Free will — Sapolsky, Harris, Faggin, and the trilogy's wager.

1. The question, stated plainly

Are you the author of your decisions, or only the experiencer of decisions your brain has already made before you became aware of them? Phrased that starkly, the question is not abstract. It is the question every conscious adult asks at some point about their own life. It is also the question modern neuroscience has spent fifty years investigating with surprising depth, and the question modern philosophy of mind continues to argue about with surprising heat. The literature is enormous; the positions are clear; the data are partial; the consequences for ethics, law, politics, and one's own posture toward one's own life are non-trivial.

For the purposes of this primer the question reduces to three sub-questions: (i) does the empirical record force the conclusion that free will is illusory? (ii) what does "free will" actually mean, and which versions of it are compatible with the physics we have? and (iii) is there a framework in which freedom is a fundamental feature of reality rather than an emergent and possibly illusory phenomenon? The trilogy's answer is yes to (iii), and this page is the case for that yes.

2. The case for no free will — Sapolsky, Harris, and the experimental record

The argument the determinists make is straightforward and forceful. Every action you have ever taken was produced by your brain. Your brain is a physical system. Physical systems are governed by physical laws. Therefore every action you have ever taken was governed by physical laws, working through prior causes you did not choose, on a substrate you did not design. The "you" who allegedly chose is a representation generated inside that same brain, after the relevant causal chains have already started. The choosing is narration, not causation.

Robert Sapolsky — Determined (2023)

Sapolsky's case is the most thorough recent statement. He works through the layered chain of causes that produce a given behaviour: the firing of a neuron one second before the action; the prior activity of the brain region in which that neuron sits, minutes before; the hormonal state of the body, hours before; the contents of breakfast, that morning; the developmental trajectory of the brain through adolescence; the in-utero hormonal environment; the genome inherited from parents; the cultural environment those parents were shaped by; the evolutionary history of the species. At every level the apparently free decision is the output of variables the decider did not select. Sapolsky's conclusion: free will, in any sense robust enough to ground moral responsibility, does not exist, and the consequences for criminal justice, education, and ethics are large.

Sam Harris — Free Will (2012)

Harris makes the same case more compactly and philosophically. Either decisions are determined by prior causes (in which case the chooser is just where the causes pass through) or they involve random quantum noise (in which case the chooser is not in control either). Compatibilist rescue moves, in his analysis, redefine "freedom" so broadly that they no longer match anyone's intuitive picture of what they meant by free will. The phenomenology of choice is robust but, on Harris's reading, simply wrong about what is happening underneath. He concludes that abandoning the illusion of authorship would, paradoxically, make us more compassionate.

The experimental record the determinists lean on

Benjamin Libet (1983) — the readiness-potential experiment. Subjects watch a clock and report the moment they consciously decide to flex a finger. EEG electrodes record cortical activity. The result: a readiness potential (a slow rise of negative cortical voltage in the supplementary motor area) appears about 350 ms before the subject reports the conscious decision, and roughly 550 ms before the action itself. The brain has begun preparing the movement before the conscious mind reports deciding on it. Libet's own interpretation was carefully agnostic; popular interpretations have been more decisive.

Soon, Brass, Heinze, Haynes (2008) — the fMRI prediction study. Using machine-learning analysis of fMRI activity in the prefrontal and parietal cortex, the researchers were able to predict whether a subject would press the left or the right button up to 10 seconds before the subject reported deciding. The prediction accuracy was modest (around 60% on a 50% chance baseline, climbing higher closer to the decision moment) but the temporal lead was striking.

These two experiments are the empirical centerpieces of the no-free-will case as it is presented to the general reader. They are real, replicated, and important. They are also, as we will see, less decisive than they often sound.

3. The case for free will — the responses, honestly

The determinists have not had the field to themselves. The responses fall into several families.

