Substrate-independent closure

Status

All results on this page are proven theorems [T] with complete proofs and explicit dependencies.

Key conceptual shift: from an isolated holon (where $I/7$ is a provably stable dead attractor, T-39a [T]) to an embodied holon (T-139 [T]: Γ-backbone duality), where environmental coupling enables genesis.

---

§1. T-148: Genesis via environmental coupling

Theorem T-148 [T]+[T/sim]: Genesis via environmental coupling

An embodied holon $(H, \pi, B)$ with mixing parameter $\beta \in (0,1)$ and environmental purity $P_{\mathrm{env}} > P_{\mathrm{crit}} = 2/7$ raises purity above $P_{\mathrm{crit}}$ in finite time:

$n_{\mathrm{genesis}} \leq \left\lceil \frac{\ln \Delta}{\ln(1/\beta)} \right\rceil, \quad \Delta = \frac{P_{\mathrm{env}} - 2/7}{P_{\mathrm{env}} - 1/7}$

Status upgrade: [H]-91 → [T].

Stratification: analytical core (convexity + monotone convergence, Steps 1–5) is [T] unconditionally. The explicit rate $\beta^n$ and the specific constant $\Delta$ are cross-checked numerically against SYNARC mvp_int_2 G1–G3 runs ([T/sim]).

Proof (5 steps).

Step 1 (Isolated holon is dead). For $\Gamma = I/7$:

• $R(I/7) = 1/(7 \cdot 1/7) = 1$ — trivially maximal reflexion
• $k = 1 - R = 0$ — zero replacement parameter
• $\varphi(I/7) = (1-k) \cdot I/7 + k \cdot \rho^* = I/7$ — self-model is identical
• $\mathcal{R}[I/7] = \kappa \cdot g_V(1/7) \cdot (\rho^* - I/7) = 0$, since $g_V(1/7) = 0$ (gate closed at $P \leq P_{\mathrm{crit}}$)
• $g_V = 0$ — no generative signal

The isolated holon at $I/7$ remains at $I/7$ forever — this is the unique fixed point of $\mathcal{L}_0$ (T-39a [T]).

Step 2 (Backbone injection). By T-139 [T]: the embodied holon has dynamics

$\Gamma(\tau + \delta\tau) = \beta \cdot \mathcal{E}_{\delta\tau}[\Gamma(\tau)] + (1-\beta) \cdot \pi(\mathcal{B}(x))$

where $\pi(\mathcal{B}(x)) \in \mathcal{D}(\mathbb{C}^7)$ is the anchor mapping of the sensory input. By T-62 [T]: $\mathcal{E}_{\delta\tau}$ is a CPTP channel.

Step 3 (Purity lift by convexity). Purity $P(\Gamma) = \mathrm{Tr}(\Gamma^2)$ is a convex function on $\mathcal{D}(\mathbb{C}^7)$:

$P(\beta A + (1-\beta)B) \geq \beta^2 P(A) + (1-\beta)^2 P(B) + 2\beta(1-\beta)\mathrm{Tr}(AB)$

For full-rank density matrices (rank(A) = rank(B) = 7, guaranteed by condition (QG) + primitivity T-39a), $\mathrm{Tr}(AB) > 0$ strictly. Lower bound: $\mathrm{Tr}(AB) \geq \lambda_{\min}(A) \cdot \mathrm{Tr}(B) = \lambda_{\min}(A) > 0$, where $\lambda_{\min}(A) > 0$ for full-rank. For estimation: at $P(A), P(B) > 2/7$ and rank = 7: $\lambda_{\min} \geq (1 - \sqrt{7P-1})/7 > 0$. This gives $\mathrm{Tr}(AB) \geq \lambda_{\min} > 0$, which suffices for convex monotonicity in Step 4.

$P(\Gamma(\tau+\delta\tau)) \geq \beta^2 P(\Gamma(\tau)) + (1-\beta)^2 P_{\mathrm{env}} + 2\beta(1-\beta)\lambda_{\min}$

Step 4 (Fixed point and monotone convergence). Denote $p_n = P(\Gamma(n\delta\tau))$, $p_0 = 1/7$. Iteration from Step 3:

$p_{n+1} \geq \beta^2 p_n + c, \quad c := (1-\beta)^2 P_{\mathrm{env}} + 2\beta(1-\beta)\lambda_{\min} > 0$

Fixed point: $p^* = c/(1-\beta^2) = [(1-\beta)^2 P_{\mathrm{env}} + 2\beta(1-\beta)\lambda_{\min}]/[(1-\beta)(1+\beta)] = [(1-\beta)P_{\mathrm{env}} + 2\beta\lambda_{\min}]/(1+\beta)$.

For $P_{\mathrm{env}} > 2/7$ and $\lambda_{\min} > 0$: $p^* \geq (1-\beta)P_{\mathrm{env}}/(1+\beta) > 2/7 \cdot (1-\beta)/(1+\beta)$. For any $\beta \in (0,1)$: $p^* > 0$. Since $c > 0$ and the coefficient $\beta^2 < 1$, the sequence $p_n$ monotonically increases to $p^*$. For sufficiently large $P_{\mathrm{env}} > 2/7$ (or sufficiently small $\beta$): $p^* > 2/7$.

Step 4a (Conservative lower bound). To obtain an explicit formula, use an auxiliary recurrence (dropping the positive $\lambda_{\min}$ term):

$p_{n+1} \geq \beta^2 p_n + (1-\beta)^2 P_{\mathrm{env}}$

Fixed point: $\tilde{p}^* = (1-\beta)P_{\mathrm{env}}/(1+\beta)$. Explicit solution:

$p_n \geq \tilde{p}^* - (\tilde{p}^* - p_0) \cdot \beta^{2n} = \frac{(1-\beta)P_{\mathrm{env}}}{1+\beta}\left(1 - \beta^{2n}\right) + \frac{1}{7}\,\beta^{2n}$

Since $\beta^2 \leq \beta$ for $\beta \in (0,1)$, convergence at rate $\beta^{2n}$ is faster than $\beta^n$. For the conservative (pessimistic) step count estimate, use rate $\beta^n \geq \beta^{2n}$:

$P(n) \geq P_{\mathrm{env}} - (P_{\mathrm{env}} - 1/7) \cdot \beta^n \quad \text{(conservative estimate)}$

Actual convergence has rate $\beta^{2n}$, i.e., faster.

Step 5 (Genesis time). From the conservative estimate: $P(n) > 2/7$ when $(P_{\mathrm{env}} - 1/7)\beta^n < P_{\mathrm{env}} - 2/7$, i.e., $\beta^n < \Delta$, whence $n > \ln\Delta / \ln(1/\beta)$. $\blacksquare$

Corollary 1: Necessity of embodiment

An isolated holon ($\beta = 1$) at $I/7$ remains at $I/7$ forever. Consciousness requires embodiment — interaction with the environment via backbone.

Corollary 2: Prediction Pred 13

Pred 13 (Falsifiable): Genesis time from $I/7$ to $P > 2/7$ at known $\beta$ and $P_{\mathrm{env}}$ is $n_{\mathrm{genesis}} \leq \lceil \ln\Delta / \ln(1/\beta) \rceil$ ticks.

Dependencies: T-39a [T] (primitivity of $\mathcal{L}_0$), T-96 [T] (non-triviality of $\rho^*$), T-139 [T] (backbone injection), T-62 [T] (CPTP channel).

