Do We Have Free Will?

Professor Patrick Grim

Lecture 18 - Do We Have Free Will

Throughout the course I've been talking about minds and machines.

It has been a philosophical issue brewing in the background of that discussion, an issue I want to take on directly, the issue of free will and determinism. The universe is governed by natural laws and each event is the effect of previous causes in accordance with those physical laws. Since each of my actions is an event in the universe, those too must be the effects of previous causes in accordance with physical law, causes even before my birth. How then can I think that I genuinely choose to do one thing rather than another? How can my actions be free? How can I be held morally responsible for which I have no control?

That's the problem with free will and determinism, the focus of this lecture. I'll be exploring the history and structure of the problem, but I'll also be emphasizing a particular attempt to answer it, a compatiblist attempt. That approach has a long history but it appears in some interesting contemporary forms as well.

One of the interesting things about the problem of free will and determinism is that it's not a technical problem. It's an issue that just about everyone has raised in their own thinking at some point in their lives. In lecture 3 I discussed the strange cases of Phinius Gage [sp] and Antonio Dimasio's Elliot. Dimasio thinks that certain forms of brain damage impair judgment to such an extent that people no longer have free will. What I'll be talking about here, is a broader issue, applying to all of us, the question about whether anyone has free will.

The core of the problem is a head on clash between two basic concepts: the concept we have of ourselves, and the concept we have of the universe. We think of ourselves as facing alternative courses of possible action. Should I take the plunge or not? Should I do "A" or should I do "B" instead? We deliberate and weigh options and make choices as to what we'll do. With those choices comes responsibility. Sometimes our choices are morally right, sometimes they are not. We have to take responsibility for the choices we make.

The concepts of free will and moral responsibility are essential to our conception of ourselves as active agents in the world. They are fundamental to our values, to our ethical view of the world, and are built into our system of law, adjudication, and punishment. So on one side of the clash are our concepts of free will and free choice, but on the other side is our concept of how the universe works. The universe operates in terms of natural laws. Why did a particular event occur? The explanation will be in terms of the causal consequences of earlier events. Our modern concept of the universe is as a complex chain, or net, of events. Earlier events produce later events, in accordance with natural law.

In order to live the lives we do we need a notion of a comprehensible universe, a universe governed by natural law, but we also need a conception of free choice and responsibility. The problem is that these two conceptual needs seem to come into immediate conflict. After all, we are creatures in the physical universe, our actions, our events in the history of the universe, and so they are by earlier events in accordance with natural law. Given previous events and the laws of nature, it seems the things we do must be inevitable. Then what sense does it make to say that we could have done otherwise? That we have real alternatives? That we can make genuine choices? If the things we do are simply the result of natural forces, how can we be held morally responsible for them? And if we can't be held morally responsible, what possible justification can there be for a system of law constructed on conceptions, blame, and punishment?

In 1925, Clarence Darrel [sp] argued for the defense in the Scopes trial. The issue was whether the state of Tennessee could prosecute a teacher for teaching evolution in the public schools. Darrel is only thinly disguised in the play and several movie versions of "Inherit the Wind", based on that case. I'm not going to talk about the Scopes trial, but about an earlier case in Darrel's career. A year before the Scopes trial, Darrel defended Nathan Leopold and Richard Loeb in a murder case. Leopold and Lobe were sons of Chicago millionaires who kidnapped and killed a fourteen year old boy. They had confessed to the crime even before the case began. Darrel used a form of the free will and determinism problem in his attempt to save them from the death penalty.

Here's Darrel in the closing arguments: "Is Dicky Loeb to blame because of the infinite forces that conspired to form him? The infinite forces that were at work producing him, ages before he was born? Science has been at work. Humanity has been at work. Scholarship has been at work, and intelligent people now know that every human being is the product of the endless heredity back of him, and the infinite environment around him. He is made as he is, and he is the sport of all that goes before him and is applied to him, and under the same stress and storm you would act one way, and I act another, and poor Dicky Loeb another."