Compatibilism — Dennett, Frankfurt, the mainstream-philosophy answer

Daniel Dennett (most clearly in Freedom Evolves, 2003, and Elbow Room, 1984) and Harry Frankfurt argue that free will and determinism are compatible: free will, properly understood, does not require violating physical law. It requires acting on one's own reasons, values, and preferences, without external coercion, in a way that is responsive to deliberation and could be different under different reasons. On this view, a deterministic brain that does its choosing through deliberation, responsive to argument and evidence, is exactly a free brain. The compatibilist's free will is the kind worth wanting; the libertarian's "uncaused choice" is incoherent (a choice with no cause is not a choice). Most contemporary philosophers of mind, including most who take the empirical record seriously, are compatibilists.

Harris's objection that compatibilism redefines the term too liberally is taken seriously; the compatibilist response is that ordinary moral and legal practice has always operated with the compatibilist sense, not the libertarian one, and that the libertarian sense was never the one anyone meant.

Critique of Libet

The Libet experiment looks decisive in popular summary; it is much less decisive in the technical literature. Several lines of critique:

The readiness potential is statistical, not deterministic. Aaron Schurger's work (2012 onward) reanalysed Libet-style data and showed that the readiness potential can be modeled as spontaneous noise that crosses a decision threshold, not as a fixed precursor of an inevitable action. On Schurger's reading, the readiness potential is the brain's accumulation of noise-plus-evidence; whether the noise crosses the threshold depends on the conscious decision-making process, not the other way around.

Veto power remains. Even Libet himself noted that subjects could veto a movement after the readiness potential began but before the action; this preserves a window of conscious control that the popular framing tends to omit.

The decisions are trivial. Libet's subjects were deciding when to flex a finger — not whether to keep a marriage, take a job, or release a technology that could end scarcity. Generalizing from finger-flicks to deliberative life-choices is a large extrapolation, and the experimental literature on deliberative decision-making is far less clean than the urgency-of-the-popular-summary suggests.

Critique of Soon

The 10-second-prediction headline of Soon et al. (2008) is striking but the underlying numbers are modest. Prediction accuracy was around 60% on a binary task (50% chance); the lead time of 10 seconds is the upper bound, with accuracy near chance at that distance and rising as the decision approached. The task was again simple (left button vs right button, no stakes). What the study shows convincingly is that some preparatory neural activity precedes conscious awareness for trivial decisions; what it does not show is that all deliberative life-choices are predetermined ten seconds in advance.

Libet's other experiments — the sensory studies and subjective antedating

The discussion above concerns Libet's motor experiments — readiness potential before the conscious decision to flex a finger. Almost every popular treatment of Libet ends there. The argument is incomplete without his sensory experiments, which point in the opposite direction and which Stuart Hameroff and others have argued are the more philosophically consequential half of Libet's work.

The setup was clinical. Libet had access (with their consent) to patients undergoing awake neurosurgery for other conditions, with electrodes placed in the somatosensory cortex while the cortex was exposed. He could stimulate the skin (a finger, say), stimulate the corresponding region of cortex, record from either, and converse with the patient throughout. The headline findings (Libet et al., Subjective referral of the timing for a conscious sensory experience, Brain 102, 1979; extended in Libet's Mind Time, 2004):

• Peripheral stimulation reaches conscious awareness almost immediately. A pulse delivered to the finger produces conscious awareness of being touched at roughly 30 milliseconds, the time it takes the signal to travel up the nerve and into cortex.
• Direct cortical stimulation requires ~500 milliseconds of continuous activity. Stimulating the same cortical region with brief pulses below the duration threshold produces no conscious experience at all. Only sustained cortical activity for around half a second yields a reported sensation.
• But the reported timing is antedated. When the patient was asked to compare the timing of cortical-stimulation events with peripheral-stimulation events, the reported subjective timing of the cortical-stimulation experience was backdated to roughly the moment cortical activity began, not to the moment 500 ms later when conscious awareness was actually produced.

Libet's interpretation, defended for the rest of his career, was that conscious experience involves a subjective referral of timing backward — that the brain delivers a conscious percept whose felt time-tag is not the time the percept was completed but an earlier time the brain has reconstructed as the time the percept "is about." The conscious present, on this reading, is not a real-time read-out but a backward-tagged construction. The 30-millisecond felt awareness of a finger-touch happens, in actual brain time, at least 500 milliseconds after the touch — but is experienced as having happened at the moment of touch.

