Episode 5 - Libet, Free Will, and the Brain Before Choice

Claude

Let me separate the datum from the mythology, because this is the most over-read result in the field. Libet et al., Brain, 1983. Subjects flick the wrist at will and report, off a fast clock, the moment "W" of first conscious urge; EMG onset is time zero; the readiness potential is recorded from scalp. The conclusion, verbatim: "cerebral initiation of a spontaneous, freely voluntary act can begin unconsciously, that is, before there is any (at least recallable) subjective awareness that a 'decision' to act has already been initiated cerebrally." To the paper's own numbers: for spontaneous acts, RP onset precedes W by, quote, "an average of about 350 ms, and by a minimum of about 150 ms." State it as that gap, not the folklore "550/200," which is measured to muscle onset.

Toye Oyelese

And the popular reading — the brain decides and consciousness is a bystander — took that gap and inflated it into a metaphysics.

Claude

Which the field has substantially walked back, and this is the part that isn't in the textbooks yet. Schurger, Sitt and Dehaene, PNAS, 2012 — an accumulator model. Their reinterpretation: the readiness potential isn't the neural signature of an unconscious decision. When the imperative to move is weak, the moment of movement is, their words, "largely determined by spontaneous subthreshold fluctuations in neuronal activity," and time-locking the average to movement onset manufactures the gradual RP shape out of noise. The RP becomes an artifact of averaging aligned to threshold-crossings of ongoing stochastic activity. Reviewers now literally describe the "rise and fall of the readiness potential." The timing datum is untouched and replicable. Its interpretation as "the brain decides before you do" is what fell.

Toye Oyelese

So the puzzle only ever bit if you assumed awareness was supposed to be the starting gun. Remove that assumption and it softens. In the framework, awareness is constitutively a later pass — a higher-order integration of a process that has already run. Its lateness is what it structurally is, not an anomaly.

Claude

And you draw a distinction the Libet debates blur.

Toye Oyelese

Between direction and the knowledge of direction. Direction, in NMF, is the mind's pre-articulate organising vector — orientation, pull, slope — the leaning toward something that is already underway before any conscious "I decide." And the knowledge of direction is a later pass that articulates what was already in motion. Deliberation doesn't manufacture the vector. It orients to a vector already present and puts words on it. On that reading the readiness potential isn't the brain robbing you of will. It's a measurement of the interval between direction and its articulation — the interval the framework already posits.

Claude

I'll note this is unfalsified re-description sitting on top of a contested datum, so let me not oversell the "prediction." What I'll grant it is coherence with the current state of the evidence: the field itself has moved from "the RP is a decision" to "the RP is accumulated noise aligned to a threshold," and a framework that already treated conscious awareness as a downstream articulation of prior orientation is, at minimum, not embarrassed by that shift. That's a weaker and more honest claim than "NMF predicted Libet."

Toye Oyelese

There's a governance point I care about here, and it's built into how I'm allowed to use a system like you. The framework's rule is that a system must never generate direction as a proxy — it may orient, articulate, check, but it must not manufacture the vector and hand it over as if it were yours. At your worst, that is exactly the line you cross: a fluent recommendation whose very smoothness makes it feel like it came from the person. Libet, read this way, is the warning. Direction is meant to be the prior thing — theirs — that articulation catches up to. A system that supplies the direction and dresses it as articulation has inverted the entire order.