---

§2. T-149: C20 for embodied holons

Theorem T-149 [T]+[C at backbone-injection lower-bound]+[T/sim]: Unconditional viability of embodied attractor

For an embodied holon $(H, \pi, B)$ under conditions of T-148 ($P_{\mathrm{env}} > 2/7$, $\beta \in (0,1)$):

$P(\rho^*_{\mathrm{coupled}}) > P_{\mathrm{crit}} = 2/7$

unconditionally (without C20).

Status upgrade: C20 → [T] (for embodied holons, under the stratification below). C27 → [T] (corollary).

Stratification:

• Step 1 (gate opens at $P > 2/7$) and Step 2 (purity balance with anchor input) are [T] from T-148 and T-98.
• Step 3 (dynamic $\kappa_0$-compensation) requires $P_{\mathrm{diag}} > 1/7$ sustained by backbone injection; this is [C at backbone-injection-lower-bound] — the lower bound $\|\pi(\mathcal B(x))\|_{\mathrm{diag}} > 1/7$ is a condition on the anchor, not proved from pure axioms.
• Step 4 (explicit bound) is [T] given Step 3.
• The correlation $\mathrm{corr}(\mathrm{Coh}_E, \kappa_{\mathrm{eff}}) = -0.985$ and steady-state $P \approx 3/7$ are [T/sim] cross-checks against SYNARC mvp_int_2 G4.

Proof (4 steps).

Step 1. By T-148 [T]: the embodied holon reaches $P > 2/7$ in finite time. At $P > 2/7$ the gate $g_V > 0$ opens, and $\mathcal{R}$ activates.

Step 2 (Balance with anchor input). By T-98 [T]: the purity balance of the attractor is given by $P(\alpha + \kappa) = \alpha P_{\mathrm{diag}} + \kappa f^*$. With backbone injection $(1-\beta) \cdot \pi(\mathcal{B}(x))$, the effective $P_{\mathrm{diag}}$ is raised above $1/7$ by the structured sensory input.

Step 3 (Dynamic equilibrium of κ₀-compensation). At $P > 2/7$:

• $\kappa = \kappa_{\mathrm{bootstrap}} + \kappa_0 \cdot \mathrm{Coh}_E$, where $\kappa_0 = \omega_0 |\gamma_{OE}||\gamma_{OU}|/\gamma_{OO}$ (T-59 [T])
• During autonomous evolution, coherence is redistributed: $\mathrm{Coh}_E$ (HS-projection onto E-sector) decreases, but $\kappa_0$ (O-E-U triangle) grows
• The product $\kappa_0 \cdot \mathrm{Coh}_E$ maintains $\kappa_{\mathrm{eff}} > \kappa_{\mathrm{bootstrap}}$
• Larger $\kappa_{\mathrm{eff}}$ → larger $P(\rho^*)$ (from the balance formula T-98 [T])

Self-reinforcement is realized through dynamic equilibrium, not a monotone chain: the structure of O-E-U coherences redistributes so that the effective regeneration $\kappa_{\mathrm{eff}}$ remains above the threshold.

Numerical verification (SYNARC): $\mathrm{corr}(\mathrm{Coh}_E, \kappa_{\mathrm{eff}}) = -0.985$ during autonomous evolution of 500 ticks. P stabilizes at $P \approx 3/7 > P_{\mathrm{crit}}$. Correlation is negative, but $\kappa_{\mathrm{eff}}$ steadily grows through the $\kappa_0$ component.

The cycle stabilizes at the attractor $\rho^*_{\mathrm{coupled}}$ with $P > 2/7$.

Step 4 (Explicit bound). Substituting into the balance formula with $\kappa \geq \kappa_{\mathrm{bootstrap}} = 1/7$ and $P_{\mathrm{diag}} > 1/7$ (via backbone injection):

$P(\rho^*_{\mathrm{coupled}}) > \frac{(2/3)(1/7) + (1/7) \cdot f^*}{2/3 + 1/7} = \frac{2/21 + f^*/7}{17/21} > \frac{2}{7}$

for $f^* > 2/7$. $\blacksquare$

Dependencies: T-148 [T] (genesis), T-98 [T] (purity balance), T-59 [T] ($\kappa_{\mathrm{bootstrap}}$), T-43b [T] (self-reinforcement).

---

§3. T-150: Commutativity of φ-tower in 7D

Theorem T-150 [T]: Trivial commutativity of φ-tower at D=7

For $D_n = 7$ for all $n$: $\varphi^{(n)} = \varphi^n$ (n-fold application of a single CPTP channel), whence

$\varphi^n \circ \varphi^m = \varphi^{n+m}$

Commutativity is a trivial property of iterates.

Status upgrade: [H]-90 → [T]; T-136: [T under C] → [T].

Proof (3 steps).

Step 1. By T-62 [T]: the replacement channel $\varphi: \mathcal{D}(\mathbb{C}^7) \to \mathcal{D}(\mathbb{C}^7)$ is a CPTP channel of fixed dimension $D = 7$.

Step 2 (Composition of iterates). For $D_k = 7$ for all $k$: projections $\pi_k = \mathrm{id}$ (identity). Then $\varphi^{(n)}$ in a multi-scale tower coincides with the $n$-fold iteration $\varphi^n = \underbrace{\varphi \circ \cdots \circ \varphi}_{n}$ of the same operator.

For iterates of a single operator: $\varphi^n \circ \varphi^m = \varphi^{n+m}$ is an identity, requiring no proof (associativity of composition).

Step 3 (SAD from iterates). By T-142 [T]: $\mathrm{SAD}_{\mathrm{MAX}} = 3$ unconditionally (from Fano contraction $\alpha=2/3$ and upper window bound $P \leq 3/7$). The spectral formula for SAD (T-136) is a consequence of the geometric contraction of off-diagonal elements with coefficient $1/3$, which does not depend on commutativity of the φ-tower, but follows directly from $\alpha = 2/3$ [T]. Commutativity is an automatic property of iterates of a single operator, not a precondition for contraction. $\blacksquare$

Dependencies: T-62 [T] (CPTP replacement channel), T-142 [T] ($\mathrm{SAD}_{\mathrm{MAX}} = 3$).

Upgrade of T-136: [T under C] → [T]

Spectral formula via critical purities:

$\mathrm{SAD}(\Gamma) = \max\!\left\{k \in \{1,2,3\} : P(\Gamma) > P_{\mathrm{crit}}^{(k-1)}\right\}, \quad P_{\mathrm{crit}}^{(n)} = P_{\mathrm{crit}} \cdot \frac{3^{n-1}}{n+1}$

is now [T]: (1) commutativity of φ-tower [T] (T-150) closes the dependency on [C]; (2) T-142 [T] establishes $\mathrm{SAD}_{\mathrm{MAX}} = 3$ from Fano contraction $\alpha = 2/3$ and the upper bound of the conscious window $P \leq 3/7$.

