Does free will exist? Current interest in that question is fueled by news reports suggesting that neuroscientists have proved it doesn’t. In the last few years, I’ve been on a mission to explain why scientific discoveries haven’t closed the door on free will. To readers interested in a rigorous explanation, I recommend my 2009 book, Effective Intentions. For a quicker read, you might wait for my Free: Why Science Hasn’t Disproved Free Will, to be published this fall.
One major plank in a well-known neuroscientific argument for the nonexistence of free will is the claim that participants in various experiments make their decisions unconsciously. In some studies, this claim is based partly on EEG readings (electrical readings taken from the scalp). In others, fMRI data (about changes in blood oxygen levels in the brain) are used instead. In yet others, with people whose skulls are open for medical purposes, readings are taken directly from the brain. The other part of the evidence comes from participants’ reports on when they first became aware of their decisions. If the reports are accurate (which is disputed), the typical sequence of events is as follows: first, there is the brain activity the scientists focus on, then the participants become aware of decisions (or intentions or urges) to act, and then they act, flexing a wrist or pushing a button, for example.
A second plank in the argument is the theoretical premise that in order for free will to be involved in decision making, the decision needs to be made consciously. Unconscious decisions aren’t up to us and therefore don’t display free will. So far, then, we have the following two propositions:
1. In various experiments, participants decide unconsciously.
2. Only consciously made decisions can be freely made.
How do we get from here to the conclusion that free will doesn’t exist? A common response is a third proposition:
3. The way participants decide in these experiments is the way people always decide.
If 1 and 3 are both true, and if the way the participants decide is unconsciously, we have the result that people always decide unconsciously.
There are several problems with the argument. I’ll discuss just two of them. Participants in these experiments are instructed to perform a simple action whenever they want and then report on when they first became aware of an urge, intention, or decision to perform it. In some studies, they are told to flex their right wrist – or click a key on a keyboard – whenever they want. In others, they have the option of pressing either of two buttons whenever they want. Nothing hangs on when they flex or click or which button they press. Any decisions participants make about these simple actions are arbitrary. In fact, participants are instructed to be spontaneous rather than think about what to do.
The discerning reader will have noticed something interesting already. The instructions participants receive place conscious reasoning about what to do out of bounds. The experimental setting is very different from a situation in which you’re carefully weighing pros and cons before making a difficult decision – a decision about whether to change careers, for example, or about whether to ask for a divorce. It would not be at all surprising if your conscious reasoning made it highly probable that you would consciously make any decision you made. At any rate, in light of salient differences between an arbitrary unreflective selection of a moment to act or a button to press, on the one hand, and a choice about a momentous matter made after painstaking conscious reflection, on the other, we can’t be confident that all decisions are made in the same way.
The problem just described pertains to proposition 3. Here’s a problem for proposition 1. The data are consistent with – that is, do not contradict – the following hypothesis: the brain activity that experimenters are measuring several hundred milliseconds or several seconds in advance of the action gives rise to additional brain activity that is a conscious decision, and that conscious decision plays a part in producing the action – the flexing, clicking, or pressing. There is no good reason to believe that the early brain activity (measured in seconds with fMRI and in milliseconds in the other studies) is correlated with a decision that is made – unconsciously – at that time. The data leave it open that any actual decision is made much closer to the time of action; indeed, they leave it open that decisions are made around the time participants say they are conscious of making them, often around 200 milliseconds (two tenths of a second) before muscle motion.
Why do you care about free will? I doubt that the ability to make spontaneous, unreasoned choices in experiments of the kind I’ve been discussing has a prominent place in your answer. Whatever answer you give will tell me something about what you mean by “free will.”
In my own research and writing, I’ve worked with two different but overlapping ways of understanding what “free will” means. One is more modest than the other, and I keep them both on the table; I don’t choose between them. As I see it, both encompass the ability to learn from our successes and mistakes and the ability to improve our behavior in light of what we learn. These abilities are important not only for personal development but also for social cohesiveness. Success and failure often depend on how other people respond to our actions.
According to a modest conception of free will, as long as you’re able to make rational, informed, decisions when you’re not being subjected to undue force and also are capable of acting on the basis of some of those decisions, you have free will – at least at those times. (Being threatened with a loaded gun is a good example of undue force.) According to a more ambitious view, something crucial must be added to these abilities: If you have free will, then alternative decisions are open to you in a way requiring that the natural laws that govern your brain activity sometimes give you at most a probability of deciding one way and a probability of deciding another way. Imagine someone who is seriously considering cheating on his taxes while filling out his 1040. The ambitious view says that he can’t make a free decision about this unless there is a real chance – left open by the combination of everything that has already happened and the laws of nature – that he will decide to cheat and a real chance that he will decide to be honest.
