical temperature T_c	Ordering threshold
P − 1/N	Order parameter	Measure of structure
λ_max ≈ 1/2	Macroscopic occupancy	Condensation into one mode

Entropic interpretation:

At $P = 2/N$:

$$S_{vN} = \log N - \frac{N}{2}\left(\frac{2}{N} - \frac{1}{N}\right) + O\left(\frac{1}{N^2}\right) = \log N - \frac{1}{2}$$

The system contains 1/2 nat less entropy than maximal chaos.

7.4 For information theorists

Channel capacity:

Distinguishing state $\Gamma$ from $I_N/N$ is equivalent to transmitting information over a channel with capacity:

$$C = D_{KL}(\Gamma \| I_N/N) \approx \frac{N}{2}(P - 1/N)$$

At $P = 2/N$: $C = 1/2$ nat = distinguishability boundary.

Holevo bound:

$$\chi(\{p_i, \rho_i\}) \leq S(\bar{\rho}) - \sum_i p_i S(\rho_i)$$

To distinguish $\Gamma$ from $I_N/N$ one needs $\chi \geq 1/2$ nat, which requires $P \geq 2/N$.

---

8. Universality of the Factor 2

8.1 Appearance in various contexts

Context	Formula	Interpretation
Detection theory	SNR = 1	Signal = noise
Quantum distinguishability	F(ρ, σ) = 1/2	Distinguishability limit
Information theory	ΔS = k ln 2	One bit of information
Statistics	2σ rule	Significant deviation
UHM	P = 2/N	Structure = chaos

8.2 Physical meaning

The factor of 2 is the arithmetic footprint of the majority criterion: with the orthogonal split $P = \|I/N\|_F^2 + \|\Delta\|_F^2$, the boundary is exactly

$$s = \frac{\|\Delta\|_F^2}{P} = \frac{1}{2} \quad\text{(structure holds half the total weight)},$$

equivalently $\|\text{deviation}\|^2 = \|\text{baseline}\|^2$, i.e.

$$P = 2 \times P_{\min} = \frac{2}{N}.$$

The "2" is thus not a tuned constant but the denominator of the parameter-free majority threshold $1/2$ — the $K=2$ instance of the same plurality-dominance principle whose $K=3$ instance gives $R_{\mathrm{th}} = 1/3$.

---

9. Conclusion

9.1 Main result

Theorem (formulation)

For a holonomic system of dimension $N$:

$$P_{\text{crit}} = \frac{2}{N}$$

This value is unique, at which:

• All five formulations of the criterion coincide (4 mathematically equivalent + 1 autopoietic)
• The factor of 2 arises naturally (signal = noise)
• The dominant mode captures ~50% of coherence

9.2 For N = 7 (UHM)

$$P_{\text{crit}} = \frac{2}{7} \approx 0.286$$

Spectral structure at the boundary:

• $\lambda_1 \approx 0.493$ (~50%)
• $\lambda_2 = \cdots = \lambda_7 \approx 0.085$ (~8.5% each)

9.3 Methodological significance

1. Convergence of independent paths confirms the fundamentality of the threshold
2. Factor 2 — universal distinguishability threshold in information systems
3. Spectral characterization connects purity with mode dominance

---

Appendix A: Complete Computations

A.1 Derivation of λ_max at P = 2/N

Problem:

$$\max \lambda_1 \quad \text{subject to} \quad \sum_{i=1}^N \lambda_i = 1, \quad \sum_{i=1}^N \lambda_i^2 = \frac{2}{N}$$

Lagrangian:

$$\mathcal{L} = \lambda_1 - \mu\left(\sum_i \lambda_i - 1\right) - \nu\left(\sum_i \lambda_i^2 - \frac{2}{N}\right)$$

Optimality conditions:

$$\frac{\partial \mathcal{L}}{\partial \lambda_1} = 1 - \mu - 2\nu\lambda_1 = 0$$ $$\frac{\partial \mathcal{L}}{\partial \lambda_k} = -\mu - 2\nu\lambda_k = 0 \quad (k = 2, \ldots, N)$$

From the second condition: $\lambda_2 = \cdots = \lambda_N = -\mu/(2\nu) = \lambda$.

Substituting into the constraints:

$$\lambda_1 + (N-1)\lambda = 1$$ $$\lambda_1^2 + (N-1)\lambda^2 = \frac{2}{N}$$

From the first: $\lambda = (1 - \lambda_1)/(N-1)$.

Substituting into the second:

$$\lambda_1^2 + \frac{(1 - \lambda_1)^2}{N-1} = \frac{2}{N}$$ $$(N-1)\lambda_1^2 + (1 - \lambda_1)^2 = \frac{2(N-1)}{N}$$ $$N\lambda_1^2 - 2\lambda_1 + 1 = \frac{2(N-1)}{N}$$ $$N^2\lambda_1^2 - 2N\lambda_1 + N - 2(N-1) = 0$$ $$N^2\lambda_1^2 - 2N\lambda_1 + 2 - N = 0$$

By the quadratic formula:

$$\lambda_1 = \frac{2N \pm \sqrt{4N^2 - 4N^2(2-N)}}{2N^2} = \frac{2N \pm 2N\sqrt{N-1}}{2N^2} = \frac{1 \pm \sqrt{N-1}}{N}$$

Taking $+$ (maximum):

$$\boxed{\lambda_{\max} = \frac{1 + \sqrt{N-1}}{N}}$$

---

Related documents:

• Viability — application of the theorem
• Axiom of Septicity — axiom context
• Coherence matrix — definition of Γ
• 7D Minimality theorem — why N = 7
• Interiority hierarchy — levels L0 → L2