Source
- Frankish (2016), “Illusionism as a Theory of Consciousness”, Journal of Consciousness Studies 23(11–12), 11–39, LIT-458, read in full in NOTE-384 from the author’s eprint (pages are the eprint’s, 1–24).
- Dennett (1996), “Facing Backwards on the Problem of Consciousness”, JCS 3(1), 4–6, LIT-397, read in full in NOTE-346.
- Frankish (2025), “Closing the Cartesian Bureau de Change”, LIT-103, read in full in NOTE-143.
What was actually shown
Three options for an apparent anomaly (LIT-458 §1.1). Radical realism (real, needs new science), conservative realism (real, explicable with existing science) and illusionism (not real; explain the appearance).
Strong and weak illusionism (§1.2). A quasi-phenomenal property is a physical property, perhaps gerrymandered, that introspection typically misrepresents as phenomenal (p. 4). Strong illusionism says experiences have only these. Weak illusionism keeps real phenomenal properties that merely seem ineffable, intrinsic, private and infallibly known. Unless it can name a residue stronger than quasi-phenomenality yet still conservatively explicable, which Frankish “doubts”, only radical realism and strong illusionism remain (NOTE-384, C1).
The arguments (§2). A presumption of conservatism, stronger when the pressure comes from introspection. Epiphenomenalism threatens radical realism under causal closure. Reductive theories explain at most why subjects would judge their experience qualitative, “a conflation of phenomenality with the representation of phenomenality” (p. 13), and so are covertly illusionist. A debunking argument, taken from Chalmers 1996 without his endorsement: our phenomenal claims would be the same without phenomenal properties. And an abduction from anomalousness plus the uncheckability of introspection.
What it keeps (§§1.6–1.7). Conscious experiences exist in an inclusive sense, and are “like something” in a functional sense, as introspective representation; in that sense “there is something it is like to be a zombie” (p. 10). Everyday phenomenal talk can be read as talk within a fiction. The audience of the illusion is no inner observer but “numerous unintelligent subsystems” (§3.3).
The earlier form (LIT-397). Dennett lists what impresses anyone about their own consciousness (delight, distraction, foreboding, the limit on items held at once) and finds only functions and dispositions: “Subtract them away, and nothing is left beyond a weird conviction (in some people) that there is some ineffable residue” (p. 5). A fundamental posit needs independent evidence, as mass and charge had; without it, adding experience to physics is like adding cuteness (p. 6).
An application (LIT-103). If hedonic valence motivates “in virtue of its feel”, there is a private place where value is “presented to the organism, felt, and reacted to”, “a tiny one-dimensional Cartesian bureau de change”, and the hard problem “remains as hard as ever” (p. 20). An evolutionary account is complete only if “feel” is metaphor or cashed out functionally: “Any state will do, provided it has the right effects in the right context – and the effects are all that evolution is sensitive to”.
What could have come out otherwise. Someone could have named a stable weak-illusionist residue, or a reductive theory could have explained phenomenality rather than its representation. Frankish argues neither has happened; he does not show that neither can.
What this does not say
- That there are no experiences, or no pain. Illusionism denies phenomenal properties, not experiences in the inclusive sense. Dennett in 1980 held that “there are no such things as pains, although of course people do feel pain” (LIT-456, NOTE-371).
- That an illusionist theory exists. The article “will not propose a detailed illusionist theory” (p. 1). How introspection represents non-existent phenomenal properties, the core of the illusion problem by its own account, gets “preliminary remarks” (§3.4).
- That the weak view is refuted. The collapse argument is a challenge: it holds until someone names the residue (NOTE-384, Limitations).
- That the abduction has a weight. “If there is even a remote possibility that we are mistaken … there is a strong abductive inference” (p. 14) does not say how anomalousness is weighed against the cost of being wrong about what introspection presents as certain.
- That Dennett’s subtraction is shown. That nothing is left after the functions are subtracted is asserted from the author’s introspective catalogue (NOTE-346, C3), and the reply leaves open what independent ground for a fundamental posit would satisfy it.
- That Dennett is a strong illusionist on every reading. Taylor reads his view as realism about a real pattern, with qualia eliminativism peripheral (THEORY-563). The readings agree only if Taylor’s “consciousness” is Frankish’s inclusive sense.
- That Veit’s account of valence is wrong. The bureau-de-change charge is conditional. If “feel” is metaphor, Frankish has “no objection”; it identifies an ambiguity, not an error (NOTE-143, Limitations).
Connections
- Rival. The hard problem (THEORY-694) holds that experience is a further explanandum beyond the functions. This account holds that the further explanandum is the seeming. Chalmers’s meta-problem paper (LIT-425, THEORY-577) agrees with Frankish that only strong illusionism dissolves the hard problem, and rejects it by a Moorean argument that the “denying the data” objection of §3.1 anticipates.
- Qualia. Dennett’s “Quining Qualia” (THEORY-415) supplies the four marks weak illusionists deflate, and argues the narrower thesis that no property of experience has all four.
- Valence. THEORY-101 (Veit’s hedonic valence as the origin of consciousness) is the account LIT-103 charges with keeping a Cartesian feel; that THEORY records Veit’s concession that the positive account of how valence motivates is owed.
- AI. If what makes experience “like something” is introspective representation, the question for a machine is whether it has the introspective mechanisms, which sharpens rather than settles THEORY-023’s mimicry point. The application is the record’s, not Frankish’s.