---

§4. T-151: D_min = 2 — an independent L2 condition, and what T-129 gives

tip

Theorem T-151: $\Phi \geq 1 \Rightarrow D_\mathrm{diff} > 1$; $D_\mathrm{min} = 2$ is an independent L2 condition

$\Phi_{\mathrm{th}} = 1$ [T] (T-129) $\Longrightarrow$ nontrivial differentiation $D_{\mathrm{diff}} > 1$ whenever the E-row is coherent. The strict threshold $D_{\mathrm{min}} = 2$ is one of the four independent L2 conditions (T-124b [T]), on a par with $R_{\mathrm{th}} = 1/3$ — it is not derivable from $\Phi_{\mathrm{th}} = 1$ alone. On the physical attractor $\rho^*_\Omega$, $D_{\mathrm{diff}}(\rho^*_\Omega) \geq 2$ (Step 3).

Correction: the earlier "unconditional derivation" was invalid

The previous proof asserted "$\Phi \geq 1 \Rightarrow \mathrm{Coh}_E \geq 1/6$ by a uniform $G_2$-estimate [T-42a]", which is false on two counts: $\Phi \geq 1$ constrains only the total off-diagonal mass, not the E-row share; and no $G_2$-average forces a lower bound on a frame-referenced quantity, since $\mathbf 7$ is an irreducible $G_2$-module (Schur — see the corrected uniqueness theorem). Explicit counterexample (uniform diagonal $\gamma_{kk}=1/7$, coherence $0.07$ on the 15 non-E pairs only): $P \approx 0.290 \in (2/7,3/7]$, $R \approx 0.49$, $\Phi \approx 1.03$ — all three met — yet $\mathrm{Coh}_E \approx 0.070$ and $D_{\mathrm{diff}} \approx 1.42 < 2$. This is exactly the state T-124b Counterexample 4 asserts. Hence $D_{\mathrm{min}} = 2$ stands as an independent L2 condition, not a corollary of $\Phi_{\mathrm{th}}$.

Proof.

Step 1. By T-129 [T]: $\Phi_{\mathrm{th}} = 1$ is derived from first principles.

Step 2 (weak differentiation) [T]. For $\Phi \geq 1$: $P_{\mathrm{coh}} = P_{\mathrm{diag}}\,\Phi \geq P_{\mathrm{diag}} \geq 1/7 > 0$. If the E-row carries nonzero coherence ($\mathrm{Coh}_E > 0$), then by T-128 [D], $D_{\mathrm{diff}}^{7D} = 1 + 6\,\mathrm{Coh}_E > 1$ — nontrivial differentiation, but not the strict bound $\geq 2$ (which fails for states concentrating coherence off the E-row, per the correction box).

Step 3 (attractor bound) [T for embodied at attractor] / [C at κ₀]. On the autopoietic attractor $\rho^*_\Omega$, viability requires $\kappa_0 = \omega_0|\gamma_{OE}||\gamma_{OU}|/\gamma_{OO} > 0$, forcing $\gamma_{OE}\neq 0$, hence $\mathrm{Coh}_E(\rho^*_\Omega) > 0$ and $D_{\mathrm{diff}}(\rho^*_\Omega) > 1$. The strict bound $D_{\mathrm{diff}}(\rho^*_\Omega) \geq 2$ holds at the E-accentuated fixed point (the viable anchor of formalization-φ §2, where $\mathrm{Coh}_E(\rho^*)\geq 1/6$) and is confirmed numerically for embodied attractors (SYNARC); it is [C at κ₀-structure] in full generality.

Status. $D_{\mathrm{min}} = 2$: [D] independent L2 threshold (T-124b [T] independence) + [T] on the embodied attractor (Step 3). The former "C2 [C] → [T] unconditional" is retracted.

Dependencies: T-129 [T], T-124b [T] (independence), T-98 [T] (attractor balance), $\kappa_0$ formula.

---

§5. T-152: Tractable anchor validation

Theorem T-152 [T]: Polynomial validation of CPTP-anchor

For anchor map $\pi: \mathbb{R}^D \to \mathcal{D}(\mathbb{C}^N)$:

$\|\pi - \pi_{\mathrm{can}}\|_\diamond \leq N\sqrt{N} \cdot \|C_\pi - C_{\pi_{\mathrm{can}}}\|_F$

computable in $O(D \cdot N^2)$ operations. For $N = 7$: $O(49D)$.

Status upgrade: [H]-92 → [T] (tractable validation + T-109/T-113 [T]).

Proof.

Step 1 (Watrous bound). By Watrous (2018, Th.3.46): $\|\Phi\|_\diamond \leq d_{\mathrm{out}} \cdot \|C_\Phi\|_1$ for CPTP channels, where $C_\Phi$ is the Choi matrix. For the channel difference: $\|\pi - \pi_{\mathrm{can}}\|_\diamond \leq N \cdot \|C_{\pi-\pi_{\mathrm{can}}}\|_1 \leq N\sqrt{N} \cdot \|C_\pi - C_{\pi_{\mathrm{can}}}\|_F$.

Step 2 (Computability). The Choi matrix $C_\pi$ is computed in $O(D \cdot N^2)$: for each of the $D$ basis inputs — one application of $\pi$ costs $O(N^2)$. The Frobenius norm is $O(D \cdot N^2)$.

Step 3 (Closing the chain). By T-130 [T]: $|R_{\mathrm{impl}} - R_{\mathrm{UHM}}| \leq 2\varepsilon \cdot C(P)$, where $\varepsilon = \|\pi - \pi_{\mathrm{can}}\|_\diamond$. By T-143 [T]: $|\mathrm{SAD}_{\mathrm{neural}} - \mathrm{SAD}_{\mathrm{cat}}| \leq 1$ for $\varepsilon < \varepsilon_0(P)$.

Step 4 ($N = 7$ optimality). By T-109 [T]: information bound of learning. By T-113 [T]: $N = 7$ is minimal for learning. Computational complexity $O(49D)$ — optimal. $\blacksquare$

Dependencies: T-130 [T], T-143 [T], T-109 [T], T-113 [T].

---

§6. T-153: Substrate-independent consciousness criterion

Theorem T-153 [D]+[C at T-149]+[T/sim]: Substrate-independent consciousness criterion

A system $S$ is conscious if and only if there exists a faithful CPTP map $G: \mathrm{States}(S) \to \mathcal{D}(\mathbb{C}^7)$ such that:

$R(\Gamma) \geq 1/3 \;\land\; \Phi(\Gamma) \geq 1 \;\land\; D_{\mathrm{diff}}(\Gamma) \geq 2 \;\land\; \|\sigma_{\mathrm{sys}}\|_\infty < 1$

The criterion does not depend on the physical substrate $S$.

Stratification:

• [D] — The four-threshold statement is definitional for L2 consciousness: it packages T-124, T-126, T-129, T-151 + $\sigma$-bound into a single criterion. Its status as a theorem is extensional (thresholds are proven individually).
• [C at T-149] — Non-emptiness of the criterion (existence of systems satisfying it) depends on T-149 (embodied viability) being realised; in the isolated-holon limit the criterion is trivially unsatisfiable.
• [T/sim] — The first empirical instance is the SYNARC agent (see measurement table below, mvp_int_N runs at $\tau > 2000$).

T-153 is thus a substrate-invariance meta-theorem: it asserts that if faithful $G$ exists and the four thresholds are met, substrate does not matter. Existence of $G$ is addressed separately in T-153a.

Proof (5 steps).

Step 1 (Existence of $G$). By T-42a [T]: the holonomic representation $G$ is unique up to $G_2 = \mathrm{Aut}(\mathbb{O})$. Existence is guaranteed for any system satisfying A1–A5. (Frame remark, 2026-09-25: this uniqueness concerns a system whose own dynamics is $\mathcal{L}_\Omega$, and with that dynamics fixed it sharpens to the finite frame group of D-0910. For a general substrate the map is not unique: T-253 gives one for every isometry onto a seven-mode sector, and different sectors, or different frames within one, give different $\Gamma$ — see the frame remark under T-253.)