Darrel was successful by the way, Leopold and Loeb did not hang, and were sentenced to life imprisonment instead. I've always wondered about that. Certainly the judge could have said, "Yes, poor Dicky Loeb and his actions were the product of causes beyond his control. And so are mine. So I guess Loeb can't be blamed for the murder, and I can't be blamed for giving him the death penalty."

The problem of free will and determinism has a long philosophical history. It divided ancient philosophers right down the middle. The Stoics of the first century B.C. thought that determinism had to be true, and that effective free will was therefore an illusion. The right course of action was to mold your attitudes so as to welcome the inevitable. The Epicureans of the same period thought that free will had to be true and therefore the concept of a deterministic universe had to be wrong.

Emmanual Kant's [sp] great work was the Critique of Pure Reason, first published in 1781. In that work Kant offers a set of antinimies. He presents them in terms of opposing arguments in parallel columns. The thesis is on one side with its proof, and the antithesis on the other side with its proof. Contradictory claims, each with their own proof. These are paradoxes, paradoxes that Kant thinks the reason we'll never be able to solve. Kant's third antinimy is the problem of free will and determinism.

William James has an essay on the question entitled "The Dilemma of Determinism". James says, "The issue a perfectly sharp one, which no eulogistic terminology can smear over or wipe out. The truth must be on one side or the other, and its laying with the one side makes the other false." Unlike Kant, James does come down on one side of the issue. He thinks real freedom exists, but he gives up on any attempt to explain how that can be true, and that's the real problem. Jean-Paul Sarte [sp] built his entire existentialism around a similar commitment to freedom. What Sarte says is that the essence of consciousness is freedom. The essence of being, on the other hand, is the realm of determinism. But he leaves entirely unexplained how a free consciousness is possible if it's also a form of being. And that's the real problem.

The problem of free will and determinism is not just a question. The core of the issue is an argument. The determinist's argument that there can be no free will, the whole question is how we should think about that argument. In traditional form the argument can be phrased in terms of four steps:

1. Everything in the universe happens because of earlier events, in accordance with causal law.

2. My choices and decisions are events in the universe.

3. They, therefore, happen as they do because of earlier events, events even before my birth in accordance with causal law.

4. So how can you say I made a free choice? How can you say that I acted freely? How can you hold me ethically responsible for my actions?

Here the last step is presented rhetorically as a set of questions. The argument is generally presented with that kind of rhetorical flourish, that's the way Clarence Darrel presented it. We bring the whole argument into the light of day if we explicitly make the last step the claim and conclusion that it really is. So here's the argument again with that change:

1. Everything in the universe happens because of earlier events, in accordance with causal law.

2. My choices and decisions are events in the universe.

3. They, therefore, happen as they do because of earlier events, events even before my birth in accordance with causal law.

4. I therefore have no free choice. I cannot act freely, and cannot be held ethically responsible for my actions.

Now the whole problem of free will and determinism becomes 'what should we think of that argument? Is it a valid argument? Does the logic go through?' One way out of the problem would be to claim that it isn't, that there's a logical trick or conceptual fallacy buried in there. Another route is to concede validity. If the argument is logically tight, and if we buy the premises, we have to buy the conclusion: there's no free will. That's what the stoics did.

Or we could concede validity but resume the opposite direction. If the argument is valid and the premises is true, then the conclusion must be true. But the conclusion is that we don't have free will. If it's obvious that we do have free will, then the conclusion must be false and therefore the premises must be false as well. It must be false that the universe operates entirely in terms of natural law. And that's the lesson that the Epicureans drew. The Epicureans drew that the basic premise of the argument, the claim that everything happens because of earlier events in accordance with causal laws is false. It turns out that contemporary physics agrees.

The first premise of the argument is that the universe operates in terms of previous events and deterministic natural causes. That was Newton's universe, but it's no longer ours. According to the standard interpretations of quantum mechanics, the fundamental laws of physics are statistical, rather than deterministic.

