Free Will That Makes Sense

As I’ve established before, there’s no good reason to believe in the supernatural. This leaves us with a lot of consequences and unanswered questions. One of these questions is “If everything is made up of atoms, is there free will?”

It’s a good question. However, in order to answer that, we first need to know: What is free will?

 

Setting Up The Problem

Here’s the classic problem. As far as we can tell, everything that exists is based solely upon arrangements of matter-energy in space-time and nothing more. But I still personally hold people accountable for their actions. I still think there are choices to be made, and choices we should make.

The key question is, without suggesting a soul, isn’t there a contradiction here? How do we go from the unthinking nature of atoms to directly blaming people for problems with their thinking? This is the problem of free will.

(Relax, we’ll get to morality in just a few more steps. First I need to demonstrate free will and show that there is moral responsibility, then I can attempt a long unravel of what morality really is. In the end over perhaps twenty or so posts [!], we will emerge with what I think is a complete moral theory. Enjoy the ride, and remember: things are always complicated in philosophy!)

 

Dilbert by Scott Adams

 

What Is Free Will?

The biggest failure of philosophy is that people are arguing over definitions instead of ideas. When someone asks Does free will exist?, the proper response is Depends on how you are defining free will.

So when it comes to free will, what are we really saying? It turns out that we’re looking for some notion of responsibility. Do we really choose our actions or are our actions completely determined for us by something beyond our personal control, like our biology and psychology or God’s unswayable will?

There are basically three camps. The hard determinists believe that there is no free will, specifically because everything we do is determined by physics. They would also agree that the future is fixed, and there is not much point in blaming people for their actions. Note that hard determinism might also be held because of beliefs akin to predestination; that God has perfect foreknowledge and it is impossible to do anything besides God’s Will.

All the way on the other side is libertarianism, the idea that free will exists because a soul allows us to initiate uncaused causes and therefore make choices that are free of any determination.

Lastly, there is the free will theory I defend and believe is obviously correct, called compatibilism. On compatibilism, we do have some sort of free will that exists despite the fact that all our actions are determined by physics.

 

Compatibilist Free Will

As I’m typing these words, I’m making choices. On the whole, I’ve decided to write this blog post instead of writing something else, or not blog at all, or just give up on this whole naturalism thing and become a Christian. I’m choosing to focus on writing these words instead of on the television. I’m choosing to organize my essay this way. I chose the word “essay” for the previous sentence instead of “article” or “blog post”.

It would be absolutely ridiculous to deny that choices exist. I could do this or do that. But all my choices are determined. But determined by what? The answer: Reason. I choose things for a reason. These reasons existed before my choice and I have very limited ability to freely change these reasons or how I reason. As Richard Carrier says, “free will is doing what you want – nothing more, nothing less” and as Arthur Schopenhauer says “Man can do what he wills but he cannot will what he wills”. Both of those very accurately describe what compatibilist free will means.

Is this free will as we think of it? It’s good enough — it accounts for the fact that we know we have choices and we can pick a choice. And as it turns out, all alternatives are crazy — both hard determinism and libertarianism are clearly false. Also, combatibilism works perfectly fine, accounting for everything we want, including moral responsibility.

 

 

Hard Determinism and The Prediction Machine

Imagine there are four cards in front of you, each of a different colour: red, blue, yellow, and green. Which one do you pick?

Well, if you have free will, you pick whichever one you want.

Now imagine we have a Prediction Machine that reads absolutely every piece of information in the universe, including your brain state, and reports a perfectly accurate prediction of which card you choose. Such a machine would be possible only if hard determinism is true, and if such a machine were impossible, hard determinism would be false.

But what happens if we actually used the machine and made a prediction? Well, we would introduce new information to the universe — namely, the prediction itself. This would change the universe considerably, altering many molecules. We would then require a new prediction. However, the issuance of this new prediction would change the universe considerably, and we would need a new prediction. Continue infinitely.

This is simply because the only way to predict something is to measure something, and the only way to measure something is to interact with it, and it is impossible to interact with something without changing it. Predicting your brain state will change your brain state. This is the foundation for, among other things, the Heisenberg Uncertainty Principle. It is physically impossible to predict something with 100% accuracy.