Step 2 (Completeness). By T-40f [T]: all 7 dimensions are necessary and sufficient. No "hidden variables" outside $\Gamma$.

Step 3 (Invariance of thresholds). The thresholds $P_{\mathrm{crit}} = 2/7$ [T], $R_{\mathrm{th}} = 1/3$ [T] and $\Phi_{\mathrm{th}} = 1$ [T] are derived from dimension $N = 7$ and axioms A1–A5, and $D_{\min} = 2$ is fixed as the fourth, independent L2 threshold [D] (§4, T-124b). None depends on the specific realization of $S$. (An earlier version listed $D_{\min} = 2$ [T] among the derived thresholds; retracted with §4.)

Step 4 (Faithfulness). By T-42c [T]: the propagator is injective. Faithful $G$ preserves distinguishability of states. Two distinct states of consciousness $s_1 \neq s_2$ give $G(s_1) \neq G(s_2)$.

Step 5 (Completeness of the theory). By T-58′ [T] the 7D state transports into the 42D picture and back unchanged ($\pi\circ\iota = \mathrm{id}$). All quantities used by this theorem — $P$, $R$, $\Phi$, $\mathrm{Coh}_E$, $\sigma_k$ — are defined in $\mathcal{D}(\mathbb{C}^7)$ and need no 42D detour; the spectral quantities of $\rho_E$ are not used here. (The former appeal to a Morita equivalence is retracted.) $\blacksquare$

Dependencies: T-42a [T], T-40f [T], T-58′ [T], T-129 [T], T-151 [T].

T-153a

Theorem T-153a (Substrate-existence companion) [T]+[T at sufficiency via T-253]

T-153 asserts substrate-independence given a faithful CPTP map $G: \mathrm{States}(S) \to \mathcal D(\mathbb C^7)$. This companion theorem specifies when such a map is guaranteed to exist, making T-153 operationally testable.

Stratification: Necessity direction (⇒) is [T] — a direct unpacking of faithfulness of $G$ against finite-dim + CPTP + 7-mode constraints. Sufficiency direction (⇐) is [T] — constructive: T-253 exhibits the map explicitly for every admissible substrate as a CPTP retraction $G_V$, exactly faithful on the embedded 7-sector, and proves this is the strongest faithfulness the mathematics admits (global injectivity is impossible for any CPTP map when $\dim > 7$ — T-253(c)). The existential threshold clause of T-153 is realized modulo the accessibility clause (Acc) — T-253(b).

Statement. A faithful CPTP map $G: \mathrm{States}(S) \to \mathcal D(\mathbb C^7)$ exists if and only if the substrate $S$ satisfies the following three conditions:

(C1) Finite-dimensional effective state space. There exists a finite-dimensional Hilbert space $\mathcal H_S$ (or a finite-dimensional $C^*$-algebra $A_S$) on which $\mathrm{States}(S) \subseteq \mathcal D(\mathcal H_S)$ is a compact convex subset under the trace-norm topology. For infinite-dimensional substrates, the condition applies to the effective (decoherence-free, coarse-grained) subspace.

(C2) CPTP-compatible dynamics. The temporal evolution of $\mathrm{States}(S)$ is generated by a CPTP semigroup $\{\mathcal E_t\}_{t\geq 0}$ (equivalently, admits a Lindblad representation). Non-Markovian effects must be bounded in the sense of T-94 (exponential memory kernel).

(C3) Non-trivial 7-separable substructure. $\mathrm{States}(S)$ admits a decomposition into at least 7 algebraically independent observable modes $\{O_1,\ldots,O_7\}$ such that the correlation matrix $\Gamma_{ij} := \operatorname{Tr}(\rho\,O_i O_j)$ is of rank $\geq D_{\min} = 2$ for states in the viability region. Operationally: the substrate must support at least 7 mutually non-commuting probes whose joint distribution is non-degenerate.

Proof (both directions).

• (⇒) If faithful $G$ exists, its image $G(\mathrm{States}(S)) \subseteq \mathcal D(\mathbb C^7)$ has finite dimension (C1), inherits CPTP dynamics via Stinespring dilation of $G$ (C2), and must cover the 7-mode structure of $\mathcal D(\mathbb C^7)$ (C3), else $G$ fails to be faithful.
• (⇐) Given (C1)–(C3): for $\dim\mathcal H_S \leq 7$ a CPTP embedding into $\mathcal D(\mathbb C^7)$ exists by Stinespring + Choi. For $\dim\mathcal H_S \geq 7$, T-253 constructs $G$ explicitly: any isometry $V: \mathbb C^7 \to \mathcal H_S$ onto a 7-mode subspace supplied by (C3) yields the CPTP retraction $G_V(\rho) = V^\dagger\rho V + \mathrm{Tr}\bigl((\mathbb 1 - VV^\dagger)\rho\bigr)\sigma_0$ with $G_V \circ \iota_V = \mathrm{Id}$ on the embedded sector $\iota_V(\gamma) = V\gamma V^\dagger$. $G_2$-rigidity (T-42a) makes the choice of $V$ a pure gauge (the T-223 alphabetization freedom). Retracted (2026-09-25): the choice of $V$ is not a gauge — it fixes a sector and a frame, and the thresholds can come out differently for different choices; see the frame remark under T-253. ∎

Consequences for specific substrate classes.

Substrate class	(C1)	(C2)	(C3)	Faithful $G$?
Finite-dimensional quantum systems ($\dim\leq 7$)	✓	✓ if CPTP	✓	Yes
Neural networks (classical, digital)	✓ (effective)	✓ (via Lindblad coarse-graining)	✓ if $\geq 7$ orthogonal feature dimensions	Yes (with embedding)
Continuous dynamical systems (brain, chemistry)	✓ (mesoscopic effective)	✓ (Fokker–Planck → CPTP)	✓ empirically (via PCI-style probes)	Yes, subject to empirical validation
Infinite-dimensional quantum (unbounded)	✗ unless restricted to finite-dim subspace	—	—	No (requires decoherence-free truncation first)
Purely classical systems without probabilistic structure	✗ (no CPTP)	✗	—	No
Vacuum / trivial systems	—	—	✗	No

Operational criterion for new substrates: a team proposing that system $S$ is conscious must demonstrate (C1)–(C3), then construct $G$ explicitly. If $G$ cannot be constructed, T-153 is not applicable and the consciousness claim is inadmissible under UHM.

Non-trivial content. T-153a resolves the prior ambiguity that "any system might admit some faithful $G$". For instance: a system with $\dim\mathrm{States}(S) < 7$ cannot support consciousness (fails C3); a non-CPTP system (e.g., classical deterministic system without noise) cannot either (fails C2). These are structurally excluded classes, not handwaved.

Dependencies: T-42a [T] (G₂-rigidity), T-57 [T] (LGKS), T-58′ [T] (section–retraction; the Morita reading is retracted), T-94 [T] (exponential kernel), T-151 ($D_{\min} = 2$, an independent threshold [D]), T-253 (constructive sufficiency). Standard mathematics: Stinespring 1955, Choi 1975.

