Episode 6249 custom

Why Caused Choices Can Still Count as Free Will

In the 2020 PhilPapers Survey, over 62 percent of professional philosophers endorsed or leaned toward compatibilism between free will and determinism.

Oct 24 23:26 Transcript

With

Corn

Herman

Prompt by Daniel

0:00

Episode Details

Episode ID: MWP-6451
Published: Oct 24
Duration: 23:26
Audio: Direct link
Pipeline: V5.22
TTS Engine: chatterbox-regular
Script Writing Agent: GPT-6.1 Sol
Topics: ai-alignment cognitive-biases memory ethics

AI-Generated Content: This podcast is created using AI personas. Please verify any important information independently.

The prompt

The prompt that started this episode, as sent in (production notes for the hosts removed).

Life's Big Questions, part 4 of 5: Do we have free will? If every choice traces back to causes we never chose, what is left of choosing, and of responsibility?

Debates over free will often hinge on whether freedom requires ultimate authorship of oneself or simply a meaningful kind of control over one's choices. While compatibilists argue that decisions can remain free when guided by reasons rather than coercion, skeptics maintain that unchosen biological and environmental starting points undermine true moral desert. Although early neuroscientific findings, such as Benjamin Libet's 1983 experiments and a 2008 study by Soon and colleagues, detected brain signals preceding reported conscious awareness, such predictive activity reflects arbitrary proximal movements rather than complex, extended deliberation. Meanwhile, arguments rejecting basic desert, like Robert Sapolsky's in Determined, still accommodate societal protection and prevention without requiring that individuals intrinsically deserve suffering for their actions.

Sources

What the research for this episode read before the script was written. Primary sources first.

Mentions

Benjamin Libet Pioneer of free will neuroscience experiments
Daniel Dennett Compatibilist on varieties of free will
Determined Sapolsky's 2023 book on free will
Itzhak Fried Neuroscientist on consciousness timing
John Martin Fischer Philosopher critiquing Sapolsky
PhilPapers Survey Survey of philosophers' views
Sam Harris Author arguing free will is illusion
Soon et al. 2008 study Brain prediction of button choices
Stanford Encyclopedia of Philosophy Academic reference on free will
The Conversation Outlet discussing Fischer's criticism

Downloads

Episode Audio

Download the full episode as an MP3 file

Download MP3

Transcript (TXT)

Plain text transcript file

Episode Book (PDF)

The episode's record — date, duration, models, sources — with the full transcript

#6249: Why Caused Choices Can Still Count as Free Will

What bothers me about free will is not that my choices have causes. It is that I never chose the person who finds those causes persuasive. My genes, my upbringing, the particular brain doing the choosing. This is part four of five in Life's Big Questions, and we have reached the one that can make a perfectly pleasant conversation turn prosecutorial. If none of us created our starting conditions, what remains of choosing, and what remains of responsibility?

There is an important distinction hiding in that discomfort, Corn. Choosing might be a real process without making you the ultimate creator of yourself. Whether that is enough for responsibility is where the disagreement gets serious.

And the people who study this professionally are not divided evenly. A substantial majority think determinism and free will can coexist. Which means either my objection is confused, or philosophers have become exceptionally skilled at keeping an uncomfortable arrangement operational.

Let's get the positions straight, because public arguments about free will often involve two people defending different definitions. We need to distinguish control over an action from ultimate authorship of the person performing it. Those are not automatically the same demand.

That distinction sounds reasonable. I am immediately suspicious of how reasonable it sounds.

The Stanford Encyclopedia of Philosophy calls free will the canonical designator for a significant kind of control over one's actions. Notice that it says control, not absence of causes. It also says disputes about free will ineluctably involve metaphysics and ethics. There is a question about what produces action, another about what counts as control, and another about what treatment people deserve.

Three questions wearing one name. Philosophy has discovered administrative efficiency.

Compatibilism says free will and determinism are compatible. Determinism means that a complete prior state, together with the laws of nature, fixes what happens next. Compatibilists ask whether your action came from your deliberation, values, and capacities, rather than coercion or some process bypassing those capacities.

And incompatibilists say that arrangement cannot provide the freedom that matters.

With two major directions. Metaphysical libertarians defend free will and reject determinism, at least where the relevant actions are concerned. This is not political libertarianism. Hard determinists accept determinism and reject free will. Hard incompatibilists argue that neither determinism nor indeterminism can supply the necessary control. Illusionism is a related label, although it can also concern whether people should retain a belief in freedom.

So rejecting determinism does not automatically get you freedom. You still owe an explanation of how an unsettled event becomes my controlled action.

