Experiment #106 · Thought experiment

Block's Blockhead

A giant lookup table passes the Turing test

Ned Block · 1981 · Philosophy of mind

First published: N. Block, "Psychologism and Behaviorism", *Philosophical Review* 90 (1981): 5–43.

A machine with a pre-recorded response for every possible conversation passes the Turing test. Does it think?

Block imagines a machine whose memory contains a pre-recorded sensible response for every possible sequence of conversational inputs (within a bounded length). The lookup table is astronomically large but finite. Such a Blockhead would pass the Turing test by simple retrieval, with no genuine reasoning. Block's argument: behavioural success is insufficient for mind; what matters is the *kind* of mechanism implementing the behaviour. The case is the canonical objection to behaviourist and Turing-style criteria of intelligence, and bears directly on contemporary discussions of large language models.

Formulation

Machine M with lookup table containing sensible response for every possible input string up to length n. M passes Turing test by retrieval. Intuition: M does not think. Conclusion: behavioural indistinguishability is insufficient for mind.

Compare this experiment

Dimensions Engaged

◎

Observer

Targets Observer · Agency: is the right test for mind behavioural, mechanistic, or something else?

⧉

Information

Engages Information · Granularity: is pure retrieval informationally equivalent to genuine reasoning?

Responses — How Schools Engage

Affirms / takes the bait 2

Dualism

Confirms what dualists already held: no functional system suffices for mind. Blockhead is the cleanest objection to functionalism.

Phenomenology

A clean illustration of why third-person behavioural criteria miss what is constitutive of mind: the lived intentional structure of cognition cannot be replaced by stored outputs.

Reframes the question 3

Naturalism

Functionalist naturalism: the right level of analysis is the cognitive architecture, not behavioural input-output alone. Blockhead's wrong architecture is what disqualifies it.

Structuralism

Mind requires the right computational structure, not just the right outputs. Blockhead's structure (pure retrieval) is the wrong kind.

Pragmatism

Mind is what does cognitive work in life; Blockhead's scale (impossibly large lookup) makes it irrelevant to actual cognitive practice. The intuition gap reflects practical infeasibility, not metaphysics.

Holds it inconclusive 1

Analytic Metaphysics / Logical Atomism

A canonical case in psychofunctionalism vs functionalism: machine-state functionalism distinguishes architectures, behavioural functionalism does not.