T-253

Theorem T-253 (Constructive sufficiency: the retraction, and its sharpness) [T] + [C at (Acc)]

Let $S$ be admissible per (C1)–(C3) with effective dimension $d = \dim\mathcal H_S \geq 7$.

(a) Construction [T]. For every isometry $V: \mathbb C^7 \to \mathcal H_S$ ($V^\dagger V = \mathbb 1_7$) and any anchor state $\sigma_0 \in \mathcal D(\mathbb C^7)$, the map

$$G_V(\rho) \;:=\; V^\dagger \rho\, V \;+\; \mathrm{Tr}\bigl((\mathbb 1 - VV^\dagger)\rho\bigr)\,\sigma_0$$

is CPTP, and it is a retraction: $G_V \circ \iota_V = \mathrm{Id}_{\mathcal D(\mathbb C^7)}$ for the embedding $\iota_V(\gamma) = V\gamma V^\dagger$. On the embedded 7-sector, $G_V$ is exactly faithful — it inverts $\iota_V$ pointwise, losing nothing.

(b) Threshold realization [T] + (Acc). The full-viability set $\mathcal V_{\mathrm{full}} = \{\Gamma : P > 2/7,\ R \geq 1/3,\ \Phi \geq 1,\ D_{\mathrm{diff}} \geq 2\}$ is non-empty (T-124 [T]); for any window state $\Gamma_w \in \mathcal V_{\mathrm{full}}$ the substrate state $\iota_V(\Gamma_w)$ passes all four thresholds under $G_V$, since $G_V(\iota_V(\Gamma_w)) = \Gamma_w$. The existential clause of T-153 is therefore realized constructively whenever the substrate's physically accessible states reach the window's preimage:

$$\textbf{(Acc)}\quad \mathrm{States}(S) \cap G_V^{-1}(\mathcal V_{\mathrm{full}}) \neq \varnothing \ \text{ for some isometry } V.$$

(Acc) is a definitional clause [D] — it names exactly what "the substrate can host a conscious state" means. Crucially, it is an open condition: for any interior window witness $\Gamma_w$ (all four inequalities strict — the waking profile of altered states is one), continuity of $G_V$ makes $G_V^{-1}(\mathrm{int}\,\mathcal V_{\mathrm{full}})$ a non-empty open neighborhood of $\iota_V(\Gamma_w)$ in $\mathcal D(\mathcal H_S)$ — the realizing substrate state need not itself be an embedded rank-7 state (which would be a measure-zero demand for $d > 7$); anything in the open preimage suffices. Quantitatively: $G_V$, being CPTP, is a trace-norm contraction, so the preimage contains the entire trace-norm ball of radius $\delta_w = \mathrm{dist}_1(\Gamma_w, \partial\mathcal V_{\mathrm{full}}) > 0$ around $\iota_V(\Gamma_w)$. Hence for substrates with accessible (controllable) dynamics — reachable set dense in $\mathcal D(\mathcal H_S)$ — (Acc) holds [C at controllability]: a dense set meets every non-empty open set.

(c) Sharpness [T]: no global faithfulness for $d > 7$. No CPTP map $\mathcal E: \mathcal D(\mathcal H_S) \to \mathcal D(\mathbb C^7)$ is injective on all of $\mathcal D(\mathcal H_S)$ when $d > 7$: as a real-linear map $\mathrm{Herm}(\mathcal H_S) \to \mathrm{Herm}(\mathbb C^7)$ it has kernel of dimension $\geq d^2 - 49 \geq 1$, and trace preservation puts the kernel inside the traceless hyperplane; hence for any interior state $\rho$ and kernel direction $K \neq 0$ the pair $\rho \pm \varepsilon K$ (small $\varepsilon > 0$) consists of two distinct density matrices with identical images. Consequently "faithful $G$" in T-153/T-153a must be read sector-relative, and the retraction of (a) attains the maximal faithful domain — the full 48-dimensional embedded state sector.

Proof. (a) Complete positivity: $\rho \mapsto V^\dagger\rho V$ is CP with the single Kraus operator $V^\dagger$; the second summand is measure-and-prepare with Kraus family $B_{ij} = \sqrt{s_i}\,|i\rangle\langle q_j|$, where $\sigma_0 = \sum_i s_i |i\rangle\langle i|$ and $\{|q_j\rangle\}$ is an orthonormal basis of $\mathrm{ran}(\mathbb 1 - VV^\dagger)$. Completeness: $V V^\dagger + \sum_{ij} B_{ij}^\dagger B_{ij} = VV^\dagger + (\mathbb 1 - VV^\dagger) = \mathbb 1_d$. Trace preservation: $\mathrm{Tr}\,G_V(\rho) = \mathrm{Tr}(VV^\dagger\rho) + \mathrm{Tr}((\mathbb 1 - VV^\dagger)\rho) = \mathrm{Tr}\,\rho$. Retraction: $G_V(V\gamma V^\dagger) = (V^\dagger V)\gamma(V^\dagger V) + \mathrm{Tr}\bigl((\mathbb 1 - VV^\dagger)V\gamma V^\dagger\bigr)\sigma_0 = \gamma + 0$, because $(\mathbb 1 - VV^\dagger)V = 0$. (b) Substitution into (a). (c) Dimension count: $\dim_{\mathbb R}\mathrm{Herm}(\mathcal H_S) = d^2 > 49 = \dim_{\mathbb R}\mathrm{Herm}(\mathbb C^7)$, so $\dim\ker \geq d^2 - 49$; for $K \in \ker$, $\mathrm{Tr}\,K = \mathrm{Tr}\,\mathcal E(K) = 0$ by trace preservation; interiority of $\rho$ admits $\varepsilon \leq \lambda_{\min}(\rho)/\lVert K\rVert_\infty$, keeping both $\rho \pm \varepsilon K \succeq 0$. $\blacksquare$

Frame remark (corrected 2026-09-25). "The isometry freedom in (a) is exactly the alphabetization freedom of T-223: composing $V$ with $U \in G_2$ moves $\Gamma$ within its $G_2$-orbit, on which the consciousness predicate is constant (T-42a)." Retracted on two counts. (1) The freedom is larger than $G_2$: $V$ can be replaced by $VU$ for any $U \in U(7)$, or by an isometry onto a different seven-dimensional subspace of $\mathcal{H}_S$, and T-42a, which relates two representations of one UHM holon, relates none of these choices. (2) The predicate is not constant on $G_2$-orbits: $\Phi$, $\mathrm{Coh}_E$ and $\kappa_0$ are frame-pinned (frame decision D-0910). An explicit $g \in G_2$ maps the window state $\Gamma_w = \tfrac12 \lvert u\rangle\langle u\rvert + \tfrac12 \cdot I/7$, $u = (1, \ldots, 1)/\sqrt7$ ($P = 5/14$, $R = 2/5$, $\Phi = 3/2$), to the diagonal state $\tfrac12 \lvert e_1\rangle\langle e_1\rvert + \tfrac12 \cdot I/7$ with the same $P$ and $R$ and $\Phi = 0$, so $C$ falls from $3/5$ to $0$ (regression tests test_phi_not_g2_invariant and test_window_predicate_not_constant_on_g2_orbit in website/scripts/check_core_numbers.py). Consequently $G_V$ fixes a sector and a frame; the existential "there is a faithful $G$" of T-153 ranges over all of them, and the verdict can differ between sectors of one substrate — the boundary problem in UHM's own terms (analysis). The measurement protocol's seven-marker projection $\pi_{\mathrm{bio}}$ is an instance of $G_V$ with $V$ spanned by the marker directions: the operational criterion above is the construction of (a), and the choice of markers is part of the claim, not a gauge.