Therefore, yes, all of our choices are caused, but they are unpredictable. Since the future cannot be determined, the future is, by definition, indeterminate. This means hard determinism is false — there are some indetermine parts of the universe, and this is where free will can exist. (I owe this point to Ebonmusing’s essay “Ghost in the Machine”.)

 

Hard Determinism and Ethical Dilemmas

Imagine you’re faced with a classic ethical/moral dilemma, such as deciding whether you need to risk your own life to save a baby from a burning orphanage. You internally battle between fear and duty, weighing in your mind the chance for you and the baby’s survival versus both of your deaths. Now imagine someone comes up to you and says “You don’t have free will, your choice is predetermined. You are forced to make the choice you will make; you cannot choose otherwise.”

That’s great, and that’s exactly what hard determinism says. However, it doesn’t tell you anything useful. The ethical dilemma doesn’t go away, and your personal battle between duty and fear doesn’t get any easier. The knowledge that your choice is predetermined has literally changed nothing. The choice, emotional panic, and guilt does not go away. If determism were true, we would be able to say Oh, ok., make the choice we were going to make regardless, and go on our merry way. But that’s silly — which choice is the one I was going to make regardless? And why are you giving me advice if you don’t think I can listen to it and change my opinions?

Therefore, hard determinism in the sense that “we have no need to think over our choices because everything is predetermined” is completely false. We still agonize over our choices. I owe this point to Eliezer Yudkowsky’s essay “Causality and Moral Responsibility”.

 

Dilbert by Scott Adams

 

Libertarianism and Uncaused Causes

Now that hard determinism has been solved, we slink on over to libertarianism. It turns out that while hard determinism is absurd for denying that choices exist, libertarianism is absurd for denying that causes exist. If libertarianism intends to depart from compatibilism in any meaningful way, it has to say that physical forces aren’t enough to create choice. …But if physical forces don’t create force, what does?

According to libertarianism, we make a choice via an uncaused cause — we… just… choose. But this doesn’t make sense — how can our actions be completely uncaused? How is that in any way distinguishable from doing things at random? If our choices aren’t caused in any way then our choices can’t even be caused by thinking or reasoning. If nothing causes our choices, I might decide to walk into a wall despite not wanting to or decide to jump out a window for no reason at all. Compatibilist free will is what keeps me in control of my actions, and prevents me from acting randomly.

If an exact, perfect duplicate of me was in an exact, perfectly duplicated situation, it would make the same choice I would, 100% of the time. There is no reason to believe that this is not true, however libertarianism denies it. Libertarianism asserts that I would always have a chance of choosing otherwise in this situation. Sure, I could have chosen otherwise if my circumstance was different, but it’s not. I owe all of these points to Richard Carrier’s essay “Moreland’s Christian Science”.

Instead, we have a simple account for all of our actions. We have a certain set of acquired beliefs and acquired desires, and we act to realize our strongest desire at any given time. Our beliefs and desires change as we interact with our environment over time. This is exactly what compatibilism says will happen. Libertarianism would have us act on a desire at random, or act on no desire at all. Hard determinism would have us make a completely fixed choice, without any consideration of our desires.

 

Moral Responsibility Requires Compatibilism

So now that we have affirmed compatibilism as the most accurate account of free will as we can understand it; the only theory that accounts for the existence of both choices and causes; how do we derive moral responsibility? The question remains, How can I blame someone if their actions were caused solely by physical interactions?. Well, I can rephrase the question to make the answer clear: How can I blame someone if their actions were caused by them?.

On naturalism, there is no humanity separate from matter-energy, but only humanity as a specific (and very meaningful) arrangement of matter-energy. This means that the physical interactions of atoms within the human are the human, and this arrangement is the target of our blame, the blame itself an arrangement of matter-energy. As Carrier writes, “it is wickedness we condemn and goodness we praise, not freedom from causation. Complaining changes nothing. But acknowledging your faults and improving yourself changes everything.”

