The Solution to the Problem of Free Will

Guest post by Danny Frederick

The traditional problem of free will arose out of theology. God was supposed to know everything, in which case He would have perfect foreknowledge. But if God knows beforehand what we are going to do, how can we be free either to do it or not?

In the Enlightenment, the problem arose out of science, specifically, the success of Newton’s theory. For here was a clear and (relatively) simple theory of the world, expressed in precise mathematical formulae, which made predictions about the motions of every kind of matter, whether large or small, whether on earth or in the heavens, whether solid or fluid; and the predictions were borne out by experimental testing. Furthermore some of the predictions were very surprising and counterintuitive, so people expected them to fail; yet they also survived testing. The theory was even used to predict the existence, size and orbit of a planet in the solar system which no one had previously known about; and when telescopes were pointed in the relevant portion of the night sky, there it was! Neptune was discovered. People could not believe that this theory could be false. It was generally accepted as firmly established truth. The success of Newton’s theory marked the transition from the open-minded and critical Renaissance to the dogmatic Enlightenment.

It is debatable whether Newton’s theory is deterministic, but it was taken to be so in the Enlightenment. It was thought that the physical world consists of atoms in space subject to Newtonian forces and that, if we know where all the atoms are at a particular time, we can, in principle, predict where they will all be at any later time just by using Newton’s laws. Since human actions involve physical motion, how can we be free to either perform or not perform any particular action? Plainly, we cannot: how a person’s body moves is determined by the position and momentum of atoms, and the forces that were acting on them, billions of years before we were born.

It was the acceptance of Newton’s theory, along with the interpretation of it as implying determinism, that led people either to deny free will or to try to show, with sophistical arguments, that free will and determinism were actually compatible, along the following lines: an agent can do as he chooses (and is thus free, in that sense), but his choice is determined.

Since the 1930s, we have had quantum theory, which has indeterministic laws. That, it might seem, solves the problem of free will. There is some spielraum for us to be able to influence physical events, and thus to act in the world, because physical events which are physically undetermined may, in some cases, be open to influence from mental acts of will. Unfortunately, the spielraum allowed by quantum theory is rather small: what we need is a conception of laws of nature as ceteris paribus and a theory of emergent evolution which permits higher, emergent, features, such as acts of will, to interfere with lower ones (‘downward causation’).  Of course, that will not avoid determinism unless the acts of will are themselves undetermined. But that seems to generate a new problem. If our acts of will are undetermined, then they are random; but if they are random, we have no control over them; but if we have no control over them, then we do not act freely. As Hobart put it:

In proportion as an act of volition starts of itself without cause it is exactly, so far as the freedom of the individual is concerned, as if it had been thrown into his mind from without – ‘suggested’ to him – by a freakish demon… it is just as if his legs should suddenly spring up and carry him off where he did not prefer to go [“Free Will as Involving Determination and Inconceivable Without it,” Mind 43 (169): 1-27 (1934)], available here.

However, that problem is a confusion because it fails to distinguish actions from mere events. A random event is one that we do not control. But an undetermined action is not simply a random event; it is an undetermined action. And an action is something that an agent does, something an agent controls.

Every action is an act of will; but some acts of will are not actions. The person who tries unsuccessfully to move his paralysed limb performs an act of will, but no action. An act of will becomes an action by bringing about a motion of the agent’s body. An act of will is under the agent’s immediate control at the time of its occurrence. That is, the agent does not perform an act of will by doing something else: he simply acts.

An act of will is essentially undetermined in this sense: there is no time before it occurs at which the total state of the universe at that time gives the act a probability equal to 1 or to 0. For, if the act of will, some time before its occurrence, had a probability of 1, it would be certain to occur, and if it had a probability of 0, it would be certain not to occur. Either way, its occurrence would not be under the agent’s control at the time of performance, so it would not be an act of will.

Acts of will are not entirely unpredictable, because they have a probability of occurrence between 0 and 1 and we may often know roughly how probable a particular action of an agent is. Indeed, most of our actions are habitual, so they have a high probability of occurrence, which makes us largely predictable. That enables us to plan moderately well in most circumstances. Further, there are things that we can do, such as offering incentives, that will raise the probability that a specific agent will act in a specific way, so we are able to influence, but not to determine, acts of will. For, an agent can always act against habits and incentives, because our actions are not determined. Even if there is a very high probability, say 0.95, that a particular agent will perform an act of will of a particular type, it is still up to the agent whether or not she acts in that way.

When my act of willing my arm to move brings about the motion of my arm, that act of will becomes an action of my moving my arm. The action is not the motion of my arm; and the act of will does not bring about the action. The act of will is the action in virtue of bringing about the bodily motion. Compare: when a woman produces a baby, she becomes a mother. The mother is not the baby; and the woman does not bring about the mother. The woman is the mother.

