nucleation

LIT-1000: Does neural computation feel like something?

2 min read

Properties5
StatusActive — read 2026-10-14 (NOTE-818) in full; worth reading as the counterfactual problem for computational functionalism made concrete: a simulated voltage clamp replaying a recorded response stays idle on the matching stimulus yet removes every other transition, so on the Turing–Putnam definition the computation changes while no activity does. Read it knowing that the simulation shows only what an ideal clamp does, that the verdict on consciousness follows from a cited premise rather than from the model, and that it concerns consciousness in general, not affect.
Version2
Tagsconsciousness, neuroscience
Cited as “paper” byNOTE-818 — Gidon, Aru and Larkum — Does neural computation feel like something?
Cited as “source” byTHEORY-252 — On computational functionalism, an intervention that changes no neural activity and no actual causal interaction can change consciousness, because a clamp that stays idle on the actual input removes the transitions on every other input

Graph View

Albert Gidon, Jaan Aru and Matthew E. Larkum (2025), Frontiers in Neuroscience 19, 1511972.

Key takeaways

From the reading, NOTE-818.

  • The model. Five Hodgkin–Huxley-type neurons in a recurrent ring, driven by two LGN nodes: green light fires all five, red or blue only one. The recorded green response is replayed either by overwriting each neuron’s voltage or by a voltage clamp that acts only on deviation.
  • The clamp supplies only what is missing. With red light it injects exactly the absent LGN input. With green light it injects nothing: activity and actual neuron-to-neuron causation are as without it.
  • The computation still changes. With the clamp on, transitions no longer depend on the input, leaving “a trivial computation consisting of a single sequence of states (i.e., universally realizable)”.
  • Nothing type I and type II. Clamp off, and clamp on but idle, give identical activity. On computational functionalism they differ in consciousness, because only the second removes the counterfactuals. The idle clamp is a “counterfactual eraser”.
  • IIT agrees without paradox. The clamp degrades the transition probability matrix and so Φ; IIT rests on what a system is, not what it does. Φ is not computed.
  • The positive suggestion is thin. “Natural counterfactuals”, embodied in the substrate, are offered in one paragraph as what matters; the paper ends that the biological brain is “the only known substrate capable of consciously experiencing the world”.

Standing in the record

Filed on 2026-10-12 from the owner’s papers-feed for QUESTION-059, as “a simulation-based test” of whether meeting the functional role suffices for feeling. Read on 2026-10-14 in NOTE-818, which corrects that description: the simulation is an illustration, the argument is from the definition of computation, and the subject is consciousness in general, not affect. Its bearing on QUESTION-059 is indirect: it poses a dilemma for any identity of feeling with a functional role, including Veit’s (THEORY-101). It does not supply the case the question asks for. THEORY-252, filed from the reading, holds the conditional it establishes, with Block’s lookup machine (THEORY-092) and Frankfurt’s idle device (THEORY-119) as its neighbours.

Backlinks