We also know that people’s actions are entirely determined by the circumstances they are in, including their desires and beliefs. This is perfect, because we can blame people in order to change their circumstances, desires, and beliefs! So blame actually does something very important, and we have every reason to continue to blame wrongdoers and praise rightdoers.

 

Moral Responsibility on Hard Determinism and Libertarianism

Now that we can get a sense of moral responsibility with compatibilism, how do we get it from hard determinism? Well, we don’t by definition. Hard determinism tells you that you cannot possibly change anyone’s actions no matter what you do, which not only prevents moral responsibility, but is absolutely untrue now that we think about it.

Again, we bounce over to libertarianism. How do we get moral responsibility here? Well, again, we don’t. Remember that with libertarianism, we are making choices at random; not because we have gone through thought processes, weighed morals, weighed reason, weighed desires, and made a conscious decision. (If this doesn’t sound like actual libertarianism, that’s because most libertarian theorists agree with me and are actually compatibilists as I have been defining compatibilism; see the Tree Falling Problem.)

As I said, Compatibilist free will is what keeps me in control of my actions, and prevents me from acting randomly. Compatibilism also says that one’s desires, beliefs, and circumstances are the sole determinants of choice. This failure to properly control my actions can be changed by changing the my desires, beliefs, and/or circumstances, which can be done via blame.

More importantly, this is also the only method that will work, since it is impossible to blame if people make choices based upon factors other than desires, beliefs, or circumstances. Can you persuade someone who is by definition unpersuadable (hard determinism)? Can you persuade someone to change their soul (libertarianism)?

 

Foxtrot by Bill Amend

 

Why Does Free Will Exist?

This is another question I imagine will be asked — how can we even account for free will without a soul? Well, I answer with only one word: evolution. Ebonmuse puts it best, and I see no reason to reinvent the wheel:

How could free will come into being? By far the most feasible answer involves the evolutionary process that created the human species. After all, free will is a highly adaptive property. A living creature without free will, or some equivalent decision-making capability, would necessarily be guided purely by preprogrammed instinct. This can work so long as that creature never encounters anything other than the limited range of situations it is programmed to deal with, but if it is faced with a situation that does not fit the assumptions of its programming, it will be unable to respond effectively and may well die. (For an excellent example of how instinctive programming can produce a creature unable to deal with novel situations, consider the sphex wasp).

By contrast, a free-willed living being would stand an excellent chance of responding appropriately no matter what type of situation it is faced with, rather than becoming inert or entering an endless loop such as the sphex wasp does. This could conceivably be a powerful selective advantage that evolution would favor for living beings, such as the ancestors of humans, that inhabited complex and unpredictable environments.
– From Ebonmuse’s “Ghost in the Machine”

Followed up by: The Folly of Debating Definitions

 

-

I now blog at EverydayUtilitarian.com. I hope you'll join me at my new blog! This page has been left as an archive.

On 4 Feb 2011 in All, Naturalism, Words. 8 Comments.

8 Comments

  1. #1 Peter Hurford (author) says:
    15 Mar 2012, 7:41 pm  

    In the comments section of “TheraminTrees’s Atheism, Part I: Incompatibility”, I talked about the following two potential arguments for the nonexistence of god:

    ARGUMENT FROM FREE WILL
    P1: If God exists, libertarian free will exists.
    P2: Libertarian free will does not exist.
    C3: Therefore, God does not exist.

    ARGUMENT FROM UNPREDICTABILITY
    P4: If God exists, God can predict the future with 100% accuracy.
    P5: It’s impossible to predict the future with 100% accuracy.
    C6: Therefore, God does not exist.

    Stephen R. Diamond and I disagree on whether these arguments succeed (I think P1 is false but all other premises are true, and Stephen thinks P5 is false but all other premises are true), and also disagree on the success of compatibilism.

    I wanted to unify our discussion of free will and predictability into this comment thread, since this is where I advance those arguments.

    ~

    Now it’s my turn to disagree with your first premise [P5 of the Argument from Unpredictability]. I know there are arguments that the future is unpredictable in principle because of the recursiveness of prediction and I’ve seen your agreement with this line of argument.