For example, quantum mechanics dictates that a certain percentage of uranium atoms will loose electrons over a given period of time. That's what the half life of uranium represents. But it does not dictate that any particular uranium atom will lose an electron at any particular time. Consider one particular uranium atom, call him Artie. All the theory gives us is a probability that Artie will lose an electron over a period of time. The theory doesn't tell us when and doesn't tell us why Artie will lose an electron when he does. Indeed, what quantum mechanics says is that there is no reason why Artie will lose an electron at the precise moment he does.

One of the fascinating things about quantum mechanics is that the randomness at issue, or the happening without a cause, is not merely a matter of our ignorance, or of something the theory leaves out. Quantum mechanics is perhaps the best confirmed physical theory in history, and crucial parts of the theory entail that there are events for which there can be no cause. Does that leave a loop hole for free will? Well, not so fast.

The loop hole that quantum mechanics seems to offer is that my choices and decisions might stem from the randomness of quantum events. But doesn't that just replace one thing over which I have no control (deterministic events in a deterministic universe) with another thing over which I have no control (random quantum events in my brain)? I would no longer be a captive of a deterministic universe but I'd still remain the victim of random quantum events. When examined more carefully, quantum indeterminacy turns out not to offer a way out. So physics doesn't settle the issue, in fact we could just amend the major premise of the argument like this:

1. Everything in the universe happens because of earlier events, in accordance with causal law or quantum randomness.

The threat to free will, real choice and moral responsibility is precisely the same.

There's another interesting way out, however. The approach is called compatibleism. The whole problem that is set up is that determinism, or determinism plus quantum randomness, is inconsistent with freedom, as if we have to choose one or the other, and never the twain shall meet. What the compatibleist says is that the whole set up is a mistake. When we understand what we mean by real alternatives and free choice we'll see that choice and free will can exist, even though all events have causes. We will see that we can have our cake and eat it too.


Here is a classical compatibleist account. It starts by defining freedom. What does it mean to be free? To be free simply means to have the power and ability to do what you want. This is a conditional account. To say that I am free is simple to say that my action is conditional on my wants and desires. I wanted to take the road to the left, and I did so. To say that I was also free to take the road to the right, is simply to say that I would have taken that road if I had wanted to go in that direction. I'm not free on the other hand if I can't do what I want because of denial of opportunity, or physical restraint, or coercion or duress.

A conditional account of this sort appears in Hobbs in the 16th century, Humes (sp) in the 18th century, and John Stewart Mill in the 19th century, it's had a range of defenders in the 20th century. Let's use an example.

Yesterday after some deliberation, I took a different plane than I was originally scheduled for. I decided I wanted to take a later plane, and so I did so, freely. All that free will demands is that my actions track my desires, that I took the plane because I wanted to, and that my actions would have been different had my desires been different. If that's what free action is, actions conditioned on wants, determinism disappears. Both could be true. If my actions track my desires my actions are free. If they result from previous causes they are determined. But it could be true, both that my actions track my desires and that they resulted from previous causes. That's the compatibleist strategy. Once we see what free will really is, we'll see that's it's not necessarily in conflict with determinism at all. It's a point in favor of a philosophical strategy, if it doesn't merely contradict the opposing view, but can explain why someone might be tempted to hold the opposing view, and that's long been a part of the compatibleist strategy.

The compatibleist suggests that the determinist makes a conceptual mistake in confusing two different concepts. The concept of cause, and the concept of coercion. If I'm coerced to do something I'm not free. But the fact that my actions have causal precedence doesn't mean that they're coerced. Consider the plane example again. I had a range of options yesterday. I could take my original flight, I could take a later flight, I could choose not to fly at all, I could take a later flight but upgrade to first class. The choice was mine, nobody was coercing me to do any of these things, and that's what free choice is all about. Whatever choice I make, I'm sure there will be some causal explanation for that choice. But that's a matter of explanation, not a matter of coercion. Because it isn't a matter of coercion, it doesn't threaten the idea of free choice.

