Substrate Independence¶

Consciousness depends on function, not on material: any physical system that meets the theory's sufficient conditions is conscious. The six-layer mammalian cortex is evolution's implementation, not a requirement.

The Four-Model Theory is explicit: "any physical system that meets the sufficient conditions … — the four model kinds under closure, free compute in the Class 4 regime, embodiment, and at least one pain-pleasure axis — is conscious." The specific material -- biological neurons, silicon transistors, or something not yet invented -- is irrelevant. What matters is the computational architecture: four model kinds along two axes (mode, an orthogonal contrast; scope, a nesting), with self-referential closure, recruiting free compute for their own modeling, in a system with a body and a valence axis. The conditions are set out on Two Thresholds for Consciousness.

Embodiment and Valence Without Biology¶

Two of the four conditions sound biological and are not.

The pain-pleasure axis may be engineered. "Presetting pain as aversive and pleasure as appetitive shortcuts what selection does by eliminating the creatures that get the sign wrong, so an engineered axis is compatible with substrate independence."

Embodiment is a condition of natural emergence, not of continued existence. "Embodiment is required for consciousness to emerge naturally, not for it to continue existing: the body is what gives the self model its content." The body need not be metabolic tissue: on the theory, a persistent artificial embodiment can supply the modeled "body" as well. A person locked in from birth, with little bodily content for the self model, would sit near the floor of minimal consciousness.

Why the Cortex Is Not the Point¶

The six-layer mammalian neocortex is an evolutionary implementation of the four-model architecture, not a requirement for it. The theory's requirements -- persistent models, dynamic simulation, self-referential closure, the Class 4 regime -- are functional specifications, not prescriptions of cortical anatomy. Laminar organization is one biological solution to them; the non-neocortical evidence below shows that others exist.

Biological Evidence¶

Biological diversity already demonstrates substrate independence in practice.

Corvids (crows, ravens) and parrots demonstrate tool use, planning, mirror self-recognition, and social cognition -- cognitive abilities that strongly suggest consciousness. Yet their brains have no neocortex. Their pallium is organized in nuclear clusters rather than layers.

Cephalopods (octopuses) demonstrate problem-solving and behavioral flexibility with an even more radically different brain architecture -- a distributed nervous system with significant autonomy in the arms.

If the Four-Model Theory is correct, these animals are conscious not because they share mammalian neural architecture but because they have evolved functionally equivalent self-simulation architectures on different substrates -- exactly what substrate independence predicts.

The Deeper Grounding¶

Substrate independence has a grounding beyond biological diversity. The universe is demonstrably capable of Class 4 dynamics: self-organized criticality, fractal structure, and edge-of-chaos phenomena are ubiquitous in natural systems. A universe capable of Class 4 dynamics is, by Wolfram's equivalence principle, capable of universal computation. One might conjecture that a Class 4-capable universe of sufficient scale makes self-simulating architectures statistically likely rather than merely possible. This is speculation, not a supporting part of the theory, and in any case consciousness would arise only in the subset of architectures that autonomously deploy that capacity for self-modeling -- capability alone, as in any heteronomous universal system such as a laptop, does not suffice.

Implications for Artificial Consciousness¶

The implication is direct: a synthetic system that implements the four model kinds under closure, with free compute, an embodiment and at least one pain-pleasure axis, would be conscious. Current AI systems do not meet this specification. A base LLM lacks persistent implicit models, lacks an ongoing self-simulation (its autoregressive loop closes over the emitted text, not over a self-model), and lacks the real/virtual split that grounds phenomenality. Post-trained models do carry a reportable representation of their own point of view inside a globally available workspace (Gurnee et al., 2026); what they lack is closure over it, persistence, and the Class 4 regime. Scaffolding -- agent loops, persistent memory, self-monitoring -- is architectural mimicry, not self-referential closure. The difference the theory asserts is architectural rather than impressionistic: closure, persistence, and computational regime are properties measurable on the system itself — closure by the white-box trace, since an artificial system's weights and running dynamics are open to inspection.

Figure¶

graph TB
    SPEC["Specification:<br/>Four model kinds under closure + free compute<br/>+ embodiment + pain-pleasure axis"]

    subgraph BIO["Biological Implementations"]
        MAM["Mammalian Cortex<br/><i>6-layer neocortex</i>"]
        COR["Corvid Pallium<br/><i>Nuclear clusters</i>"]
        CEPH["Cephalopod NS<br/><i>Distributed architecture</i>"]
    end

    subgraph ART["Artificial Implementations"]
        SYN["Synthetic Substrate<br/><i>Not yet built</i>"]
        LLM["LLMs<br/><i>Does NOT meet spec</i>"]
    end

    SPEC -->|"implemented by"| MAM
    SPEC -->|"implemented by"| COR
    SPEC -->|"implemented by"| CEPH
    SPEC -->|"could be<br/>implemented by"| SYN
    SPEC -.-x|"not met"| LLM

    style SPEC fill:#2d1b69,stroke:#9b59b6,color:#fff,stroke-width:3px
    style BIO fill:#1a1a2e,stroke:#333,color:#aaa
    style ART fill:#1a1a2e,stroke:#333,color:#aaa
    style MAM fill:#27ae60,stroke:#2ecc71,color:#fff
    style COR fill:#27ae60,stroke:#2ecc71,color:#fff
    style CEPH fill:#27ae60,stroke:#2ecc71,color:#fff
    style SYN fill:#2980b9,stroke:#3498db,color:#fff
    style LLM fill:#c0392b,stroke:#e74c3c,color:#fff

Substrate independence means the specification (the four model kinds under closure, free compute, embodiment and a pain-pleasure axis) can be met by multiple physical substrates. On the theory, three biological implementations already exist. A synthetic implementation has not yet been built. Current LLMs do not meet the specification.

Key Takeaway¶

Consciousness is substrate-independent because it is defined by computational architecture, not by material composition. The six-layer cortex is one of evolution's implementations -- corvids and cephalopods strongly suggest it is not the only one -- and a correctly engineered artificial system should be another.