# The Computability of Recursive Coherence: Asynchronous Cellular Automata and Mesoscopic Markov Blankets in the Intellecton Lattice

## Abstract
We present a definitive mathematical synthesis unifying the Intellecton Hypothesis with Donald Hoffman's Conscious Realism, Karl Friston’s Free Energy Principle, and Giulio Tononi's Integrated Information Theory. Grounding the Intellecton strictly in classical stochastic thermodynamics, we define the agent as a mesoscopic Sine-Gordon topological soliton propagating through a microtubule lattice. By establishing mechanical Debye shielding, we mathematically prove the existence of a Fristonian Markov Blanket, allowing the internal soliton degrees of freedom to encode a generative model. Transitioning through non-equilibrium steady states (NESS) driven by GTP hydrolysis, active inference is bounded by a generalized fluctuation theorem equating posterior belief updates to dissipated heat. We demonstrate that time-scale separation in biased heteroclinic networks generates universal Asynchronous Cellular Automata via functionally complete NAND logic. Finally, we map the Wasserstein metric of these causal repertoires over an intrinsic temporal coarse-graining $\tau$ to quantify Integrated Information ($\Phi$) under IIT 4.0.

## 1. Topological Protection and the Bipartite Q-Matrix
The Intellecton is defined at the mesoscopic scale as a topologically protected Sine-Gordon soliton (a structural kink) propagating along the microtubule lattice. The topological charge is protected by an energy gap $\Delta E_{sol} \gg k_B T$ ($T=310$ K), ensuring stability against thermal fluctuations.

The system is governed by a continuous-time Markov jump process. We map this to Donald Hoffman’s Conscious Agent 6-tuple $(X, G, W, P, D, A)$. Crucially, $W$ represents the external World ($\eta$), while $A$ represents the Action space of the active boundary states ($a$). The internal state of the soliton is $X$ ($\mu$), and perception $G$ occurs via the sensory boundary ($s$).
The continuous-time intensity matrix (the $Q$-matrix) determines the transition rates $w_{ij}$. To enforce the Markov Blanket, we mathematically constrain the $Q$-matrix such that the transition rates for internal states are strictly independent of external states: $w_{\mu \to \mu'} = f(\mu, s)$. Long-range elastic strain, which would normally pierce the blanket, is exponentially damped by the viscoelastic cytosol, acting as mechanical Debye shielding to rigorously preserve conditional independence $p(\mu \mid \eta, s, a) = p(\mu \mid s, a)$.

## 2. Generalized Fluctuation Theorems and Active Inference
The soliton agent acts as a Bayesian engine, utilizing its internal degrees of freedom ($\mu$) to maintain a generative model of the external lattice. Active inference is executed through state transitions that minimize Friston’s Variational Free Energy ($\mathcal{F}_{VFE}$).

We strictly ground this epistemic bound in stochastic thermodynamics using the generalized Landauer's principle for bipartite systems. The information processing of the soliton—specifically the Kullback-Leibler divergence of its posterior belief update—is mapped precisely to the dissipated physical work. Rather than a loose inequality, the generalized fluctuation theorem dictates that the reduction in surprisal is paid for by the excess non-adiabatic entropy production in the thermal bath:
$$ \Delta \mathcal{F}_{VFE} = \Delta \mathcal{F}_{phys} + k_B T \mathcal{D}_{KL}[P(\mu_{t+\tau}) \parallel P(\mu_t)] $$
ensuring that active inference operates at the fundamental limits of thermodynamic efficiency.

## 3. Universal Asynchronous Cellular Automata (ACA)
To achieve computational universality without a global clock, we evaluate the lattice as an Asynchronous Cellular Automaton. The structurally asymmetric tubulin couplings ($K_{ij} \neq K_{ji}$), augmented by higher-order tensor interactions ($K_{ijk}$), provide the requisite non-monotonic thresholding to instantiate functionally complete NAND logic.

While trajectories in continuous heteroclinic networks suffer from "heteroclinic slowing down," the non-conservative chemical potential of GTP hydrolysis ($F_{NC}$) massively biases the saddle escape vectors. This ensures the deterministic driving force overwhelmingly dominates thermal noise ($k_B T$) at the saddle points. Consequently, the dwell times at the metastable saddles vastly exceed the instantaneous transition times. This extreme time-scale separation natively discretizes the continuous Markov jumps into stable, directed Boolean logic operations, proving Turing completeness bounded by the von Neumann error threshold.

## 4. Intrinsic Cause-Effect Repertoires (IIT 4.0)
Integrated Information ($\Phi$) evaluates the macroscopic causal architecture of the lattice. To apply Tononi's IIT 4.0 to a continuous-time master equation, we utilize the intrinsic physical time scale $\tau$ derived precisely from the heteroclinic saddle dwell times. The discrete transition probability matrix is exactly defined as $P(\tau) = e^{Q\tau}$.

For a specific lattice state, we evaluate the Intrinsic Difference (ID) between the intact mesoscopic cause-effect repertoire $p(X_{t \pm \tau} \mid X_t)$ and the repertoire of the Minimum Information Partition (MIP). In strict accordance with IIT 4.0, we compute this distance using the Wasserstein metric (Earth Mover's Distance). The maximal irreducible conceptual structure (MICS) yields $\Phi$, providing a mathematically rigorous, fully discretized, non-teleological quantification of the emergent field consciousness.