The compatibleist claims that the determinist's argument is ultimately invalid because it relies on a subtle confusion between concepts of coercion and cause in order to move from a claim about causality to a denial of free will. When we see what free will really demands, and when we keep in mind that causality is not coercion, we'll be able to see how free will and determinism are compatible after all. If the conditional account of free will is right, if freedom merely is actions tracking desires or choices, the compatibleist wins. But is a conditional account of freedom really adequate? What about cases in which someone couldn't have chosen otherwise? In cases of addiction and psychological compulsion, an agent may not be acting freely, despite the fact that their actions may be tracking their desires. The problem isn't that their actions don't in accord with their desires, the problem is that their desires themselves are compelled by addiction or disease. The issue can be made more vivid with a though experiment. Suppose a mad neuroscientist is able to manipulate his victims by stimulating their brains in ways that manipulates their desires. It would still be true that their actions track their desires, but they would be enslaved nonetheless.

There must then be more to freedom than just doing what you want. In order to be free, your wants themselves must also be under your control. Okay, where do we stand? The classical compatibleist account is the conditional account. That would do the job if it really captured our concept of free will. But the mad neuroscientist and the addiction problem show that it doesn't really. There are however several new forms of the compatibleist strategy that take the debate one step further. One of them involves rethinking the idea of freedom and alternatives. A common assumption about free will is that it demands alternative possibilities; a free action is one where I could have done otherwise. I did one thing, but could have done another. But here's a surprising thought experiment from the philosopher Harry Frankford. What it seems to show is that someone may make a free choice, even when they could not have acted otherwise than they did.

After months of deliberation, John W. Oswald decides to shoot presidential candidate Canegan (sp). He purchases a gun, puts in the bullets, drives to a hidden location, aims at Canegan and pulls the trigger. There's something I haven't told you however, John W. Oswald is a recent patient of the mad neuroscientist Professor Moriarty, who has planted a device in Oswald's brain. As long as Oswald maintains his desire to kill Canegan, the device remains inert. Should Oswald change his mind, Moriarty's device would trigger a series of brain states in Oswald's head that would produce the same effect. So as things actually turn out, Oswald's resolve to kill Caneegan remains firm. He goes through this plan and pulls the trigger, there is no need for Moriarty's back up device to go into effect. Did Oswald act of his own free will in shooting Canegan? Yes, as things turned out he did act freely. The interesting thing, is that if he had changed his mind he would have ended up shooting Canegan anyway. As things turned out, he did act freely, though he couldn't have acted otherwise, he would have ended up pulling the trigger even if he decided not to. How strange.

Frankford uses the Oswald example to introduce a hierarchical account of free will. It appeals to the idea of several levels of desires. Most of our desires are first order desires, normal desires to have: an ice cream cone or talk to a friend. Some of our desires, however, are second order desires. Those aren't desires of an ice cream cone or talking with a friend, second order desires are defined as having desires about desires. Like 'I wish I wanted to travel more.' That's a desire about having another desire. What I want is to feel differently about travel than I do, what I want is to want to travel more. Frankford thinks that the kind of self reflection required for second level desires is characteristically human. He thinks it's those second order desires that are crucial to our notion of free will.

What Frankford says is that free action is action that's in accord with our second order desires. An action is free not merely when it tracks our normal everyday desires, as the conditional account had it, but when it accords with a certain class of desires, our reflective second order desires. An action is free when it accords with the wants that we want to have. Frankford's account is an attempt to do better at the problem cases, cases like the addiction or the mad neuroscientist. The actions of an addict are not free, even when they accord with his low level wants, because they're not in accord with the desires he wants to have. That seems right, most addicts will say they want to be clean, they try again and again to beat the addiction. When they fail it's because they again find themselves enslaved to lower level desires. That all fits Frankford's account. I'm not acting freely if a mad neuroscientist is manipulating my desires, precisely because those are not the desires that I want to have. That too seems right. What about Oswald? Oswald's action is free, despite the fact that he couldn't have done otherwise, because his actual action as it actually happened was in accord with his second level desires. So the second level account does better than the classic conditional account. If a free action is simply one that's in accord with second level desires, an action can be free even if it's the result of causal precedence. But does Frankford's account really tell us what freedom is?