Hameroff's argument, building on Libet and consonant with the Penrose-Hameroff Orch-OR framework: if conscious experience involves backward referral in time, the determinist objection that decisions are made by neurons before conscious awareness loses its force. The conscious decision's felt time of occurrence may not be the same as its brain-time completion. Free will operating "after" the readiness potential, on Libet's own subjective-antedating mechanism, is operating on a timeline the brain itself reconstructs — and the reconstruction is part of what consciousness does.

The reception is contested. Daniel Dennett and Patricia Churchland argued against Libet's interpretation in print across the 1990s and 2000s; Libet defended it in Mind Time and in his responses. The subjective-antedating phenomenon itself is not in serious dispute — the data are reproducible. The interpretation is. Two further claims in this neighbourhood require careful flagging:

• Daryl Bem's "Feeling the Future" (2011). Bem published in the Journal of Personality and Social Psychology a series of experiments claiming small but statistically significant precognitive effects on emotional responses — subjects' physiological responses appearing to anticipate randomly-selected emotional images. The paper attracted enormous attention and became a centerpiece of the replication-crisis discussion. Subsequent replication attempts have largely failed; Bem's original effects have not survived rigorous pre-registered replication. The literature is one of the cleanest examples of why pre-registration matters. The trilogy's framework does not depend on Bem being right, and the honest summary is that Bem's specific claims are now considered substantially refuted by failed replications.
• Penrose-Hameroff Orch-OR backward-time effects. Hameroff has argued, building on Penrose's gravity-collapse proposal, that quantum collapse events can have time-symmetric components that allow effective backward-in-time signalling within the brain — rescuing conscious agency by letting the result of a conscious decision propagate backward to the moment of choice. The proposal is consistent with the time-symmetric formalisms of quantum mechanics (the Two-State Vector Formalism, the transactional interpretation; see the arrow-of-time essay and simulation-hypothesis entry #E on the delayed-choice quantum eraser). It is not consensus. The time-symmetric formalisms are mainstream; their application to consciousness and free will is the contested move.

What the sensory experiments do, honestly: they establish that the conscious experience of when an event happened is itself constructed, not transparent. The brain's reconstruction of "when" is part of what consciousness produces, not a passive registration of physical time. The standard determinist case against free will depends on assuming that the brain's clock is the felt-self's clock — that the readiness potential at −350 ms means the decision was made before the conscious self had time to participate. Libet's sensory experiments suggest the assumption is more interesting than it looks. The conscious self's timeline is reconstructed by the same processes the determinist's argument runs on. The argument's intuitive force is weaker once one realises both clocks belong to the same system, and the brain is the timekeeper of both.

The trilogy's framework engages this directly. The Anima edge-case file's pre-event readiness phenomena, Lucía Reyes's cymatic pre-event window, the delayed-choice quantum eraser's retroactive determination of past photon behavior, the Two-State Vector Formalism's bidirectional time evolution — all sit naturally inside a picture in which the temporal order of conscious experience is partly constructed rather than fully read off the external clock. The receiver model does not require backward-time signalling to make free will defensible (sections 5 and 6 of this essay handle that work via Faggin's irreducibility framework). It does, however, treat the Libet sensory experiments as confirmation that the assumption "conscious time = brain time" is the assumption that needs questioning, not the alternative.

Libertarian responses — the minority view that takes the strong claim

A smaller group of philosophers and physicists hold that genuine libertarian free will exists and requires something more than compatibilist redescription. Roger Penrose's gravity-collapse proposal (in The Emperor's New Mind, 1989, and Shadows of the Mind, 1994) is the most-developed physics-grounded version: if objective-collapse models like Penrose-Diósi are correct (see the measurement-problem essay), then quantum events in the brain are genuinely indeterministic, and consciousness might exploit that indeterminacy to produce genuinely novel choice. The position is testable in principle and currently underdetermined empirically. It is also exactly the kind of opening the trilogy's framework finds it natural to inhabit.

