Contribution
A four-page reply that states, more plainly than Brainstorms did, what the intentional stance concedes. Stich had argued that Dennett’s two themes, eliminative materialism and the reconciliation of persons with science, pull apart: the stance can rescue belief and desire but not pain. Dennett accepts the diagnosis and answers it in two ways. For belief, he makes the idealisation explicit. For pain, he declines to give a theory and treats the choice among options as practical.
Key insight
The intentional stance protects our status as believers from science because being a believer is defined by an ideal we already know we only approximate. Nothing found inside heads can show we fail a standard we were never thought to meet. The price is that “person” is an idealisation too, and Dennett pays it. Pain gets no such protection, because nothing plays the stance’s role for it, so for pain he offers policy instead of theory.
Assumptions
- Stich’s summary is accurate (“I have no quarrel with Stich’s summary of my views”, p. 73). The reply argues only against Stich’s objections.
- Quine’s interdependence of ascriptions: evidence for one belief ascription is the best evidence against a contradictory one (p. 74).
- Pain is not a good theoretical entity in a mature psychology, “as I argue” (p. 75), referring to “Why You Can’t Make a Computer that Feels Pain” (Brainstorms ch. 11). Assumed here, argued there.
Key results
- Stich’s charges (pp. 65–73, as the reply summarises them):
- the evolutionary grounding of rationality is “simply confused”;
- the norm of full rationality would disqualify every natural believer;
- pains are not “legitimized” by the stance, so the eliminativist must deny that pain can truly be attributed, while the theory of persons needs states of consciousness. Stich: “Can the contradiction be resolved?”
- On evolution (pp. 73–74). Dennett admits the claim is left “open-ended”, calls this a deliberate refusal of “premature precision”, and points to the forthcoming “Three Kinds of Intentional Psychology”. His reliance is minimal: belief and desire structures, even malignant or suicidal ones, are understood as extensions of control systems “selected by nature because they are benign in the long run”.
- On full rationality (pp. 74–75). No one is fully rational, but irrationality cannot be stably described as believing p and believing q where p implies not-q, because the evidence for each ascription counts against the other. Researchers on human inference who ascribe contradictory beliefs uncautiously “will indeed be confused — not ‘simply’, but subtly”. Ordinary practice tolerates the slack (“he sort of believes that p”) at real cost, notably in judgments of responsibility. A norm-free naturalistic theory of belief is “not ruled out” but would revise our concepts and be acceptable only if it kept most stance-based ascriptions.
- The concession (p. 75). “The conception of ourselves as intentional systems is granted to be an unrealizable idealization”, and so is the concept of a person built on it, and it is this idealisation, not a realisable counterpart, that figures in ethics.
- On pain (pp. 75–76). Three options, with uses and costs:
- (A) bite the bullet: “curiously short-sighted”, but it reminds the questioner that something is “misbegotten about the question”;
- (B) reform: replace the incoherent concept, accepting counter-intuitive verdicts;
- (C) common sense: pain is like voices, holes (Lewis and Lewis) and haircuts, which no mature science defines either. But not all perplexity about pain is “merely philosophical”, so a theory is still needed. Let it find its terms “where it will”, and call it a theory of pain where that helps, for instance in disputes about treatments or animal pain. The choice is “in some measure a matter of policy”. The philosopher’s contribution ends once the options are clear: “philosophical medicine for the philosophical puzzles and science for the rest”.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Grounding belief in the intentional stance makes our status as believers and moral agents well-nigh invulnerable to scientific discovery | moderate | p. 73; follows from defining belief by predictive success under a stance, granted by Stich; “not ‘in principle’” (Martian puppets) |
| C2 | The evolutionary grounding of rationality is minimal: only that belief structures extend control systems selected as benign in the long run | assertion | pp. 73–74; deferred to “Three Kinds of Intentional Psychology” |
| C3 | Irrationality has no stable description as holding contradictory beliefs | moderate | p. 74; Quinean argument from the interdependence of evidence |
| C4 | The intentional system, and so the person, is an unrealisable idealisation, and it is the idealisation that ethics uses | assertion | p. 75; stated as a concession, not argued |
| C5 | No entirely satisfactory theory of pain is possible, so the choice among A, B and C is partly policy | weak–moderate | pp. 75–76; rests on the incoherence of the ordinary concept, argued in Brainstorms ch. 11, not here |
| C6 | There are no pains, yet people feel pain, and the rest of his observations dissolve the paradox | assertion | p. 76; the reply’s last sentence; it does not say how |
Concepts
- Intentional system — a system whose behaviour is predictable by assuming it believes and desires what it rationally should (Stich’s gloss, accepted). Here explicitly “an unrealizable idealization”.