If I flip a switch by moving my arm, my act of will becomes an action of flipping a switch. If I intended to flip the switch by moving my arm, then my moving my arm and my flipping the switch are both intentional. If I intended to move my arm but flipped the switch by accident, my flipping the switch is unintentional. Unintentional actions are done ‘unknowingly,’ so they are not free actions. Intentional actions are free if performed by a critically rational agent, that is, one who can question and criticise his inherited views and habits and decide for himself how to act. So my one act of will might be both a free (intentional) action of moving my arm and an unfree (unintentional) action of flipping a switch. Non-human animals perform undetermined acts of will and some of their actions are intentional; but they do not have free will because they are not critically rational.

Contemporary philosophical writing on free will is confused. First, most contemporary philosophers seem to assume that determinism is true: they generally adhere to the view that human behaviour has a deterministic explanation, though the laws will not involve psychological terms such as ‘belief,’ ‘desire,’ ‘intention’ and such like, but rather descriptions in terms of neurophysiological events. In short, the old dogma that was a response to the astounding success of Newton’s theory is retained by contemporary philosophers. But Newton’s theory has been rejected as false since the advent of relativity theory in the early part of the twentieth century, and determinism is rejected in the most fundamental physical science (quantum theory).

Second most contemporary philosophers confuse action with bodily motion, though bodily motion of a special kind, usually distinguished by its causes (beliefs and attitudes, or intentions). But nothing can make a bodily motion into an action. Only an act of will can be an action. A bodily motion is not an act, though it may be brought about by one.

Third, most contemporary philosophers think that an undetermined act, even if it has a specific probability between 0 and 1, is a random event outside of the agent’s control, and thus not a free act. But that confuses an undetermined act of the agent with a random mere event outside of the agent.

Fourth, some contemporary philosophers maintain that agents determine their actions. But agents can do no such thing: agents act and their acts usually bring about bodily motions which in turn have other effects.

For more detailed discussion see my ‘Free Will and Probability,’ Canadian Journal of Philosophy 43 (1): 60-77 (2013), available (gated) here, and (ungated) here.

Also, my ‘Popper and Free Will,’ Studia Philosophica Estonica 3.1: 21-38 (2010), available here and here.

Danny Frederick

This entry was posted in Blog. Bookmark the permalink.

15 Responses to The Solution to the Problem of Free Will

  1. Mark Friedman says:

    Hi Danny:
    Thanks for this excellent post, which also serves to introduce your “Free Will and Probability,” which I believe should be read by anyone interested in this subject. I would welcome your thoughts on the following: is the anti-free will stance is compatible with the practice of philosophy. To be (I hope) more precise, it seems to me that philosophers routinely engage in acts of will, i.e. judging, deciding, concluding, etc. that arguments are sound or not. How do those philosophers who deny free will plausibly describe what they are doing?

    I would think they would say that while they are under the mistaken impression that they are doing such things, in reality their output is merely the product of some causal chain they do not control. But then why should anyone engage with them, since if they are not capable of intentionally seeking the truth, there is no obvious reason to conclude that they are producing anything other than noise. By the same token, on their view, other philosophers are not capable of being persuaded by their arguments, so the whole exercise would seem rather pointless.

  2. Hi Mark,

    The objection you articulate has been made by Karl Popper in his ‘Language and the Body-Mind Problem’ (Chapter 12 of ‘Conjectures and Refutations’), ‘The Self and its Brain’ (with J. Eccles, sections 20-21 of Popper’s contribution), and ‘The Open Universe’ section 24. But the basics of the argument, as Popper points out, go back to Haldane, Descartes, St Augustine and Epicurus (he could also have mentioned Kant, ‘Groundwork of the Metaphysics of Morals,’ p. 448).

    Here is what Epicurus says: “He who says that all things happen of necessity cannot criticise another who says that not all things happen of necessity. For he has to admit that the assertion also happens of necesity” Aphorism 40 of the Vatican Collection.

    Some philosophers respond by referring to computers that perform logical operations. That, though, seems to beg the question: computers are engaged, at best, in a simulation of thought, not in thought itself. John Searle argues this in his famous thought-experiment of the Chinese Room.

    I think that the determinist philosophers would have to argue that thought can be reduced to mechanical operations of a sufficiently complext sort. You get that sort of approach in the work of J M Fischer. But I think it can convince only those who are already committed.

  3. Simon says:

    Let me say at the outset, that I don’t believe I have a solution to this seemingly never-ending problem, just a few thoughts.