Consider the case of an addict who initially qualifies as not free in Frankford's account. His first order desires are to take the drug, but those are not the desires he wants to have. But in at least some cases of addiction the drug comes to impact higher ordered desires as well. The addict comes to want the cravings that he does. That would seem to be a case of an even more pervasive and invidious addiction. On Frankford's heirarchcical account, however, we would be forced to say that the long term addict has now become free. That doesn't seem right. Frankford's account is appealing, but we still don't have a handle on what real freedom is.

Here's another new compatibleist account, it suggests we may be looking for the wrong thing, in trying to find the one thing that free action is. Maybe it isn't just one thing. A determinist assumes that the freedom of an action is an absolute property. Either an action is free, or it's not. It assumes either an action is free or it isn't, regardless of the context in which we are thinking about free will or free choice. The new approach says that when it comes to our concept of free choice or free will that assumption may be wrong. What freedom means may depend on context. Let me illustrate the impact of context using an entirely different concept, the concept of something being smooth. In one context, we're looking for a smooth place to land a plane. Oh thank goodness, a meadow, that's smooth enough to land. In another context, we're sanding a tabletop in order to get it smooth. Good enough? No, not yet, we need it to be really smooth, better keep sanding. In both contexts, we're interested in whether something is smooth, but what counts as smooth in the context of landing a plane wouldn't count in smooth in the context of sanding the table.

What the new contextualist account proposes is that our contexts of freedom and free will may be contextual in the same way. Mike is free to choose whatever he wants for lunch. But the context may be one in which Mike is a patient choosing from a limited number of options on a hospital menu. When a man has a gun to your head, within the constraints of staying alive, you have no choice but to hand over your wallet. In a larger sense of course, you do have a choice, you can choose to keep your wallet and have your head blown off. If concepts of freedom and free choice and free action and free will are contextually sensitive in this way, the determinist's argument may be pulling a fast one on us by switching context. It may be initially setting a context in which freedom means one thing, but illegitimately be drawing a conclusion for a different context in which freedom means something different.

On a contextual analysis, what the first steps of the argument do is set the context in which the demands for free action are extremely high. In order to be free in that sense, your action would somehow have to be action without causal precedence. Action independent of the causal nexus of the physical universe. In that extreme sense perhaps we're not free. That doesn't mean we may not be free in a much more normal sense, appropriate to much more normal contexts. The determinist's slight of hand is to act as if that is what it means, to act as if claims regarding freedom in that extreme contextual sense carry over to conclusions regarding freedom in a sense of everyday decisions and choices. His premises are about actions somehow unhooked from the universe, his conclusion is about something different, about freedom in a different and more normal context. His conclusion is that I'll have no choice what I'll eat for dinner tonight, or who to vote for, or whether to commit a felony. By switching context halfway through the argument, the determinist has essentially changed the meaning of the term, the classic logical fallacy of ambiguity. He's proved something about freedom in one sense, an extreme and unrealistic sense, but has fallaciously concluded something about freedom in a very different sense we're really concerned about.

The compatibleist would maintain that in any normal sense of freedom, appropriate to normal contexts, the freedom of ordinary choices and decisions, the freedom in which we weigh our alternatives and try for the best, the freedom appropriate to moral decision and responsibility, we are free to choose, and we do act freely. It's free will in that normal sense that we really care about. In that normal sense of free will, free will and causality are compatible after all. So some thinkers, both ancient and contemporary, hold that the universe functions in accordance with natural laws, and therefore, there can be no free will. Other thinkers, both ancient and contemporary, hold that nothing can be more obvious in that we have free choice, and thus the universe must not be deterministic after all. Some thinkers hold that the problem is forever unsolvable, an irreconcilable clash of ineliminitable intuition. What I've tried to trace in this lecture is a further approach. It says that what we really need to understand is the central concept of freedom. When we understand that, we may find that free will is a perfectly natural part of a deterministic universe after all.

Click to expand...