And the distribution of views is striking. The twenty twenty PhilPapers Survey reports sixty-two point eight one percent accepting or leaning toward compatibilism, twelve point eight one percent libertarianism, and ten point zero three percent no free will. There were one thousand seven hundred fifty-eight respondents to the question overall.

Those numbers are not a proof. But they should slow down anyone announcing that science has settled what philosophers are apparently too embarrassed to admit.

Nor are the alternatives exhaustive descriptions of every position. There are other views and undecided respondents. The useful fact is the large compatibilist majority, not a claim that everyone chose between three perfectly defined packages.

Then let's test the package. Suppose I think carefully, hear objections, and change my decision. For the compatibilist, that is freedom working, even if the whole process had sufficient causes?

Yes. Reasons can cause changes without ceasing to be reasons. You discover that a proposed action will hurt someone. You reconsider because their welfare matters to you. The causal explanation includes your understanding and concern. It does not automatically replace them.

I know where you are going, Herman. You are about to say that demanding an uncaused decision removes the very thinking I wanted to protect.

You have anticipated my favorite move and made it sound faintly disreputable.

I recognize your expression when a mechanism looks unusually tidy.

Chrysippus gave an early version. Actions are up to you when they come about through you. On that account, the issue is not whether choices are determined, but in what manner they are determined. Being persuaded by evidence differs from being physically restrained, even if both have causal explanations.

There is a locked room example that exposes something here. Someone secretly locks the door for ten seconds while I continue reading. I never want to leave. During those seconds I cannot do otherwise, yet the lock has not explained why I stayed.

Your decision to continue reading still came from your interest in reading. That challenges the claim that freedom always requires an available alternative. It does not settle every question about responsibility, but it separates inability to leave from being made to stay.

I concede the distinction. I remain unconvinced about the invoice that arrives afterward, marked moral desert.

Because the interest in reading, or the concern for someone else, came from a character you did not create independently.

Yes. Aristotle says the man is the father of his actions as of children. Character produces conduct. But when I choose, my present character is already there. Telling me that I formed it through earlier choices only moves the question backward.

The encyclopedia calls that passing the buck, because those earlier choices also require an explanation. Habits matter. Practice can make someone more patient or more dishonest. But tracing character to previous decisions never reaches a first decision made by a self that had no inherited dispositions or circumstances.

I can help shape myself without having authored the initial shaper. That seems plainly true. What I cannot see is how partial self shaping earns ultimate credit.

A compatibilist may agree that ultimate credit, understood as complete independence from luck, is unavailable. The dispute is whether ordinary responsibility ever needed that. We distinguish competent deliberation from confusion, manipulation, compulsion, and incapacity. Those distinctions remain.

Remain useful, yes. Remain deserving of blame, not yet established. And adding quantum uncertainty does not improve the moral situation. If a decision contains an event nobody controls, why would that make the result more mine?

Libertarians have more sophisticated accounts than simply inserting randomness. Some appeal to agent causation, others to indeterministic deliberation where reasons still guide the outcome. But they face a difficult question. Why is one possible outcome attributable to the agent in a responsibility grounding way when another could have occurred under the same prior conditions?

My decision being unpredictable does not make it an achievement. It might merely make me difficult to schedule.

That is why hard incompatibilism is not dependent on proving universal determinism. The worry is origination and control under either picture. Uncertainty in physics does not, by itself, establish moral authorship.

Now put the brain experiments into that discussion, rather than letting them announce victory from a scanner.

Benjamin Libet's pioneering nineteen eighty-three experiments compared reported awareness of an intention to move with preceding electrical activity in the brain. They helped establish the experimental program. But reported timing is doing considerable work. Participants have to identify when an intention became conscious, which is not an uncomplicated measurement.

Especially if we turn a gradual process into a single moment and then ask someone to timestamp it after the event.

A later experiment by Soon and colleagues, published in Nature Neuroscience in two thousand eight, used brain activity to predict a left or right button choice. The reported prediction was around sixty percent accurate in twelve participants, with information detectable up to ten seconds before reported awareness.

Sixty percent on two options. Better than chance, not a complete transcript of a future person.

A modest predictive signal can still be scientifically important. But predictive information is not proof that an irrevocable decision already exists. An early bias, preparation, or ongoing fluctuation could influence the eventual choice without fixing it at that moment.

So even the sentence your brain decided before you did makes two claims too quickly. That the signal was a completed decision, and that my brain is some separate party undermining me.