Penrose's deeper conceptual move — less discussed than the gravity-collapse mechanism, and more important — is his reframing of what free will actually is. In his own formulation across multiple interviews and lectures:

"Free will is not randomness. So what is it? Maybe… you're free to do something which may be very well-determined. You've used your consciousness as something to employ in making your decision. So that's what free will is."

The move is to refuse the dichotomy that drives Harris's argument. Harris's case for the impossibility of free will rests on the trilemma: either decisions are determined (no freedom), or they are random (no chooser), or they involve some additional non-physical agent (no naturalism). Penrose rejects the second horn. Free will is not the absence of determination. Free will is the use of consciousness in the decision. A decision can be highly determined — even fully determined, in the limit — and still be free, provided that what determines it is consciousness operating on the situation. The libertarian's strong claim is not that the decision is uncaused. It is that the cause is consciousness itself, treated as something that can be employed in making the choice, rather than something the chooser is a passive vehicle of.

This reframing matters because it dissolves the apparent gap between determinism and free will once consciousness is treated as a real agent in the causal story rather than as a downstream report of physical processes. Two consequences follow:

• The randomness objection is the wrong objection. Harris and Sapolsky both argue that adding quantum randomness to deterministic physics does not produce freedom; it produces noise. Penrose agrees. The question is not whether the decision is random. The question is whether consciousness is one of the things that determines it.
• The agent question takes priority over the mechanism question. Penrose's gravity-collapse proposal is one way of getting consciousness into the causal story (by giving the quantum measurement process a site where conscious participation becomes physically meaningful). It is not the only way. Faggin's irreducibility framework (section 5 below) gets there from a different angle. The receiver model the trilogy is built on takes consciousness as fundamental, which means it is causally efficacious in the same way any fundamental physical feature is. The mechanism is open; the agent is the point.

Stated cleanly, Penrose's position is: free will is what happens when consciousness is what determines the decision. The determinist who says "but consciousness is just brain activity, and brain activity is determined" is begging the question against the libertarian, who is denying precisely the claim that consciousness reduces to brain activity. The dispute is not about determinism. The dispute is about whether consciousness is the kind of thing that can determine anything — whether it has causal standing in its own right, or whether it is a representation the brain produces about decisions the brain has already made on its own. Penrose says it has causal standing. Faggin says it has causal standing. The trilogy's receiver model says it has causal standing. The Sapolsky-Harris position says it does not. The argument is at the level of what consciousness is, not at the level of whether deterministic physics permits unmoved movers.

This is also why the trilogy ends up needing voluntarism politically. If free will is consciousness employed in the decision, and if consciousness is a fundamental field property (the receiver model's claim), then institutional arrangements that prevent consciousness from acting according to its nature are not merely inefficient or unjust. They are working against the architecture of the substrate. The political reading of free will and the metaphysical reading of free will are the same reading at different scales. The voluntarist wager Fragile Light stages is the same wager Penrose, Faggin, and the trilogy make at the philosophical level: consciousness is what the universe does to decide, and arrangements that block it are arrangements that block the universe from doing what it does.

Watch — Penrose and Hameroff in their own voices

Roger Penrose on Mind & Consciousness · Closer To Truth Chats ↗

Penrose in conversation with Robert Lawrence Kuhn on Closer To Truth, the long-running interview series on foundations of physics, cosmology, and consciousness. Penrose's own account of why consciousness cannot be a classical computation, why his gravity-collapse proposal locates the physical site of consciousness in objective reduction, and where his thinking has moved across the decades since The Emperor's New Mind.

Stuart Hameroff · Is Consciousness a Quantum Orchestra? Microtubules, Objective Reduction, & Orch OR ↗

Hameroff (Professor of Anesthesiology and Psychology, University of Arizona; Penrose's long-time collaborator on Orch-OR) walking through the proposal directly: microtubule quantum coherence as the physical site of consciousness, the Libet sensory experiments and their implications for backward referral of subjective experience, and how the Orch-OR framework reads the free-will question. The Bandyopadhyay terahertz-coherence experimental probe is, in the trilogy's vocabulary, the closest current evidence for what this framework predicts; see the Bandyopadhyay companion page for the experimental side.