    That’s something I’ve held to as you can see in this essay, but I wonder if the predictions might be able to fall into stable timelines in which all actions inspired by the prediction end up either not interfering with the prediction or being part of what inspires the prediction to take place, as seen in the movie Paycheck.

    I think if you predict the future by actually glimpsing the future itself, the future is unavoidable, but if you predict the future by trying to compute it, you would end up being able to prevent it by acting differently. Does that make sense?

    ~

    It isn’t something I’ve thought about much, but it seems to me that the argument might expose another chink in compatibilism. I think the failure of the argument shows that even the limited free will compatibilism offers is illusion: we don’t necessarily have volitional control over our voluntary acts. (Which, it just occurs to me, also seems to be demonstrated empirically by the existence of contra-volitional hypnotic suggestion.)

    Compatibilism doesn’t require us to have volitional control over voluntary acts, it just requires us to have volitional control over other acts, such as my decision to write this sentence. The fact that we can distinguish between voluntary and involuntary acts in such a way seems to indicate that there is a compatibilist free will.

    ~

    To consolidate the argument, 1) if compatibilism were true, prediction of the future would be impossible. 2) No principle of logic or physics precludes perfect prediction on principle 3) Physicalism is true. 4) Therefore, nothing precludes perfect prediction on principle 5) Therefore, compatibilism is false.

    I disagree with both (1) and (2). I think that Observer Effects (for example, the Heisenberg Uncertainty Principle) likely preclude perfect prediction on principle. And even if they don’t, I think we still have volitional control over our acts, it’s just that the choice we make could be predicted in advance. In fact, this is a key premise of Newcomb’s Problem.

    ~

    Compatibilism implies the impossibility of prediction, because it concedes voluntary action, which allows us to defeat any prediction we know of.

    Why does voluntary action concede the impossibility of prediction? Often, we can make very good predictions about what people will do in certain situations, and this doesn’t undermine our volition.

  2. #2 Peter Hurford (author) says:
    18 Jun 2012, 3:45 pm  

    (Update to above: I now think P1 is true.)

  3. #3 Stephen R. Diamond says:
    18 Jun 2012, 5:29 pm  

    I think I missed this.

    Why does voluntary action concede the impossibility of prediction? Often, we can make very good predictions about what people will do in certain situations, and this doesn’t undermine our volition.

    A particular scenario involving the possibility of *perfect* prediction is, I contend, incompatible with volition. It involves the feedback loops I think you believed ruled out perfect prediction. One or the other doesn’t exist, the possibility of perfect prediction or voluntary action, but there’s a principled basis for getting rid of voluntary action but not perfect prediction.

    Say you can predict your behavior and determine you will do A. Doing B is otherwise just an ordinary, trivial choice. On hearing the prediction, you might just decide to do B, say to prove your autonomy. The potential reflexivity of knowledge (without there being any machinery to block its operation at the crucial juncture) shows that either perfect prediction is impossible in principle or truly voluntary action is. There are at least some potential circumstances where what we will is not what we do. So our “voluntary” actions are not necessarily the products of our will. (In practical terms, we usually invent a suitable reason for our supposedly voluntary acts after the fact.)

    Yes. Newcomb’s problem is *based* on this possibility of perfect prediction; which is why there are two perfectly “good” solutions: the problem is incoherent. See my “Another refutation of compatibilism: Newcomb’s paradox shows that free will is entirely illusory” http://tinyurl.com/cdl69lk Leave a comment if you disagree and have a minute.

    I now think P1 is true

    P1: If God exists, libertarian free will exists.

    Hooray! I think monotheism is based squarely on using free will as the basic explanation for the universe. You can take this argument in two directions. 1) As you say, the nonexistence of (libertarian) free will implies no God; 2) Determination by pure free will, conceived as unconstrained, has no predictive power, no actual implications, whatsoever. (The hypothesis that God is good is either a constrain on God not subject to his will, raising new explanatory problems and destroying the perceived elegance of theism OR the hypothesis that God is good is a tautology in that whatever God wants–and necessarily gets–is what’s defined as “good.”)

