Daniel C. Dennett (1981), “True Believers: The Intentional Strategy and Why It Works”, in A. F. Heath (ed.), Scientific Explanation: Papers Based on Herbert Spencer Lectures Given in the University of Oxford, Oxford: Clarendon Press, 1981. Revised as chapter 2 of The Intentional Stance (MIT Press, 1987; LIT-430), and reprinted in J. Haugeland (ed.), Mind Design II (MIT Press, 1997), pp. 57–80, which is the text read here.
On the citation. The brief’s citation (1981, Heath (ed.), Scientific Explanation, Oxford; reprinted in The Intentional Stance) is right as far as it could be checked: the volume, editor, publisher and year are confirmed by its Open Library record, and the reprint by the 1987 book’s catalogue contents. Three things to add:
- Pages. The 1981 page range could not be confirmed; the 1981 volume was not reachable and PhilPapers was blocked.
- Which text. The text read is the revised 1987 version, as reprinted in 1997. Its own citations postdate 1981. What changed between the printings is not known here.
- Identifier. The 1981 Clarendon volume has no DOI. The DOI here is
the Mind Design II chapter, the text actually read;
published:is the first appearance, 1981, by year only, since no source gives a month.
Key takeaways
- The thesis (p. 59). “While belief is a perfectly objective phenomenon (that apparently makes me a realist), it can be discerned only from the point of view of one who adopts a certain predictive strategy, and its existence can be confirmed only by an assessment of the success of that strategy (that apparently makes me an interpretationist).”
- The criterion (pp. 59, 72). “Any object … whose behavior is well predicted by this strategy is in the fullest sense of the word a believer. What it is to be a true believer is to be an intentional system, a system whose behavior is reliably and voluminously predictable via the intentional strategy.”
- Free choice, objective success (p. 67). “The decision to adopt the intentional stance is free, but the facts about the success or failure of the stance, were one to adopt it, are perfectly objective.” Dennett says his earlier presentations “sometimes invited” the radical interpretationist reading, and he “now want[s] to discourage it”.
- The lectern is excluded (p. 66) because ascribing it beliefs gives no predictive power we did not already have. People, animals and computers are included because often “the only strategy that is at all practical is the intentional strategy”.
- Martians (pp. 68–71). Laplacean super-physicists who predict Wall Street particle by particle, but do not see us as intentional systems, “would be missing something perfectly objective: the patterns in human behavior that are describable from the intentional stance, and only from that stance”. Where there are intelligent beings, “the patterns must be there to be described”.
- Thermostat to person (pp. 72–75). Enriching a thermostat’s links to the world shrinks the class of satisfactory interpretations until one is “practically (but never in principle)” dictated. “There is no magic moment in the transition from a simple thermostat to a system that really has an internal representation of the world.”
- Why it works (pp. 75–77). The easy answer is that evolution designed us to be rational. The hard answer, how the machinery works, “we just do not know”. The language of thought is a good bet, “what else could it be?”, but its rightness is “far from assured”.
Standing in the record
Filed on 2026-10-02 at the owner’s request for Dennett. It sits between two papers already here: “Intentional Systems” (LIT-440, read in NOTE-350), whose observer-relative definition it retracts in favour of objective success, and “Real Patterns” (LIT-220, read in NOTE-190), which turns this essay’s Martian moral, that the intentional patterns are real, into a compression criterion. NOTE-353 is the close reading, and placed it Active.
On groups. Its criterion is indifferent to what a system is made of, and it is generous: clams, plants and, “a little bit”, lightning are systems on which the strategy works (pp. 65–66), and the lectern is out only because the strategy buys no prediction. So on its own terms a nation or a corporation is a true believer exactly if treating it, as a unit, as a rational agent yields reliable and voluminous predictions that nothing cheaper yields. The essay’s examples of institutional prediction (a Cabinet that “would support his decision”, the consequences of letting Oxford colleges grant degrees, a stock market whose “societal sockets” absorb variation) are all reached through the reasons of individual people, and none ascribes a belief to the institution as such. Dennett does not say, in this essay, that a nation or a corporation is a true believer. He never raises the question.
That matters for THEORY-043, which says the readings on the society-of-mind bridge agree that agency survives in both directions. This is the criterion that makes the agreement cheap: on it, the question whether a group believes is the empirical question whether the group-level stance pays. List (LIT-401, read in NOTE-340) grants organised groups agency in a thin functionalist sense on independent grounds. It says nothing about experience, so it leaves the bridge’s dispute where that account puts it. Schwitzgebel (LIT-159, read in NOTE-131) answers in §6 a pragmatic objection Dennett sent him in correspondence; this essay’s insistence on predictive success “reliably and voluminously” is the nearest published ancestor of that objection, though the objection there concerns consciousness, not belief.
Other neighbours. Goldstein & Lederman (LIT-206, read in NOTE-138) state an “objective interpretationism” for language models that builds on Dennett; they deny thermostats beliefs, where this essay grants a thermostat “about half a dozen” deinterpreted ones (p. 72). Roberts (LIT-207, read in NOTE-157) argues the stance “does not go far enough” for chatbots, citing this book’s p. 49. Searle (LIT-400) answers McCarthy’s thermostat beliefs, which this essay recommends (n. 4), with “the mind is everywhere”. Levin (LIT-439) takes the stance’s level-choice rule from the 1987 book. Its bearing on the AI-consciousness account THEORY-023 is the same as the 1971 paper’s: belief ascription to a machine, however well earned, is not evidence of experience on the essay’s own terms. On Chalmers’s hard problem (LIT-406) it is silent.
It carries no instruction for machine-learning practice, and the anthology does not hold it. Its bearing on language models is through LIT-206 and LIT-207, which are filed here.