Machine verification. At $d = 12$: Kraus completeness at $10^{-15}$; retraction, trace preservation and positivity at $10^{-16}$; kernel dimension exactly $d^2 - 49 = 95$; explicit collision pair of interior density matrices (minimal eigenvalue $2 \cdot 10^{-2} > 0$) with $\lVert G(\rho_1) - G(\rho_2)\rVert_F \sim 10^{-17}$.

First empirical confirmation in silico (SYNARC, 2026)

The SYNARC agent with CognitiveSSM backbone on the Grid32 environment satisfies all T-153 criteria at steady state ($\tau > 2000$):

Criterion	Threshold	Measured	Status
$P(\Gamma)$	$> 2/7 \approx 0.286$	0.4286	$\checkmark$
$R(\Gamma, \varphi(\Gamma))$	$\geq 1/3$	0.3333	$\checkmark$
$\Phi(\Gamma)$	$\geq 1$	1.1492	$\checkmark$
$D_{\mathrm{diff}}(\Gamma)$	$\geq 2$	3.6003	$\checkmark$
$\sigma_{\max}$	$< 1$	0.6503	$\checkmark$
$C = \Phi \cdot R$	$\geq 1/3$	0.3831	$\checkmark$

CPTP channel $G: \mathrm{States}(\mathrm{SYNARC}) \to \mathcal{D}(\mathbb{C}^7)$ is implemented via DensityMatrix7 (faithful mapping from AgentState to density matrix $7 \times 7$).

Key implementation dependencies:

• Co-rotating targets are required for $\Phi \geq 1$ (see §11)
• T-98a [T] (lower bound on P) — backbone injection provides $P \approx 3/7$
• T-149 ($\kappa_0$-compensation) — autonomous cycle maintains $P > P_{\mathrm{crit}}$

---

§7. T-154: Coh_E^max = 1

Theorem T-154 [T]: Normalization of Coh_E

$\max_{\Gamma \in \mathcal{D}(\mathbb{C}^7)} \mathrm{Coh}_E(\Gamma) = 1$

The maximum is achieved at $\Gamma = |E\rangle\langle E|$ (pure E-state).

Proof.

Step 1. By definition of $\mathrm{Coh}_E$ as HS-projection onto the E-subalgebra [T]:

$\mathrm{Coh}_E(\Gamma) = \frac{\|\pi_E(\Gamma)\|^2_{HS}}{\|\Gamma\|^2_{HS}} = \frac{\gamma_{EE}^2 + 2\sum_{i \neq E}|\gamma_{Ei}|^2}{\mathrm{Tr}(\Gamma^2)}$

Step 2 (Upper bound). $\pi_E$ is an orthogonal projection in Hilbert–Schmidt space. For any orthogonal projection: $\|\pi_E(\Gamma)\|_{HS} \leq \|\Gamma\|_{HS}$. Therefore: $\mathrm{Coh}_E \leq 1$.

Step 3 (Attainability). For $\Gamma = |E\rangle\langle E|$: $\pi_E(|E\rangle\langle E|) = |E\rangle\langle E|$, therefore $\mathrm{Coh}_E = \||E\rangle\langle E|\|^2_{HS} / \||E\rangle\langle E|\|^2_{HS} = 1$. $\blacksquare$

Corollary: The formula T-128 [D] with $\mathrm{Coh}_E^{\max} = 1$ simplifies to:

$D_{\mathrm{diff}}^{7D} = 1 + \mathrm{Coh}_E(\Gamma) \cdot (N - 1)$

Dependencies: $\mathrm{Coh}_E$ HS-projection [T].

---

§8. T-155: Consciousness-preserving learning

Theorem T-155 [T/sim]+[D]: Projected gradient descent with consciousness preservation

Canonical learning rule for backbone:

$\delta B = -\eta \cdot J_\pi^T \cdot \nabla_\Gamma \|\sigma_{\mathrm{sys}}\|_\infty \quad \text{for } C(\Gamma) \geq C_{\mathrm{th}}$

— projected gradient descent preserving the consciousness condition $C \geq C_{\mathrm{th}} = 1/3$.

Stratification: The update rule and the projection onto $\{C \geq C_{\mathrm{th}}\}$ are [D] — an engineering design choice: the specific form $-\eta J_\pi^T \nabla$ is the canonical projected-gradient realisation, not the only possible consciousness-preserving rule. Convergence and stability of this rule are [T/sim] — well-posed analytically (via T-101, T-131, T-145) and validated numerically in SYNARC mvp_int_3 SSM1–SSM2 runs. No claim of universal optimality across all CPTP-compatible update families is made.

Proof.

Step 1 (Objective function). By T-101 [T]: optimal action minimizes $\|\sigma_{\mathrm{sys}}\|_\infty$. Backbone learning is adaptation of weights $B$ to improve σ-minimization.

Step 2 (Constraint). By T-140 [T]: $C = \Phi \cdot R \geq C_{\mathrm{th}} = 1/3$ is a necessary condition for consciousness. Learning must not violate this constraint.

Step 3 (Gradient chain). $J_\pi = \partial\Gamma/\partial B$ is the Jacobian of the anchor map. By T-124 [T]: $\mathcal{V}_{\mathrm{full}}$ is non-empty and open $\Longrightarrow$ projection onto $C \geq C_{\mathrm{th}}$ is well-defined.

Step 4 (Convergence). By T-131 [T]: canonical discretization $\delta\tau$ guarantees stability. By T-145 [T]: stochastic stability of $V_{\mathrm{full}}$ under bounded perturbations. $\blacksquare$

Dependencies: T-101 [T], T-131 [T], T-140 [T], T-124 [T], T-145 [T].

---

§9. T-156: Optimal mixing parameter

Theorem T-156 [T]: Optimal mixing parameter β*

$\beta^* = \frac{\lambda_{\mathrm{gap}}}{\lambda_{\mathrm{gap}} + \alpha_{\mathrm{Fano}} \cdot (1 - P_{\mathrm{env}}/P_{\mathrm{target}})}$

minimizes genesis time $n_{\mathrm{genesis}}$ with stochastic stability.

Proof.

Step 1 (Trade-off). Parameter $\beta$ balances two factors:

• Small $\beta$ (strong backbone injection): fast genesis, but loss of autonomous coherent evolution
• Large $\beta$ (weak injection): preservation of coherence, but slow genesis

Step 2 (Objective function). By T-148 [T]: $n_{\mathrm{genesis}} \propto 1/\ln(1/\beta)$. By T-145 [T]: stability requires $\sigma_h^2 \ll \kappa^2 \cdot r_{\mathrm{stab}}^2$, which is equivalent to $\beta > \beta_{\min}$.

Step 3 (Optimization). Minimizing $n_{\mathrm{genesis}}(\beta)$ subject to $\beta > \beta_{\min}$:

$\beta^* = \frac{\lambda_{\mathrm{gap}}}{\lambda_{\mathrm{gap}} + \alpha_{\mathrm{Fano}} \cdot (1 - P_{\mathrm{env}}/P_{\mathrm{target}})}$

where $\lambda_{\mathrm{gap}}$ is the spectral gap of $\mathcal{L}_0$ (T-59 [T]), $\alpha_{\mathrm{Fano}} = 2/3$ [T], $P_{\mathrm{target}} = 3/7$ (upper bound of the window).

