NOTE-143: Frankish — Closing the Cartesian Bureau de Change

Contribution

This is a short commentary in a special issue on Veit’s 2023 book. It adds a clear two-paradigm framing, Cartesian versus post-Cartesian, to the debate over evolutionary accounts of consciousness. It also names a specific failure mode, the “Cartesian bureau de change”. An account falls into it when it makes valence explain action by being felt, and that relocates the hard problem rather than dissolving it. The piece is the source of the challenge that Veit’s reply (LIT-105) takes up at length.

Key insight

If a state of hedonic valence motivates because of an intrinsic feel, then there is a private place where value is “presented to the organism, felt, and reacted to” (p. 20). Shrinking the Cartesian theatre to a one-dimensional exchange counter does not make it any less Cartesian. An evolutionary account can therefore only be complete if “feel” is either metaphorical 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” (p. 20).

Assumptions

• The two-paradigm taxonomy. The Cartesian paradigm (Block 1995, LIT-643, Chalmers 1996, Nagel 1974) holds that experiences have intrinsic, non-functional, irreducibly subjective feels (qualia). The post-Cartesian paradigm is qualia eliminativism or illusionism (Dennett 1988, 1991, 2005; Frankish 2017), on which experiences are functional states “fully characterized in terms of the causal role they play” (p. 18).
• The causal premise (p. 18). “anything that makes a causal difference is experimentally detectable and can be treated as a theoretical posit defined by its causal role”. It follows that anything irreducibly subjective “will slip the net of evolutionary explanation”. This is asserted, not argued.
• Introspection has no special authority. Introspective reports are data, not insight into the nature of consciousness: “consciousness is what our best empirical theory of consciousness says it is” (p. 18).
• Philosophers’ terms are not empirical concepts. “Qualia”, “phenomenal properties” and “phenomenal consciousness” were defined against function (Block 1995), so they cannot be revised the way empirical concepts can (p. 19).
• Incommensurability. The two paradigms cannot be combined (p. 20). This is stated as a conclusion but functions as a premise of the recommendation.

Key results

These are the positions argued, since this is a philosophical commentary.

• Placement (p. 19). Veit decentres introspection, focuses on function and seeks evolutionary explanations, which puts him “firmly” in the post-Cartesian paradigm.
• Terminological charge (pp. 19–20). Repurposing “qualia” invites misunderstanding and “accusations of failing to address the real problem”. If a neutral term is wanted, coin a new one; Frankish’s own proposal is “quasi-phenomenal properties” (Frankish 2016).
• Substantive charge (p. 20). Veit’s phrases are quoted from the book: “a commanding sensation” (74), “an imperative plus or minus ‘feel’” (78–79), “a psychologically real felt common currency” (84), “‘quasi-intrinsically’ motivating states” (85). Together with his talk of reducing the explanatory gap (87, 114), these suggest a hedonic quale that directly moves the organism. If so, “the hard problem remains as hard as ever”.
• Diagnosis (p. 20). Frankish suspects “a little fudging”, meaning rhetorical concessions to Cartesians. He cites Veit 2023: 100 (“it is precisely this subjective experiencing that does the functional work”) and 117 (“Things feel a certain way, because they have to feel that way to be functional”).
• Conditional remedy (p. 20). If “feel” is metaphorical, Frankish has “no objection”. But Veit should then give “a positive naturalistic account of how states of hedonic valence motivate action – an account of their nature and causal powers, the systems that consume them”.
• Note 4 (p. 21). The “hybrid materialist-functionalist identity thesis” is either incoherent, if its relatum is irreducibly subjective qualia, or it is simply functionalism in a broad sense, since fine-grained functions may need specific neural implementations (Dennett 2005).

Claims

id	claim	strength	support
C1	Under the Cartesian paradigm no evolutionary explanation of consciousness is possible, because anything with selective effects is functionally definable.	weak	informal argument in one paragraph (p. 18); the premise that causal efficacy implies functional definability is asserted
C2	Under the Cartesian paradigm, consensus on animal consciousness is unlikely, because extrapolating from humans depends on a contested metaphysics.	moderate	informal argument (p. 18), illustrated by the spread of candidate substrates (quantum effects, information structure, higher-order monitoring)
C3	Veit’s methodology places him in the post-Cartesian paradigm.	moderate	textual evidence: quotations from Veit 2023: xi–xiii (p. 19)
C4	Veit writes as if valence motivates in virtue of its felt quality.	moderate	four quoted phrases (p. 20); Frankish concedes that the scare quotes on “feel” may signal metaphor
C5	Using “qualia” for functional properties causes confusion; a new term is better.	weak	informal argument (pp. 19–20); no evidence that the confusion occurs
C6	The Cartesian and post-Cartesian paradigms are incommensurable.	weak	assertion (p. 20); rests on C1
C7	Veit is “fudging” to appeal to Cartesians.	weak	interpretive conjecture (“My suspicion is…”, p. 20)

Concepts

• Cartesian paradigm: the view that experiences have intrinsic feels, accessible only to their subject, generating the hard problem and the explanatory gap (Levine 1983). Named for Descartes via Dennett’s “Cartesian Theatre”.
• Post-Cartesian paradigm: qualia eliminativism or illusionism. Experiences are real but are “complexes of activated perceptual sensitivities and reactive dispositions” (p. 18). They are subjective, internal and private only in deflated senses.
• Illusionism (note 1): belief in qualia results from “an introspective illusion or distortion of some kind”.
• Cartesian bureau de change: a minimal Cartesian theatre in which intrinsic value (valence) is presented, felt and converted into action.
• Quasi-phenomenal properties: Frankish’s (2016) neutral term for whatever “what it’s like” talk actually tracks.
• Hard question (Dennett 1991, 2018): “What difference does consciousness make?”, the post-Cartesian replacement for the hard problem.

