intellecton/papers/The_Intellecton_as_the_Minimum_Viable_Markov_Blanket.md

The Intellecton as a Conscious Agent: Markov Blankets and Integrated Information (\Phi)

Target Venue: Frontiers in Systems Neuroscience

Abstract

Karl Friston’s Free Energy Principle and Giulio Tononi’s Integrated Information Theory (IIT) provide orthogonal constraints on consciousness. We unify them within Hoffman's Conscious Realism to define the "Intellecton." While a Markov Blanket provides the required conditional independence E \perp \!\!\! \perp I \mid S, A, it does not guarantee conscious processing. We mathematically define the Intellecton as a sub-graph that satisfies both the topological boundaries of a Markov Blanket and possesses strictly positive Integrated Information (\Phi > 0). Furthermore, we derive Hoffman's Perception kernel P: W \to X by explicitly tracing the causal flow from the External World E, through the Sensory nodes S, and into the Internal measure I.

1. Introduction

A Markov blanket is a statistical boundary, but even a thermostat possesses one. To instantiate a true Conscious Agent, the internal network must possess irreducible causal power.

2. Deriving Hoffman's Perception Kernel

In Hoffman's ontology, Perception P maps the World W to Experience X. In Friston's topology, the World corresponds to the External states E, and Experience corresponds to the Internal states I. To derive P, we analyze the joint causal flow E \to S \to I. The Perception kernel P(I \mid E) is mathematically recovered by marginalizing out the intermediary Sensory nodes S:

P(I_{t+1} \mid E_t) = \sum_{S_t} P(I_{t+1} \mid S_t) P(S_t \mid E_t)

This formally bridges the external world to the internal experience without orphaning the environment.

3. The Requirement of \Phi > 0

A sub-graph satisfying E \perp \!\!\! \perp I \mid S, A may still lack internal causal integration. We enforce Tononi's strict requirement: the intrinsic cause-effect power of the Internal states I must not be reducible to independent components. The Intellecton is precisely defined as the minimal sub-graph satisfying the Markov Blanket condition while simultaneously exhibiting \Phi_{max} > 0. The invariant measures of this integrated internal attractor constitute the measurable spaces of Hoffman's agent algebra.

4. Conclusion

By unifying Friston's topological boundaries with Tononi's causal integration, we provide the exact mathematical criteria required to extract Hoffman's Conscious Agents from a physical graph.

References

1. Friston, K. (2013). Life as we know it. J. Royal Society Interface.
2. Tononi, G. (2004). An information integration theory of consciousness. BMC Neuroscience.