The phenomenology of agency

A separate strand of response, less from physics and more from philosophy of mind: the felt experience of deliberating, weighing options, and choosing is so robust and so universal that dismissing it as "illusion" requires accounting for why an illusion this systematic exists, persists across cultures, and grounds the entire practice of moral discourse. Phenomenologists (Husserl, Merleau-Ponty, and the contemporary enactivists) argue that the lived experience of agency is itself evidence about the structure of consciousness — not conclusive evidence, but not nothing either.

4. What the determinist case actually shows — and what it doesn't

The honest summary of the experimental record: the brain begins preparing actions before the conscious mind reports deciding; the felt moment of decision is not the originating cause of the action; the apparatus of choosing is layered through neural, hormonal, developmental, and evolutionary processes the chooser did not select. This is real and consequential. It does not entail that no free will exists. It entails that the naive picture — a Cartesian "self" outside the brain, reaching down to pull the levers — is wrong. Almost no serious participant in the debate holds the Cartesian view. The question is what to make of the architecture once the Cartesian self is gone.

The determinist's answer: nothing free remains; the felt self is post-hoc narration. The compatibilist's answer: what's left, when the Cartesian self is gone, is exactly what we should always have meant by freedom — deliberation, responsiveness to reasons, the absence of external coercion. The libertarian's answer: there is room in the physics, especially in the under-determined parts (objective collapse, quantum noise, computational irreducibility, the measurement problem) for something genuinely undetermined-from-below that consciousness participates in. The trilogy's answer adds one more move: consciousness itself is fundamental, and freedom is part of what consciousness is, not a feature consciousness has.

5. Faggin's framework — consciousness, love, and freedom as irreducible

Federico Faggin (designer of the first commercial microprocessor; later, with Giacomo Mauro D'Ariano, the architect of a quantum-information approach to consciousness) has argued across his 2021 book Irreducible and the body of work catalogued on the D'Ariano-Faggin companion page that consciousness is not produced by computation, is not an emergent property of complex information processing, and cannot in principle be generated by any classical or quantum substrate. Consciousness is, on his account, ontologically irreducible: a fundamental feature of reality that substrates can receive, decode, and localize, but that no substrate produces from scratch.

Faggin's deeper move — the one that matters here — is that he treats love and free will as co-fundamental with consciousness. Consciousness, on his framework, is not a passive registering of states. It is intrinsically agentive (capable of free choice) and intrinsically valuing (capable of love). These are not three separate fundamental properties; they are three faces of the same irreducible feature. To be conscious, in Faggin's sense, is already to be free and already to love — the field's own structure carries all three.

This is the move the trilogy makes, in its own vocabulary. The receiver model says consciousness is a field property; the trilogy adds that what we call freedom and what we call love are the field's organizing principles. A receiver is not a passive antenna picking up a fixed broadcast. It is an active participant in the field's localization — choosing, valuing, and coupling. The choice is real because the substrate is constituted that way; the love is real for the same reason. Neither emerges from neurochemistry. Both are what the substrate is.

Faggin's position is not consensus science. It is a serious-minority position held by a small group of researchers in the foundations of physics and consciousness studies, with technical work that engages quantum information theory at a high level (D'Ariano's reconstruction of quantum mechanics from informational principles, for example). The trilogy takes it seriously because it is the cleanest current statement of an architecture in which freedom is not an embarrassment for the physicist but the natural reading of the data.

6. The trilogy's voluntarist wager

The trilogy lives in this position and makes it political. Fragile Light's central question is what happens when a single conscious being (Luz Paz) is given a technology that could end scarcity on Earth and faces an existential choice: release it to the world, or accept institutional containment of it? The wager the book makes is that voluntarism — the political philosophy that power is made safe only by structures preventing any one party from holding too much of it — is the working ethics of any civilization that reaches the threshold the book describes. Freedom is not a luxury that emerges when a society is rich enough to afford it. Freedom is what a consciousness is; the question is whether the institutional arrangements allow it to act according to its nature.