    1. One of the first things we need to do is decide just what is meant by the concept of free will. Hume, for example, seems to see it as a matter of lack of external constraints. It is fairly easy to believe in free will in this sense. It is what we mean when we say something like “I did it of my own free will”. What we mean, roughly, is nobody else forced me to do it.

    2. Another conception is that an action is undetermined, yet not random. This is a far more difficult idea to grapple with, since all we know of the world is the idea of cause and effect. And for each cause there is an anterior effect or effects. Free will then seems to suggest there is something beyond this. An effect that is not caused (determined) and yet is somehow caused as well. Yes, we can suggest there is something about the human mind that is able make choices without anterior causes being the sole reason for the choice, but it still appears quite mysterious.

    3. When we think about cause and effect, we often think of the billiard ball example. One ball hits another at a particular angle and velocity and we can make exact inferences about the velocity and angle of the ball that is hit. It’s an obvious example of determinism in action. But not all cause and effect is like that. Often the causal field is huge and the cause and effect relationships are difficult to understand. When we, ourselves are at the centre of a situation of this nature, do we think it is free will in action, simply because we are unable to see or understand the cause and effect relationship?

    4. There can be little doubt that some of the actions we believe we take of our own free will have anterior causes preceding them, some of which occurred before our births. This must make us think that the area where libertarian free will is involved might be quite small. This is not to say it is unimportant.

    5. There is little doubt that we believe we have free will when we consciously choose to do x or ¬x. The question is whether this is an illusion. Maybe part of the reason we see the choice as free is that we are unaware of the different events taking place in our brains. We know that neurons move etc. This much science can now tell us. But we are certainly not aware of events at the neuronal level. Experiments by Libet seemed to indicate that actions are initiated before we are aware of having made a choice. I accept later evidence may prove that this is not really the case, since the experiments were quite crude.

    6. I’m not convinced that arguments relating to quantum theory are very helpful. We don’t even know whether quantum behavior has a deeper underlying non quantum basis. It may simply be an aggregate of the latter. Certainly quantum behavior doesn’t seem anything like what we mean when we talk of making free choices.

    7. None of this gets us much closer to solving the problem and I’ve reached the conclusion that I’m not clever enough to solve it. But lately I’ve also wondered more whether it makes much difference one way or the other. Maybe, if we think that we make choices freely and it is important to us that we are able to make choices without constraints imposed by others, that is the more important consideration.

    • Mark Friedman says:

      Hi Simon:
      I’ll let Danny respond to the more technical aspects of your thoughtful comment, but I believe that one of the primary conventional answers to your last question is that the ascription of moral responsibility depends on free will. Do you not see this as a posing a problem for the free will denier?

      • Simon says:

        Yes, I see it as a problem, but to suggest that this means we must therefore believe in free will, surely would be to get things the wrong way round.

        • Mark Friedman says:

          Hi Simon:
          Didn’t mean to suggest that, but the ability to ascribe moral responsibility is an important “difference” between the two positions. Also, I would suggest that since across cultures most people appear to think that it is just to punish based on the underlying assumption of moral responsibility, they are implicitly endorsing free will. Obviously, free will might yet be an illusion, but perhaps this essentially universal impulse offers some support for “the ability to choose otherwise.”

  4. Hi Simon,

    Here are my responses to the issues you raise.

    1. Lack of external constraints concerns freedom rather than free will.

    2a. I think that free acts of will are random, in the sense of undetermined. But they are not random events, over which we have little control. They are undetermined acts. As acts, they are things that we do, not things that merely happen to us. And to say that they are undetermined is just to say that we could have acted otherwise. It seems to me that undetermined acts of will are mysterious only if one confuses them with undetermined events which are not acts.

    2b. It is false that all we know of the world is cause and effect. For one thing, we know about free acts, which are not effects of a cause, because they are undetermined. For another, we know that sub-atomic particles are indeterministic: we can often not say that a particle WILL behave in a particular way, only, at best, that theres is a specific numerical probablity that it will do so.

    3. The question you raise at the end of this point was answered affirmatively by Spinoza and Schopenhauer, amongst others. But they were creatures of the Enlightenment and its attachment to Newtonian determinisim, whose foremost exponent was perhaps Laplace. But things have moved on since then (see point 6, below).

    4. I do not know why you say “There can be little doubt that some of the actions we believe we take of our own free will have anterior causes preceding them, some of which occurred before our births.” That seems patently false to me. An act of will is an event which is immediately under the control of the agent whose act of will it is. That means it cannot have been determined; otherwise, it would not have been under the agent’s immediate control. Suppose that I perform the action of turning on the light. How did I do that? I did it by flipping the switch. And how did I do that? I flipped the switch by moving my arm. And how did I do that? I moved my arm by willing it to move. And how did I do that? I just did it. To avoid an infinite regress we have to stop at an act that is done ‘just like that,’ rather than by doing something else. So there can be no action unless there are acts which are under our immediate control. So, if determinism were true, there would be no action.