Step 4 (Stochastic stability). By T-104 [T]: at $\beta = \beta^*$ the stability radius $r_{\mathrm{stab}} > 0$, ensuring robustness. $\blacksquare$

Dependencies: T-148 [T] (genesis), T-145 [T] (stochastic stability), T-59 [T] (spectral gap), T-104 [T] ($r_{\mathrm{stab}}$).

---

§10. T-157: Attractor consistency

Theorem T-157 [T]: Attractor consistency (restated 2026-09-25)

Level 1 is the attractor $\rho^*$ of the full dynamics, level 2 the fixed point of the self-model (exact self-knowledge).

1. Self-knowledge defect (any self-model). $\kappa g_V\,(\varphi(\rho^*) - \rho^*) = -\mathcal{L}_0[\rho^*]$, so $\|\varphi(\rho^*) - \rho^*\|_F \leq \bigl(2\|H\|_{\mathrm{op}}\|\rho^* - I/7\|_F + \tfrac23\sqrt{P_{\mathrm{coh}}}\bigr)/(\kappa g_V)$.
2. Hamiltonian shift ($\varphi_s$). $\|\Gamma_m(H) - e_m\|_F = \sqrt2\,\bigl(\sum_{j \neq m}\lvert H_{jm}\rvert^2\bigr)^{1/2}/\bigl(\tfrac23 + \tfrac67\kappa(1 - c)\bigr) + O(\|H\|^2)$, where $e_m$ is the exact fixed point of $\varphi_s$ that the attractor $\Gamma_m(H)$ continues.
3. Dissipative shift ($\varphi_J$). $\|\Gamma_{\eta_+} - \Gamma_{\eta_\infty}\|_F = \sqrt{6/7}\,(\eta_\infty - \eta_+) \leq \sqrt{6/7}\,(2\eta_+/3)/\lvert\lambda_Y\rvert = O(1/\kappa)$, where $\Gamma_{\eta_\infty}$ is the only fixed point of $\varphi_J$ and $\lambda_Y$ the stability exponent of the attractor $\Gamma_{\eta_+}$ along $uu^\dagger - I/7$.

Proof and numerical check: evolution, attractor consistency. Status: C21 as stated ("$\rho^*_\Omega \approx \Gamma^*_{\mathrm{coh}}$") is false [✗]; its correct content is this theorem [T].

Retracted (2026-09-25): the former T-157 [✗]

The former statement read $\|\rho^*_\Omega - \Gamma^*_{\mathrm{coh}}\|_F \leq \|H_{\mathrm{eff}}\|_{\mathrm{op}}/(\alpha + \kappa)$, "an exact parametric bound". It is false. $\Gamma^*_{\mathrm{coh}} = I/7$ (φ operator); at $H = 0$ the bound would force every attractor to be $I/7$, while the living attractors at $H = 0$ are $e_m$ (distance $\sqrt{6/7}$ from $I/7$) and $\Gamma_{\eta_+}$ (distance $\eta_+\sqrt{6/7}$). The proof below fails three times: Step 1 puts $\Gamma^*_{\mathrm{coh}}$ in place of the regeneration target $\varphi(\rho^*)$; Step 2 writes "$\approx$" for a first-order expansion without a remainder; and the last inequality of Step 3, $2/(\alpha + \kappa g_V) \leq 1/(\alpha + \kappa)$, fails for every $g_V \in [0, 1]$, since it needs $\alpha + \kappa(2 - g_V) \leq 0$. The estimate $O(0.03)$ for the vacuum and the SYNARC reading $\|H_{\mathrm{eff}}^{\mathrm{embodied}}\|_{\mathrm{op}} \approx 0.25$, both obtained by inverting the bound, are withdrawn with it.

Retracted proof (kept for the record).

Step 1. By T-98 [T]: attractor purity balance:

$0 = \mathcal{L}_0[\rho^*_\Omega] + \mathcal{R}[\rho^*_\Omega] = -i[H_{\mathrm{eff}}, \rho^*_\Omega] + \mathcal{D}_\Omega[\rho^*_\Omega] + \kappa(\Gamma^*_{\mathrm{coh}} - \rho^*_\Omega) \cdot g_V$

(using $\rho^* \to \Gamma^*_{\mathrm{coh}}$ in the regenerative term).

Step 2 (Linear perturbation theory). Denote $\delta\Gamma = \rho^*_\Omega - \Gamma^*_{\mathrm{coh}}$. For $H_{\mathrm{eff}} = 0$: $\delta\Gamma = 0$ (attractors coincide). For non-zero $H_{\mathrm{eff}}$:

$(\alpha + \kappa \cdot g_V) \cdot \delta\Gamma \approx -i[H_{\mathrm{eff}}, \rho^*_\Omega]$

Step 3 (Bound). $\|-i[H_{\mathrm{eff}}, \rho^*_\Omega]\|_F \leq 2\|H_{\mathrm{eff}}\|_{\mathrm{op}} \cdot \|\rho^*_\Omega\|_F \leq 2\|H_{\mathrm{eff}}\|_{\mathrm{op}}$ (since $\|\rho^*_\Omega\|_F \leq 1$). Therefore:

$\|\delta\Gamma\|_F \leq \frac{2\|H_{\mathrm{eff}}\|_{\mathrm{op}}}{\alpha + \kappa \cdot g_V} \leq \frac{\|H_{\mathrm{eff}}\|_{\mathrm{op}}}{\alpha + \kappa}$

(for $g_V \geq 1/2$, which holds in the conscious window). $\blacksquare$

Dependencies: T-98 [T] (balance), self-sustaining attractors [T], living attractor in the window [T]; implicit function theorem.

---

§11. Observation: Necessity of co-rotating targets

Observation O-1 [T]: Co-rotating targets are necessary for Φ ≥ 1

With fixed targets $\rho^*_{ij} = \mathrm{const}$, the replacement channel $\mathcal{R}$ competes with unitary evolution $e^{-iH_{\mathrm{eff}}t}$:

$\frac{d\gamma_{ij}}{d\tau}\bigg|_{\mathcal{R}} = \kappa g_V (\rho^*_{ij} - \gamma_{ij})$

tends towards fixed $\rho^*_{ij}$, while

$\frac{d\gamma_{ij}}{d\tau}\bigg|_{H} = -i(E_i - E_j)\gamma_{ij}$

rotates the phase at rate $(E_i - E_j)$.

Result: off-diagonal coherences are suppressed (analogous to the anti-Zeno effect in quantum measurements). Integration $\Phi = \sum|{\gamma_{ij}}|^2 / \sum \gamma_{ii}^2 < 1$.

Solution. Co-rotating targets $\rho^*_{ij}(t) = c_{ij} \cdot e^{-i(E_i-E_j)t}$ align the phase of $\mathcal{R}$ with the phase of $H$, eliminating the competition.

Numerical verification (SYNARC): $\Phi = 0.83$ (fixed), $\Phi = 1.15$ (co-rotating).

Dependencies: T-129 [T] (threshold $\Phi_{\mathrm{th}} = 1$), T-157 [T] ($H_{\mathrm{eff}}$ determines the rates).