Connections

The account of consciousness held here is illusionism / qualia eliminativism: a deflationary, fully functionalist view on which access consciousness “is the whole story” (note 2). It is sceptical of phenomenal consciousness as the Cartesian defines it, but not of experience.

• LIT-096 (Nagel 1974) is named as a founder of the Cartesian paradigm. Its “something it is like” criterion is exactly what Frankish treats as a philosophers’ notion that should not be repurposed.
• LIT-185 (IIT): Frankish says introspective claims about qualia serve as “unrevisable axioms” in Integrated Information Theory (p. 19). NOTE-122’s reading of LIT-185 concerns IIT’s formal ill-definedness rather than its axioms’ epistemic status, so the two notes criticise different parts of IIT. They do not duplicate each other.
• LIT-105 (Veit’s reply) answers each charge. Veit accepts the demand for a positive account and offers a short “cybernetic” reading of the currency metaphor. He rejects the terminological advice, saying concepts are revisable and Cartesians should not hold “definitional rights”. On note 4, he defends leaving room for identity theory. He also turns the charge around, saying Frankish is the one “caught in Cartesian gravity”. Read the two together.
• LIT-120 (Bruckner, welfare subjects without sentience) and LIT-146 (Gruen, moral status of animals) both touch Frankish’s view indirectly. If Frankish is right, any sentience-based ground of moral status has to be restated functionally. The commentary does not draw this consequence.
• LIT-056 (Butlin et al., indicator properties for AI consciousness) takes computational functionalism as a working assumption. Frankish’s paradigm would make that assumption the whole story rather than a hedge. This is a reader’s connection; Frankish does not discuss AI.

Consciousness tag: justified. The piece is about the nature of consciousness and the metaphysics of qualia.

Bearing on the record

• It supplies a clean statement of the illusionist pole for any THEORY document in the record about what consciousness is, or about whether valence explains action through feel. It is the natural counterweight to biological-naturalist or IIT-based readings (LIT-135, skim only; LIT-185).
• The bureau-de-change argument applies to any proposal that grounds welfare or moral status in felt valence, including AI-welfare arguments. Anyone relying on “valence matters because it feels like something” should answer it.
• ML practice: none. There is nothing here for the Anthology of the SOTA.

Limitations

• It is a five-page commentary, so its arguments are compressed. The key premise (C1) and the incommensurability claim (C6) are asserted, not defended. Illusionism is taken as the correct baseline, not argued for here; the reader is referred to Frankish 2016/2017 and Dennett.
• The charge against Veit is explicitly conditional (“Perhaps Veit intends ‘feel’ to be understood in a metaphorical sense… If so, I have no objection”). The piece therefore identifies an ambiguity rather than a demonstrated error.
• It engages only Veit’s book (2023) and gives no detailed discussion of the book’s empirical content (“I won’t dive into the details here”, p. 17).

Open questions

• Can a functional account say how valence motivates, meaning its consumers, causal powers and architecture, in enough detail to meet Frankish’s demand? Veit’s reply (LIT-105) offers a one-paragraph cybernetic sketch and defers the rest to neuroeconomics.
• Is the dispute about the word “qualia” substantive or verbal? Frankish’s and Veit’s positive views of valence converge once “feel” is read functionally. What remains is conceptual engineering: whether to revise the term or replace it.

Corrections to the seeded skim

• The dossier guessed that Veit’s reply might be “Pathological Complexity and the Function of Consciousness in Nature: Part 2” (unverified). That is wrong. The reply is “On the Evolution, Science, and Metaphysics of Consciousness”, Adaptive Behavior 34(1):53–71 (LIT-105). “Pathological Complexity and the Function of Consciousness in Nature” is the special issue’s title. Its Part 1 introduction is Veit 2025b, Adaptive Behavior 33(3):165–168, as Veit’s own reference list gives it.
• The dossier leaves out the notes, and two of them carry argument. Note 2 says that for post-Cartesians “some form of access consciousness is the whole story” (p. 21), which makes the illusionism explicit. Note 4 presses a second charge of “fudging”, aimed at Veit’s “hybrid materialist-functionalist identity thesis”: what is identified with material-functional states, if not qualia? And why is the identity partly material rather than wholly functional? Veit answers note 4 directly in LIT-105, so it is not a side remark.
• Minor: the dossier’s page ranges are close but slightly off. “The Pull of Cartesian Gravity” runs pp. 19–20. The bureau-de-change section and the Conclusion are both on p. 20. The notes and references are on p. 21.
• Suggested extra tag, not a change of primary: metaphysics. The piece is about the metaphysics of consciousness (qualia, the hard problem, the explanatory gap, identity theses), and a reader browsing metaphysics would expect it. consciousness stays first.