The Two Axes: Scope and Mode¶

The four models arise from two orthogonal dimensions — scope (world vs. self) and mode (implicit vs. explicit) — and this 2x2 structure identifies the minimum number of model kinds any conscious system must maintain.

The Four-Model Theory does not posit four models as an empirical discovery about the brain's internal organization. Instead, it derives four model kinds from two independent dimensions that any self-simulating system must distinguish. The resulting 2x2 matrix is a principled minimum — a constraint that follows from the logic of self-simulation itself.

The Scope Axis: World vs. Self¶

The first dimension separates what the system models about its environment from what it models about itself. This distinction is not optional. A system that models only the world has no self-awareness; a system that models only itself has no perception. Consciousness as the theory defines it — ongoing self-simulation embedded in a world — requires both.

World scope covers everything external to the system: objects, other agents, spatial layout, causal regularities, physical laws as learned through experience. In the brain, world-scope modeling spans perception, semantic memory, spatial navigation, and causal reasoning.
Self scope covers the system itself: its body, its states, its history, its capabilities, its social identity. In the brain, self-scope modeling spans proprioception, interoception, autobiographical memory, personality, and the sense of agency.

The scope axis is a continuum, not a binary switch. Many real neural models blend world and self (a social interaction model encodes both other-knowledge and self-assessment), but the endpoints — pure world-knowledge and pure self-knowledge — define the axis.

The Mode Axis: Implicit vs. Explicit¶

The second dimension separates how the models exist in the system:

Implicit mode (learned, substrate-level): Information stored in the system's architecture — in the brain, synaptic weights, connectivity patterns, dendritic morphology. Implicit models are the accumulated product of lifetime learning. They are structural, persistent, and non-conscious. They operate "in the dark."
Explicit mode (generated, phenomenal): Information actively constructed as a running simulation — in the brain, transient patterns of electrochemical activity. Explicit models are generated dynamically from the implicit models and current sensory input. They are experience. "Lights on."

The mode axis captures a distinction familiar from computer science: the difference between data stored on disk (implicit) and data currently loaded into RAM and being processed (explicit). The analogy is imperfect — the brain is not a von Neumann architecture — but the structural principle holds: there is a categorical difference between information that exists as structure and information that exists as process.

Why These Two Axes?¶

The scope and mode axes are not arbitrary. They follow from the definition of consciousness as self-simulation:

Self-simulation requires a self-model. That model must be distinguished from the world-model it is embedded in. This forces the scope axis.
A simulation must be generated from something. The running process (explicit) must draw on stored knowledge (implicit). This forces the mode axis.
The two axes are orthogonal. Scope (what is modeled) is independent of mode (how it exists in the system). World-knowledge can be either implicit or explicit; self-knowledge can be either implicit or explicit.

Any system that collapses either axis — a system with no implicit/explicit distinction, or a system with no world/self distinction — lacks the architecture for consciousness as the theory defines it.

Why the 2x2 Is the Minimum¶

Two axes with two values each yield four quadrants. The claim is not that four is the right number of models but that it is the smallest number that satisfies the architectural requirements for self-simulation. A system with fewer than four model kinds is missing something essential:

Without an IWM: no accumulated world-knowledge to generate perception from.
Without an ISM: no accumulated self-knowledge to generate a self from.
Without an EWM: no conscious perception — the "lights" stay off for the world.
Without an ESM: no conscious self — experience with no subject.

The brain's actual modeling ecology is vastly richer — an effectively uncountable number of overlapping models that blend across both axes. The four canonical models are extremal points in a continuous space, not discrete boxes in the brain.

Figure¶

graph TB
    subgraph quad["The 2×2 Model Space"]
        direction TB

        subgraph implicit["MODE: Implicit (Learned, Substrate-Level)"]
            direction LR
            IWM["<b>IWM</b><br/>Implicit World Model<br/>━━━━━━━━━<br/>Synaptic weights<br/>Perceptual regularities<br/>Causal models<br/>Motor programs"]
            ISM["<b>ISM</b><br/>Implicit Self Model<br/>━━━━━━━━━<br/>Body schema<br/>Habits & skills<br/>Personality traits<br/>Autobiographical structures"]
        end

        subgraph explicit["MODE: Explicit (Generated, Phenomenal)"]
            direction LR
            EWM["<b>EWM</b><br/>Explicit World Model<br/>━━━━━━━━━<br/>Conscious perception<br/>Unified scene<br/>Transient, virtual<br/>🔆 Lights ON"]
            ESM["<b>ESM</b><br/>Explicit Self Model<br/>━━━━━━━━━<br/>Conscious self<br/>Unified narrative<br/>Transient, virtual<br/>🔆 Lights ON"]
        end
    end

    IWM -.->|"SCOPE: World"| EWM
    ISM -.->|"SCOPE: Self"| ESM

    style IWM fill:#4a6785,color:#fff
    style ISM fill:#4a6785,color:#fff
    style EWM fill:#c9a227,color:#000
    style ESM fill:#c9a227,color:#000
    style implicit fill:#2c3e50,color:#ecf0f1
    style explicit fill:#7d6608,color:#ecf0f1

Key Takeaway¶

The two axes — scope and mode — are not arbitrary classification dimensions but necessary consequences of what self-simulation requires. Any conscious system must distinguish world from self (scope) and stored knowledge from running simulation (mode). The resulting 2x2 structure defines the minimum architecture; real brains implement far more models, but no conscious system can have fewer than these four kinds.