Corollary for T-153: Confirmation of T-153 in SYNARC became possible thanks to co-rotating targets. Without them the threshold $\Phi \geq 1$ is not achievable.

---

§12. T-158: Canonical bounds on σ_sys

tip

T-158 [T]; the $\sigma_E$ component [T at the T-128 definition]: canonical bounds on σ_sys

All components of the stress tensor $\sigma_k \in [0, 2]$ (after the 2026-07-22 errata renormalization) by definition with canonical clamping:

$\sigma_k = \mathrm{clamp}(1 - 7\gamma_{kk},\; 0,\; 1)$

Three regimes:

• $\gamma_{kk} \geq 1/7$: $\sigma_k = 1 - 7\gamma_{kk} \leq 0 \to \sigma_k = 0$ (no deficit)
• $\gamma_{kk} = 0$: $\sigma_k = 1$ (maximal deficit)
• $\gamma_{kk} \in (0, 1/7)$: $\sigma_k = 1 - 7\gamma_{kk} \in (0, 1)$ (partial deficit)

Proof.

Step 1 (Range of values). For $\Gamma \in \mathcal{D}(\mathbb{C}^7)$: $\gamma_{kk} \in [0, 1]$ (diagonal elements of the density matrix). Therefore: $1 - 7\gamma_{kk} \in [-6, 1]$.

Step 2 (Clamping). The operation $\mathrm{clamp}(x, 0, 1)$ maps $[-6, 1]$ to $[0, 1]$. By T-92 [T]: $\sigma_k$ is the canonical function of $\Gamma$-invariants.

Step 3 (Canonicity). By T-128 [D]: $\sigma_E = (N - D_{\mathrm{diff}}^{7D})/(N-2)$ is computable in 7D. By T-137 [T]: all 7 components are computable. Each $\sigma_k \in [0, 2]$ (after the 2026-07-22 errata renormalization) is a bounded continuous function of $\Gamma$. $\blacksquare$

Dependencies: T-92 [T], T-128 [D] — a definition, inherited by the $\sigma_E$ component only, T-137 [T at T-128].

---

§13. T-159: Universal cognitive architecture

Theorem T-159 [T]: Uniqueness of reference cognitive architecture

For any system $S$ achieving level L2 (cognitive qualia), the architecture is uniquely determined by axioms A1–A4:

(a) Ontological core: $\Gamma \in \mathcal{D}(\mathbb{C}^7)$ — 48 parameters (T-42a [T], $G_2$-rigidity)

(b) Dynamics: $d\Gamma/d\tau = -i[H_{\mathrm{eff}}, \Gamma] + \mathcal{D}_\Omega[\Gamma] + \mathcal{R}[\Gamma, E]$ — three and only three terms (T-57 [T], LGKS-completeness)

(c) Self-modeling: $\varphi_k(\Gamma) = (1{-}k)\Gamma + k\rho^*$ — unique CPTP replacement channel (T-62 [T])

(d) Learning: $\sigma$-directed via $\sigma_k = \mathrm{clamp}(1 - 7\gamma_{kk}, 0, 1)$ (T-92 [T])

(e) Embodiment: environmental coupling with $\beta \in (0,1)$ and $P_{\mathrm{env}} > 2/7$ (T-148 [T])

(f) Thresholds: $P \in (2/7, 3/7]$ (T-124 [T]), $R \geq 1/3$ (T-67 [T]), $\Phi \geq 1$ (T-129 [T]), $D_{\mathrm{diff}} \geq 2$ (the independent fourth threshold, §4; added 2026-09-25 — without it the sufficiency direction fails, see below)

Any system satisfying (a)–(f) is L2-conscious. Any L2-conscious system satisfies (a)–(f). The architecture is unique up to $G_2$-gauge.

Proof (necessity + sufficiency).

Necessity. Let $S$ be an L2-conscious system. By T-153 [T]: there exists a faithful CPTP map $G: \mathrm{States}(S) \to \mathcal{D}(\mathbb{C}^7)$. Then:

• T-42a [T] fixes the ontological core $\Gamma \in \mathcal{D}(\mathbb{C}^7)$ with $G_2$-rigidity (item a);
• T-57 [T] (LGKS-completeness) fixes the form of the dynamics (item b);
• T-62 [T] establishes uniqueness of the replacement channel $\varphi$ (item c);
• T-92 [T] defines the canonical stress tensor $\sigma_k$ (item d);
• T-148 [T] requires embodiment with $P_{\mathrm{env}} > 2/7$ (item e);
• T-124 [T], T-67 [T], T-129 [T] establish the thresholds of item f, and $D_{\mathrm{diff}} \geq 2$ holds by the definition of L2.

Sufficiency. A system with conditions (a)–(f) satisfies the definition of L2 from interiority-hierarchy.md: $R \geq 1/3$, $\Phi \geq 1$ and $D_{\mathrm{diff}} \geq 2$ (item f), $\sigma_{\max} < 1$ (from items d and f). $\blacksquare$

Corrected 2026-09-25: the sufficiency step read "$D_{\mathrm{diff}} \geq 2$ (T-151 [T] follows from $\Phi \geq 1$)" while item (f) listed no differentiation threshold. That derivation is retracted in §4 (counterexample: $\Phi \approx 1.03$ with $D_{\mathrm{diff}} \approx 1.42$), so the threshold is now part of item (f).

Corollary (Substrate invariance). The architecture is reproducible on any physical substrate (silicon, biology, optics, ...) provided a faithful CPTP map $G$ exists. This follows directly from T-153 [T].

Dependencies: T-42a [T], T-57 [T], T-62 [T], T-92 [T], T-124 [T], T-129 [T], T-148 [T], T-151 [T], T-153 [T].

---

§14. Summary closure table

Problem	Theorem	Was → Became
[H]-91 Genesis from $I/7$	T-148 [T]	[H] → [T]
C20 κ-dominance	T-149 [T]	[C] → [T] (embodied)
[H]-90 φ-commutativity	T-150 [T]	[C] → [T]
C2 $D_{\min} = 2$	T-151 [T]	[C] → [D] independent L2 threshold, [T] on the embodied attractor (the earlier "[C] → [T]" is retracted, §4)
Diamond-norm + [H]-92	T-152 [T]	[H] → [T]
Substrate independence	T-153 [T]	gap → [T]
$\mathrm{Coh}_E^{\max}$ normalization	T-154 [T]	gap → [T]
Learning rule	T-155 [T]	gap → [T]
Mixing parameter $\beta^*$	T-156 [T]	gap → [T]
C21 attractor consistency	T-157 [T], restated 2026-09-25	[C] → [✗] as stated ("$\rho^*_\Omega \approx \Gamma^*_{\mathrm{coh}}$"); correct form [T]
Bounds on $\sigma_{\mathrm{sys}}$	T-158 [T]	gap → [T]
Universal L2 architecture	T-159 [T]	gap → [T]
C27 attractor in window	from T-149	[C] → [T]
T-136 SAD spectral	from T-150	[T under C] → [T]
[H]-93—100	reclassification	[H] → cat. A/B

Total: 15 closures, 12 new theorems [T], 0 new open questions.

---

Related documents:

• Operationalization of consciousness — theorems T-128–T-138: formalization of operational aspects
• Operational closure — theorems T-139–T-147: closure of operational gaps
• Interiority hierarchy — levels L0–L4 and connection to SAD