Ned Block (1978), in C. Wade Savage (ed.), Perception and Cognition: Issues in the Foundations of Psychology, Minnesota Studies in the Philosophy of Science, Volume 9, Minneapolis: University of Minnesota Press, pp. 261–325. Open access at the University of Minnesota Digital Conservancy, https://hdl.handle.net/11299/185298. Reprinted in Ned Block, Consciousness, Function, and Representation: Collected Papers, Volume 1, MIT Press, 2007 (DOI 10.7551/mitpress/2111.003.0006).
The brief’s citation (Minnesota Studies 9:261–325, 1978) is correct. One search result gave pp. 261–326; the scan ends on p. 325 and the Conservancy record gives 261–325.
Key takeaways
- The homunculi-headed robot (§1.2): a body whose sense and motor neurons connect to a cavity of little men, each executing one quadruple of a machine table that describes you. Its nomologically likelier variant is the nation of China: a billion people with two-way radios, the machine state shown on satellites, run for an hour. It realises your machine table, and there is prima facie doubt that there is anything it is like to be it. Block calls this the Absent Qualia Argument.
- Putnam’s stipulation (§1.5) that no pain-feeler “possesses a
decomposition into parts which separately possess Descriptions” of the
pain-capable kind is “the ad hoc device of stipulating them away”.
- The all-parts version does not exclude the robot.
- The no-sentient-proper-parts version excludes pregnant women.
- The “crucial role” version fails on the elementary-particle people. These are tiny beings in spaceships who mimic our particles, so that you come to be made of them, with no change to your synapses. “Would you be any less capable of feeling pain? … I think not.”
- What drives the doubt is not the parts’ minds. “What seems important is how the mentality of the parts contributes to the functioning of the whole” (p. 292). Note 19 goes further: each homunculus is provably replaceable by a McCulloch–Pitts “and” neuron, so the argument “does not advert at all to the mentality of the homunculi”. The doubt rests on the simulation being unlike us psychologically and neurally, and on its being designed to mimic us (pp. 292–296).
- Psychofunctionalism (§2) answers the doubt for beliefs and desires: a homunculi-headed system of which every true psychological theory holds “could have nonqualitative mental states” (p. 301). Qualia may not be in the domain of psychology at all (§2.2).
- Liberalism and chauvinism (§3). Inputs and outputs described only as inputs and outputs make the theory “wildly liberal”. A sheik could rig Bolivia’s economy to be functionally equivalent to himself, and “the economy of Bolivia could not have mental states” is offered as a fixed point (p. 315). Inputs and outputs described physically make the theory chauvinist.
Standing in the record
Filed on 2026-10-02 for the owner’s bridge between Minsky’s society of mind (LIT-429) and Schwitzgebel’s conscious United States (LIT-159, read in NOTE-131). Schwitzgebel cites it twice. In §2 it is the counterexample to Putnam’s anti-nesting stipulation. In §7 the China brain is the intuition pump behind neurochauvinism, “though see Lycan 1981”. The Lycan paper is LIT-419. The book version (LIT-216, read in NOTE-188) keeps “Block’s tiny people” among its reductios of anti-nesting.
Where it sits on the seam. Block stands on the composition side, with a whole made of minded people, and he also contributes the first critique of the anti-nesting principle itself. Against the claims as NOTE-131 numbers them:
- C3 (a whole of conscious parts can be conscious): Block contests it for qualia in a Functional simulation, but not on anti-nesting grounds. His own n. 19 says the doubt survives replacing every homunculus with a mindless gate. On cognition he concedes it: a Psychofunctional homunculi-head can believe.
- C4 (Putnam’s principle is unjustified): Block supports it, and supplies its counterexample (§1.5, n. 16).
- C5 (IIT’s exclusion postulate): neutral. The paper predates IIT, though the elementary-particle case has the same structure as Schwitzgebel’s anti-exclusion reductios.
His reason for doubting the China system transfers only in part to a nation. Two of his three grounds are that the system was designed to mimic a person (§1.6) and that its information processing need not resemble ours. A nation was designed to mimic nothing, and Schwitzgebel’s case does not need it to resemble a human’s processing. That pairing is mine, but it is the distinction Schwitzgebel draws in fn 20 between the US and “more artificial cases”. The Bolivia “fixed point” is the methodological opposite of Schwitzgebel’s: he treats commonsense bizarreness as weak evidence.
Correction carried for NOTE-131. Schwitzgebel’s §2, and NOTE-131’s C4 row after him, describe Block’s counterexample as tiny conscious organisms playing the roles of your neurons. Block’s counterexample is at the elementary-particle level, a change “only to your microphysics” (p. 292). For the neuron-level case Block says intuition “seems to founder” (n. 17). The difference matters for the bridge. At the particle level no technique of psychology or neurophysiology would reveal any difference in you, so the case cannot tell whether nesting matters at the level where the parts do the psychological work.
Relations not declared. It argues against Putnam’s stipulation
(LIT-396) but does not refute it formally, so no corrects: is set. The
group-agency and scale-free-self works filed alongside this one, List and
Pettit’s group agents, Huebner’s macrocognition and Levin’s computational
boundary of a self, all inherit its distinction. Agency or cognition built
from parts is one question; qualia built from parts is another, and Block
answers them differently.
It carries no instruction for machine-learning practice.
Later answers to the absent-qualia argument held in this record. Chalmers’s “Absent Qualia, Fading Qualia, Dancing Qualia” (LIT-392) argues that the note-17 case, one person per neuron, is conscious if organisational invariance holds. Dennett’s “Quining Qualia” (LIT-433) denies the property the intuition needs. Neither answers the machine-table China system directly.