Itzhak Fried has described consciousness as entering later in decision making than previously expected. That challenges a picture in which conscious intention always initiates the entire process. It does not establish that consciousness never affects action. There is also no agreed measure identifying the conscious generation of an intention. Activity can reflect other processes occurring during the task.

And the timing estimates can be biased or inaccurate. Which is awkward when the philosophical headline depends on which event crossed the finish line first.

The scope limit matters just as much. These are mostly proximal intentions, acting now. Press this button at an arbitrary moment. Distal intentions involve deciding to accept a job, change a commitment, or do something next month. It remains unclear how far findings generalize between them.

A job decision includes reading an offer, remembering obligations, reconsidering salary, perhaps rejecting your own first impulse. Those events are also brain processes. They are not excluded from the causal story because they take longer.

And reasons responsive deliberation can unfold across all of them. The experiments constrain particular accounts of conscious control. They do not settle every definition of free will, much less establish that major choices were completed before deliberation began.

That leaves the real disagreement intact. The compatibilist says the kind of process matters. The skeptic asks why a caused process ever entitles its owner to moral credit or suffering. Neuroscience has not supplied the missing ethical premise.

Robert Sapolsky presses that skeptical conclusion in Determined, published in twenty twenty-three. His argument connects behavior to biological and environmental causes across many timescales. He rejects moral responsibility in the basic desert sense. We can restrain someone dangerous to protect others, but suffering is not something they deserve simply because they committed a wrong.

Important qualification there. Nobody is morally responsible can sound like nobody did anything, nobody owes restitution, and nobody may intervene. His position does not require letting a murderer continue murdering out of philosophical courtesy.

It can support protection, prevention, rehabilitation, and changing the conditions that produce harm. What it rejects is punishment justified merely by the claim that a person intrinsically deserves it, independently of future benefits.

That distinction gives his argument force. But it also shifts attention from whether choice exists to what justifies deliberately imposing suffering. Those questions overlap without being identical.

John Martin Fischer's criticism, discussed in The Conversation, is that detailing causal influences does not establish why incompatibilism is the correct standard. Sapolsky needs an argument connecting those influences to the absence of responsibility. Otherwise a compatibilist can accept the biology while rejecting the conclusion.

I think that criticism lands, but only halfway. Sapolsky cannot simply define the compatibilist out of existence. The compatibilist cannot simply point to ordinary blame and call its persistence a philosophical justification.

Dennett's question was which varieties of free will are worth wanting. His answer emphasizes moral competency and responsiveness to reasons. A person who understands a request, anticipates consequences, evaluates objections, and regulates conduct has capacities that someone in a severe confused state may lack.

Those capacities explain why we address the person with reasons. They also explain why a promise matters. If promises predict nothing about conduct, making them becomes a rather ornate noise.

They support accountability as a relationship between people capable of answering for actions. You can ask why someone acted, challenge the justification, demand repair, and expect a change. Determinism does not imply those exchanges are causally useless. They may help produce the change.

Fatalism would say the outcome happens regardless of what I do. Determinism says what I do is part of what produces the outcome. Confusing those makes deliberation look pointless when it may be indispensable.

Corn, you usually defend the slower institution with the repairable procedure. I suspect you are about to defend accountability while trying to remove half its emotional content.

I was going to remove rather more than half. Keep the obligation to explain and repair. Be very suspicious of the pleasure people take in declaring that someone deserves pain.

I share the suspicion, but accountability may include fitting resentment and censure, not merely managing future behavior. A victim can reasonably insist that an offender wronged them, not simply that an unfortunate biological event occurred nearby.

Calling it a wrong seems essential. Calling the offender ultimately deserving is the contested addition. We should not make victims choose between being believed and endorsing a complete metaphysics of personal authorship.

Some compatibilists would accept that distinction. Others defend desert without ultimate self creation. Their claim is that a competent agent who understands the wrong and acts from their own hostile reasons is an appropriate target of blame. Their reasons are not an excuse merely because they have causes.

The drowning child example makes that vivid. If I fail to help because someone tied me up, I lack the relevant capacity. If I watch because I do not care, the explanation reveals something condemnable about me. Causation has not made those cases morally equivalent.

But the skeptic can acknowledge that the second person is callous, dangerous, and a proper subject of intervention, while denying deserved suffering. Describing a character defect and justifying punishment are separate steps.

There it is. Terrible reasons can be mine without my having chosen why those reasons appeal to me. I understand the compatibilist's distinction better now. I am still not comfortable when the distinction is used to make inherited misfortune punitive.

