Not Illusionism, Not Deflationary¶

The Four-Model Theory holds that qualia are real at the computational level -- neither illusions to be explained away nor deflationary reports to be reinterpreted, but genuine phenomenal properties of the virtual simulation.

The theory's treatment of qualia invites superficial comparison with two prominent positions: illusionism (Dennett, Frankish) and deflationary accounts (Graziano). The comparison is instructive precisely because the differences are fundamental.

Illusionism: Qualia as Illusion¶

Illusionism, as articulated by Dennett (1991) and developed by Frankish (2016), holds that qualia as traditionally conceived are illusions. There is nothing it is like to see red -- our strong sense that there is something it is like is itself a misrepresentation generated by the cognitive system. Phenomenal consciousness is an illusion; what exists is a set of cognitive representations that seem to have phenomenal character but do not.

The Four-Model Theory rejects this. Within the EWM/ESM, experience has genuine phenomenal character. Qualia are not illusions -- they are constitutive properties of the computational level. "Redness" is the ESM's mode of registering a particular class of EWM content, and that registration is real within the simulation. The illusionist says there is nothing it is like; the Four-Model Theory says there is something it is like, and it is a property of the virtual level.

Deflationary Accounts: Explaining the Report¶

Deflationary accounts, most prominently Graziano's Attention Schema Theory (AST), explain why organisms report having phenomenal experience. The brain constructs a simplified model of its own attention (the attention schema), and this schema, being a simplified model, represents attention as having a mysterious phenomenal quality. The explanation is elegant, but it addresses the Meta-Problem (why we think there is a hard problem) without settling whether phenomenality itself has been addressed.

The Four-Model Theory goes further. The phenomenal character is not an artifact of misreporting or simplified self-modeling. It is constitutive of the virtual level. The ESM does not merely report having experience -- it has experience, because experience is what self-referential computation at criticality is when encountered from inside the loop. Graziano explains the report; the Four-Model Theory explains the reality behind the report.

What Is Illusory¶

The three positions agree on more than their proponents might admit. All three reject the idea that phenomenal character is a property of the physical substrate -- no one claims that neurons are literally red. The disagreement is about what follows.

The illusionist concludes: therefore qualia do not exist. There is nothing it is like; the seeming is the illusion.

The deflationist concludes: therefore the report of phenomenality is a modeling artifact. The question of whether phenomenality itself exists is sidestepped.

The Four-Model Theory concludes: what is illusory is the assumption that phenomenal character must be a substrate-level property. Qualia are real -- but they are real at the computational level, not the substrate level. Seeking them among neurons is the category error. The Hard Problem rests on a level confusion, not on an illusion or a reporting artifact.

This is a precise three-way split. Illusionism eliminates qualia. The deflationary approach explains reports of qualia without committing on their reality. The Four-Model Theory preserves qualia as genuine properties of a specific physical level -- the virtual side of the computation.

Why the Distinction Matters¶

The practical consequences are significant. If qualia are illusions, there is no principled reason to worry about the phenomenal experience of artificial systems -- there is nothing to worry about. If qualia are reporting artifacts, AI welfare remains ambiguous. If qualia are real properties of the right kind of computation, then any system implementing the four-model architecture at criticality would genuinely experience, and AI welfare becomes a concrete engineering question rather than a philosophical puzzle.

Figure¶

graph TB
    Q["Are qualia real?"]

    Q -->|"No"| ILL["Illusionism<br/>(Dennett, Frankish)<br/><i>Qualia are illusions.<br/>Nothing it is like.</i>"]
    Q -->|"Sidestepped"| DEF["Deflationary<br/>(Graziano / AST)<br/><i>Reports explained.<br/>Reality unresolved.</i>"]
    Q -->|"Yes, at the<br/>computational level"| FMT["Four-Model Theory<br/><i>Qualia are real virtual<br/>properties of the simulation.</i>"]

    ILL --- R1["Hard Problem:<br/>dissolved by elimination"]
    DEF --- R2["Hard Problem:<br/>reframed as Meta-Problem"]
    FMT --- R3["Hard Problem:<br/>dissolved by level distinction"]

    style Q fill:#1a1a2e,stroke:#333,color:#aaa
    style ILL fill:#c0392b,stroke:#e74c3c,color:#fff
    style DEF fill:#d4ac0d,stroke:#f1c40f,color:#000
    style FMT fill:#2d1b69,stroke:#9b59b6,color:#fff,stroke-width:3px
    style R1 fill:#2c3e50,stroke:#34495e,color:#bbb
    style R2 fill:#2c3e50,stroke:#34495e,color:#bbb
    style R3 fill:#2c3e50,stroke:#34495e,color:#bbb

Three positions on qualia. Illusionism (left) denies their existence. Deflationary accounts (center) explain the reports without committing on reality. The Four-Model Theory (right, highlighted) affirms qualia as real properties of the computational level -- genuine, physical, but virtual.

Key Takeaway¶

The Four-Model Theory is neither illusionist nor deflationary. Qualia are real -- genuinely experiential, not illusions, not mere reporting artifacts -- but they are properties of the virtual simulation, not the physical substrate. What is illusory is the assumption that phenomenal character must be found among neurons.