- “Strictly speaking” — the register in which there are no pains (p. 76), set against the register of common sense in which people feel them. The reply’s resolution is to keep both registers and assign them to different purposes.
- Manifest image — Sellars’s term, used for pains as “distressingly familiar items” (p. 75).
Connections
- “Artificial Intelligence as Philosophy and as Psychology” (LIT-393,
unread). The reply never discusses it; Stich’s notice does (§2, pp.
66–67), and the record’s account of the essay comes from that section.
Checked here:
- every quotation LIT-393 gives (pp. 123–124 of Brainstorms) is in Stich’s text as LIT-393 gives it;
- Stich says AI organises “armies of such idiots” to “simulate” intelligent behaviour; LIT-393’s takeaways paraphrase it as “produce”;
- Stich’s own conclusion extends discharge to physiology: homunculi so stupid “that what they do could be done (not by a machine but) by a nerve” (p. 67);
- Stich calls Dennett’s case for the homunculus point “clear and compelling” (p. 67). The objections of §§3–4 concern only pain and intentional systems, so the AI essay is the one part of Brainstorms Stich endorses without reservation, and Dennett’s acceptance of the summary covers it only as part of the whole.
- “Intentional Systems” (LIT-440, NOTE-350). The 1971 rationality assumption is the norm Stich calls too demanding. The reply defends it by conceding it is an idealisation.
- “True Believers” (LIT-398, NOTE-353). The later full defence of the stance against the same kind of objection.
- “Quining Qualia” (LIT-433, NOTE-357). The pain essay’s method (drive “a wedge of contingency” between features of a concept, then impeach the concept) is the method “Quining Qualia” applies to qualia.
- Frankish 2016 (LIT-458, NOTE-384). Frankish’s §1.6 says “eliminativism” wrongly implies a wider rejection of folk psychology. Dennett in 1980 is that wider eliminativist, about beliefs and desires as theoretical entities as well as pains, and rescues belief by the stance; illusionism keeps the pain half and drops the rest.
- Taylor 2024 (LIT-461, NOTE-379). Stich’s charge of “prima facie contradiction” anticipates Block’s (1993), which Taylor answers. Option (C)‘s haircuts reappear in the CE line Taylor builds on (“as real as haircuts and dollars …”).
Bearing on the record
- LIT-393. Confirmed, not changed. Two refinements for its next revision: “produce” should be Stich’s “simulate”; and Dennett’s acceptance is of the summary as a whole, the AI essay not being discussed in the reply. LIT-393 stays Deferred: this text is not the essay.
- THEORY-043. Indirect. The reply’s concession that belief rests on an idealised rationality is part of why the intentional stance grants agency so cheaply, which THEORY-043 already says (“Why agency is cheap to agree on”). It says nothing about composition or groups.
- No ML instruction. Nothing belongs in the anthology.
Limitations
- A reply to a review. It answers objections and points elsewhere for argument (Brainstorms ch. 11; “Three Kinds of Intentional Psychology”).
- The pain paradox is left standing by design. Dennett ends where Stich found him, “trusting that the rest of my observations on the subject dissolve the air of paradox”. Whether they do is not argued.
- The evolutionary grounding is admittedly left vague (“one of those disconcerting bits of soft-pedalling of which Stich rightly accuses me”).
Open questions
- Is “there are no such things as pains, although of course people do feel pain” consistent once “pain” is read in two registers, or does it need the later real-pattern machinery (Taylor) or the illusionist’s separation of experiences from phenomenal properties (Frankish)?
- If irrationality cannot be stably described, what do experimental findings of inferential error measure? The reply says they find something other than “unproblematic cases of people believing contradictions”, without saying what.
Corrections
- No dossier existed, so there is no seeded skim to correct.
- Citation as requested. “Reply to Professor Stich”, Philosophical Books 21: 73–76 (1980), replying to Stich’s “Headaches”, a critical notice of Brainstorms: right. Issue 2, April 1980, from the draft’s reproduction of the journal heading.
- The Tufts catalogue misspells the title as “Reply to Professor Stitch”.
- No DOI. Crossref has no record for either piece (see the LIT’s history).