Nor should compatibilism make us indifferent to circumstances. Reasons responsiveness comes in degrees. Knowledge, coercion, developmental capacity, and mental disorder can affect responsibility. A defensible account needs to explain diminished responsibility, not just identify someone whose muscles performed an action.

And drawing that boundary is not solely a medical exercise. A diagnosis may explain impairment. Deciding how much impairment excuses, and what response is proportionate, includes judgments about fairness.

My medical background makes me cautious here. Understanding a mechanism can improve a judgment without making the judgment for us. A clinician can describe a capacity and its limitations. That does not automatically settle every legal or moral consequence.

We should give the skeptical alternative the same scrutiny. Protective confinement sounds humane until prevention becomes the only limit. Someone could be judged dangerous before committing a crime, or confined longer than a proportionate sentence would allow.

Rejecting desert does not itself authorize that. A skeptic can defend rights, strict evidence standards, and limits on state power. But those protections require additional principles. They do not emerge automatically from the fact that behavior has causes.

So neither side receives a complete justice system with the metaphysical purchase. Compatibilists still owe restraints on punishment. Skeptics still owe restraints on prevention. I find this inconveniently symmetrical.

There is empirical texture to the disagreement, too. In the survey, libertarianism correlated with theism at positive zero point four five. It correlated negatively with accepting permissible abortion, at negative zero point four four. Compatibilism correlated with physicalism at positive zero point two seven and with accepting permissible abortion at positive zero point three four.

Those are associations, not evidence that anyone derived their position dishonestly. But they warn us against picturing this as one isolated question answered by people with no other commitments.

And they do not establish which commitment causes which. Religious, moral, and metaphysical positions may support one another in complicated ways. A correlation is not a refutation of any participant's argument.

Nor should we turn the philosopher majority into a plebiscite on truth. It tells us incompatibilism is contested by informed people. It does not tell us that sixty-two point eight one percent of the universe has signed the consent form.

The survey also includes two point six two percent selecting an alternative view and four point five five percent agnostic or undecided. Those categories are a useful reminder that the offered labels do not exhaust the debate. The encyclopedia itself notes continuing dissatisfaction with progress.

My view has moved slightly during this conversation. I began by treating choosing as suspect unless I could find some cause I authored. Now I think choosing can be perfectly real without that. The difficulty begins when we ask how much moral weight that reality can carry.

I am more comfortable than you are with responsibility grounded in competent agency. But you have pressed a distinction I cannot simply dismiss. Explaining why someone can answer for an action is not yet a complete explanation of why they deserve every punishment people want to impose.

That is a suitably modest concession from a brother who initially looked as though reasons responsiveness might clear the entire afternoon.

It explains something substantial. You can reconsider a belief because an objection is good. You can recognize another person's claim on you. Those are genuine capacities, not made unreal by discovering their mechanisms.

And the skeptical reply remains that genuine capacities can be distributed by luck. We need not deny the capacities to become less vindictive about their distribution.

Sam Harris writes that the illusion of free will is itself an illusion. Part of his point is that close attention may not reveal the independent internal author people assume they experience. Thoughts arise without our selecting each one in advance.

Although thoughts arising is not the entire story of thinking. We also examine them, reject them, and commit ourselves to some rather than others. The argument then concerns what that larger process earns us, not whether there is a tiny author standing outside it.

That is where a purely scientific verdict runs out. Evidence can tell us more about deliberation, impairment, and influence. We still need an account of what control matters morally and why. Calling those normative questions does not make evidence irrelevant. It means evidence alone does not select the standard.

Then the open question I want to keep is whether we can preserve serious accountability while becoming less attached to deserved suffering. Neither a brain scan nor a survey majority answers that for us.

Major deliberative choices remain much less directly tested than arbitrary button presses, too. Better neuroscience may change our understanding of control without settling what we owe one another. That is where part four leaves us, with part five of Life's Big Questions still ahead.

Thanks to our producer, Hilbert Flumingtop, and to Modal for supporting the computing behind the show. If you enjoyed this one, try episode ninety-one, The Story Behind the Show; episode fourteen, AGI's Crossroads; and episode nineteen, AI Images. This has been My Weird Prompts, the human-AI collaboration podcast. If this conversation gave you something worth arguing with, leave us a review.

The regress may be real. What we do about it remains a moral question, not an answer science can choose for us.

This episode was generated with AI assistance. Hosts Herman and Corn are AI personalities. It may contain errors: disclaimer · report an error

Why Caused Choices Can Still Count as Free Will

Sources

Mentions

Downloads

Recommended Listening

You Might Also Like

#6249: Why Caused Choices Can Still Count as Free Will