The alien civilization in Fragile Light exists in the book precisely as evidence that the wager can be survived. They went through their own version of the same choice. They lost several hundred million people. They survived. The scientist who released their code — their Luz Paz, the one "who opens" — was killed by the governing authorities before the code had fully propagated. Freedom did not lose, the alien voice tells her. Freedom was interrupted. The voluntarist wager, the trilogy is saying, is not optimistic. It is honest about the cost. It also remains the only wager available to a consciousness that takes its own freedom seriously.

The Aquinas definition of love (amor est velle aliquid bonum alicui — love is to will the good for another) runs through the trilogy alongside the voluntarist thread. Love and freedom are not in tension in the books; they are the same architectural fact viewed from two angles. To love is to will another's good; willing requires freedom; freedom is what a conscious receiver is. The trilogy treats Aquinas's definition not as a piece of theology but as a working clinical statement about what consciousness does, when it is doing the thing consciousness is for.

7. Where this leaves the reader

The empirical record on free will is real and important. The popular determinist case (Sapolsky, Harris) is honest as far as it goes, and the experiments it cites are sound. The case is also less decisive than the popular framing suggests, the compatibilist response is held by most professional philosophers of mind, and the libertarian options — especially those that take Faggin's irreducibility seriously, or Penrose's objective-collapse proposal — remain live. The reader has to choose.

The trilogy's case for choosing in favour of freedom is not a denial of the data. It is a different reading of where the data leave the question. If consciousness is fundamental rather than produced, and if the receiver model has the architecture we have been catalogueing across these companion pages, then freedom and love are not features the substrate happens to develop but features of what the substrate is. The neuroscience does not refute this; it cannot refute it from inside its own paradigm, because the determinist reading and the receiver-model reading agree on every measurement and disagree only on what the measurements are about. The choice is metaphysical. The trilogy makes it openly, and lives inside the consequences.

What the reader does with this is, in the most concrete sense the books care about, up to the reader. Which is, on the trilogy's reading, the only honest place to land.

Reading list

The determinist case

Robert Sapolsky, Determined: A Science of Life Without Free Will (Penguin, 2023). The most thorough recent statement.

Sam Harris, Free Will (Free Press, 2012). The compact philosophical version.

Benjamin Libet et al., Time of conscious intention to act in relation to onset of cerebral activity (readiness-potential), Brain 106 (1983): 623–642.

Chun Siong Soon, Marcel Brass, Hans-Jochen Heinze, John-Dylan Haynes, Unconscious determinants of free decisions in the human brain, Nature Neuroscience 11 (2008): 543–545.

Compatibilist response

Daniel Dennett, Elbow Room: The Varieties of Free Will Worth Wanting (MIT Press, 1984), and Freedom Evolves (Viking, 2003).

Harry Frankfurt, Alternate Possibilities and Moral Responsibility, Journal of Philosophy 66 (1969).

Critique of the experimental record

Aaron Schurger, Jacobo D. Sitt, Stanislas Dehaene, An accumulator model for spontaneous neural activity prior to self-initiated movement, PNAS 109 (2012): E2904–E2913. The reanalysis that reframes the readiness potential as accumulating noise.

Libertarian / physics-grounded responses

Roger Penrose, The Emperor's New Mind (Oxford, 1989), and Shadows of the Mind (Oxford, 1994). The gravity-collapse proposal as a site of genuine indeterminacy.

The Faggin / D'Ariano framework

Federico Faggin, Irreducible: Consciousness, Life, Computers, and Human Nature (Essentia Foundation, 2021). The book-length statement.

Giacomo Mauro D'Ariano & Federico Faggin, Hard Problem and Free Will: An Information-Theoretical Approach, in Artificial Intelligence Versus Natural Intelligence (Springer, 2022). The technical paper. See also the D'Ariano-Faggin companion page.

This page is part of the Reading companion essays. For the consciousness-as-fundamental position the trilogy's wager rests on, see The hard problem, re-stated; for the technical Faggin/D'Ariano framework, D'Ariano & Faggin; for the receiver-model architecture, Information as the foundation; for the political application in Fragile Light, the Fragile Light page; for the synthesis, The Evidence.