  4. #4 Stephen R. Diamond says:
    18 Jun 2012, 5:32 pm  

    On the quantum mechanics argument. Even if precise prediction is impossible (that is, qm is a true theory) the _way_ in which quantum mechanics fails bears no relation to the way voluntary action fails. In other words, the possibility of voluntary action shouldn’t depend upon whether qm actually marks the real limits of our knowledge.

  5. #5 Stephen R. Diamond says:
    18 Jun 2012, 7:56 pm  

    SRD: we don’t necessarily have volitional control over our voluntary acts. (Which, it just occurs to me, also seems to be demonstrated empirically by the existence of contra-volitional hypnotic suggestion.)

    PH: Compatibilism doesn’t require us to have volitional control over voluntary acts, it just requires us to have volitional control over other acts, such as my decision to write this sentence.

    I don’t understand your statement, but to elaborate on my hypnosis argument: if you were hypnotized to write that sentence, and wrote it without intending to, “automatically”–perhaps even seeming to yourself as not intending to at all–this, I say, would constitute voluntary action without volitional control. (And its what I contend would have to happen when a person tries to act contrary to prediction. His “voluntary” behavior would have to be contrary to his intentions.

  6. #6 Justin Keith says:
    6 Jul 2012, 4:07 pm  

    Does hard determinism demand predictability, or does it merely demand that things are determined (which may not even be true if ontological randomness is true)? I can watch a movie and not know what will happen at the end (unpredictable) but I can assert that what happens at minute marker 60 will be the same whether I’m at minute marker 10 or minute marker 59. Essentially predetermination is not prediction.

    Clearly humans make choices; some of the input/output processing occurs within a subset of the universe delineated as “me.” Non-human animals and computers do that do. Free will defined that way tends to be very unsatisfying for people, even if it’s factually correct.

    As far as moral responsibility goes, I subscribe to the “tool to correct misbehaving humans” view. Most humans assert that computers have no free will yet they still program them when things go wrong. The even will use the term “wrong” or “at fault” to point out one computer out of many to direct scarce corrective efforts for. They correct the computer with hardware or software fixes.

    Similarly, other humans can do things which some don’t like. They can reprogram these other humans through praise, condemnation, reward, and punishment.

  7. #7 Stephen R. Diamond says:
    7 Jul 2012, 11:21 pm  

    Justin,

    Does hard determinism demand predictability, or does it merely demand that things are determined

    No, but it demands predictability in principle. I think Peter is on the right track in concluding that compatibilism hinges on unpredictability, but the argument for that conclusion isn’t so straightforward. It involves predictions that are reflexive. I develop this idea in my just posted Reflexive prediction, determinism, and the impossibility of free will. (http://tinyurl.com/6tn973k)

    And I think Peter’s mistaken in applying the argument to prediction generally–it concerns the prediction of behavior. The mere fact that a prediction requires an interaction doesn’t make the result unpredictable. In my view, Peter here exposes that he thinks something is special by an interaction initiated by humans, since the world is filled with interactions–which don’t preclude prediction. (This probably doesn’t make sense unless you read my essay.)

    Similarly, other humans can do things which some don’t like. They can reprogram these other humans through praise, condemnation, reward, and punishment.

    From my standpoint, what’s wrong with this is that it doesn’t explain what distinguishes “moral” praise and condemnation from other forms of approval.

  8. #8 Stephen R. Diamond says:
    3 Aug 2012, 9:00 pm  

    I’ve posted an essay that more or less completes my “free will series.” It’s on qualitative experience, which is another issue non-naturalists are likely to press.

    “The supposedly hard problem of consciousness and the nonexistence of sense data: Is your dog a conscious being?” (http://tinyurl.com/c3zq8ht)

Leave a Reply

Comment HTML: You can use HTML in comments. I recommend <blockquote>Quote</blockquote> for quoting what others have said. <b>Text</b> is for bold, <i>Text</i> is for italic, and <a href="url">text</a> is for making links.