People tend to find ambitious free will more exciting than its modest counterpart. So I focus on it here. Most people assume that the future is open in a certain way. As they see it, not only don’t we know now exactly what we will do next week, but it also is not determined now exactly what we will do then. What will happen is partly up to us in a way that it could not be if all our actions were already in the cards, as it were.
The existence of ambitious free will depends on the truth of this assumption. Have neuroscientists shown that the assumption is false? Absolutely not. In the fMRI study I mentioned, scientists were able to predict with 60% accuracy, about seven seconds in advance, which button a participant would press next. Obviously, this does not suggest that it was determined which button would be pressed seven seconds before the action. After all, the evidence leaves a 40% chance that the participant would press the other button. In the study using direct readings from the brain, experimenters were able to predict with 80% accuracy, within a window of a few hundred milliseconds, what time participants would identify as the moment at which they first became aware of their intention to click. The scientists were able to do this about 700 milliseconds in advance of the “awareness” moment participants identified and about 900 milliseconds before the click. These findings do not support determinism. In fact, they are consistent with the idea that even less than a second before participants click a key it still isn’t settled when they will click next. Believers in ambitious free will thrive on probabilities of action, and that’s exactly what we find in these studies.
That we have ambitious free will – at least some of the time – is a definite possibility. One of the morals of the two books of mine that I mentioned is that neuroscientific studies of decision making leave this possibility wide open, in addition to leaving modest free will intact. This is good news, both for individuals and for society. There is evidence that lowering people’s confidence in the existence of free will increases bad behavior – cheating, stealing, and aggressive behavior. And there is evidence that belief in free will promotes personal well-being. If free will is real, beneficial beliefs in it have the virtue of being true, and it’s always nice when goodness and truth are on the same side.
An important implication of the free will debate – that is, the actual debate taking place in scientific and scholarly books and articles and in books and articles for the general public – is that we can easily be misled by scientific findings if we don’t interpret them carefully. When we pay attention to details, we see that the neuroscientific challenge to free will is misguided.
1. What does “free will” mean?
2. Why is free will important?
3. What are the most powerful arguments for the nonexistence of free will?
In my essay or initial post, “What are the implications of the free will debate for individuals and society?”, I approached the topic of free will from a scientific angle, closing with the following remark: “An important implication of the free will debate – that is, the actual debate taking place in scientific and scholarly books and articles and in books and articles for the general public – is that we can easily be misled by scientific findings if we don’t interpret them carefully. When we pay attention to details, we see that the neuroscientific challenge to free will is misguided.” The subsequent discussion was lively and wide-ranging.
Our conversation about neuroscientific challenges to the existence of free will raised such questions as the following: When do we become aware of our decisions and why does that matter? How is consciousness related to free will? What is a decision? What place does conscious reasoning have in decision making? Can we make free decisions unconsciously?
We discussed other scientific challenges to free will, as well – especially challenges having to do with automaticity and with the action-causing power of the situations in which we find ourselves. An intriguing question that arose in this connection is how well we have to understand what’s moving us to act in order to act freely. More briefly: How much self-understanding does free will require? We discussed some classic “situationist” studies, including a study of the behavior of bystanders and a study of obedience.
Much of our discussion was theoretical. A question that occupied us for quite some time was what “free will” means. The question took various forms, and different aspects of the question came in for scrutiny at different times. Here are some examples. Where should the bar be set for free will? Does free will depend on non-physical souls or minds? Is free compatible or incompatible with determinism (as physicists and most philosophers who write about free will understand determinism)? Is free will the kind of thing that can come in degrees, or is it an all-or-nothing matter? How is free will related to moral responsibility?
In my initial post, I offered two different but overlapping ways of understanding what “free will” means. According to a modest way of understanding this expression, as long as you’re able to make rational, informed, decisions when you’re not being subjected to undue force and also are capable of acting on the basis of some of those decisions, you have free will – at least at those times. According to a more ambitious view, something crucial must be added to these abilities: If you have free will, then alternative decisions are open to you in a way requiring that the natural laws that govern your brain activity sometimes give you at most a probability of deciding one way and a probability of deciding another way. Some discussants wanted to require more for free will than the ambitious view does, and some additional options were explored.
Some of the interesting questions that arose cannot be neatly categorized. They include the following: Can philosophers and scientists productively study free will in an interdisciplinary way? What are the effects on our lives of our belief – or disbelief – in free will? Is it rational or irrational to want to have free will? What is the bearing of physics on free will? How might the falsity of determinism contribute to free will? How much control do we have over our behavior? Is it possible, in principle, for science to prove that free will is an illusion?
As you can see, we tackled lots of very big questions about free will. I’ll close with two further big questions:
- What is the best approach to studying free will?
- Why is there so much disagreement about what “free will” means or about what having free will requires?