    5. You return here to the position of Spinoza and Schopenhauer. i concede that free will may be an illusion: later scientific discoveries might conceivably show us that. But it would mean that we never act. There is quite a bit of discuission of Libet: the consensus seems to be that his experiments do not show what he claims they show.

    6. I agree that quantum theory might later be replaced by a deterministic theory; but even if that happens, we should also concede that the deterministic theory may later be replaced by an indeterministic one. Quantum theory concerns sub-atomic events, which are not actions, but are mere events outside of our control. We can set up an experiment in which there is a 0.5 chance of a particular event happening: that much is under our control. But whether the specific event happens or not is outside of our control – unlike our undetermined acts. The significance of quantum theory is that it shows that it is not the case that science necessarily presumes determinism.

    7. Your last point sounds like Strawson, ‘Freedom and Resentment.’

    If you want to retort to any of these comments, please do.

  5. Simon says:

    1. Quite probably – I’m making a comment about natural language I suppose, with the thought that this is something that often causes errors in the discussion.
    2a. If the key thought is that we could have done otherwise, then I concede that in one sense, I could have done otherwise, but perhaps ultimately I could not. I could have done otherwise in the sense that I was capable of doing the one or the other and I could have done otherwise in the sense that I think I made a free choice. But neither implies ultimately that I did. That, it seems to me, is what we really wish to explore. What is it about me, about my brain, that allows a genuinely free choice, without anterior causes playing their part?
    2b. You say we know about free acts, but is this not surely precisely what we are seeking to explore? As stated previously, quantum indeterminacy seems to me to be in a totally different category to free will. If we concede quantum indeterminacy, does it really tell us anything useful about free will? I wonder if we think dogs have free will. Certainly they might decide to go up one path rather than another, but do we want to put it down to free will? Why not habit or the smell of food or whatever, which would seem to suggest the choice is determined in some way. If only humans have it, how do we explain that. If dogs have it as well, where does it stop? Insects? Amoebas?
    4. Do you not think that some aspects of my education, say, have an influence on decisions I now make?
    5. I concede the Libel point. On the other, I concede that, under one definition, I do not act. My point is, so what?
    7. I shall investigate the Strawson paper.

    • 1. I agree. The idioms of ordinary language are often confused and even more often false.

      2a. The question of whether you could have done otherwise seems simple to me. If what you did was determined, that is, had a probability of 1 relative to some prior state of the universe, then you could not have done otherwise. Since you did it, it could not have had a probability of 0. Therefore, if you could have done otherwise, the probability of what you did, given any actual prior state of the universe, was between 0 and 1; which is to say that what you did was undetermined. What makes it possible for you to act freely? Two things. First, that some of your behaviour is undetermined; second that some of your undetermined doings are acts. I do not try to define ‘act.’ I assume it is a notion we are all familiar with; though the implications of the notion can be clarified in discussion. If you are capable of undetermined acts, then you are capable of acting otherwise than you do.

      2b. As I said before, quantum theory is important because it shows that science does not assume determinism. As you say, it does not (explicitly) concern free acts, but it gives us at least an analogy. I argued in section 5 of my paper on Popper and Free Will, which is available here,

      that only acts of agents which are capable of self-consciousness qualify as free acts. Dogs do not appear to be self-conscious, though dolphins and chimpanzees and even some birds do seem to have a rudimentary form of self-consciousness which makes them borderline cases of free agents (assuming that they are not just determined pieces of clockwork). Habits are not deterministic: they are propensities. We can act against our habits.

      4. Yes, I think that some aspects of your education have an influence on the free decisions you make. All sorts of things have such an influence. But such influences only make some decisions more probable than others. So long as the probability is between 0 and 1, your decision is not determined and you have the freedom to decide otherwise.

      5. You express insouciance over whether you act or not. Fair enough. Some people express insouciance over whether there is life after death or whether Newton’s theory is true or whether the human race will destroy itself in one way or another. I cannot make anyone care about the problem of free will. But for those who are interested, I have offered a solution which, contrary to the contemporary consensus in academic philosophy, shows that the existence of free will is possible and plausible.

      7. Strawson’s paper is available here:

  6. Simon says:

    2b. Yes, I can understand that tself-consciousness does seem to be associated with this issue. Somehow, perhaps, our ability to reflect on issues and apparently take conscious decisions seems very bound up with the idea of having free will. I will have to reflect more on that. Ditto, the thought that the probability is <1.

  7. Here is a nice piece, by professor Helen Steward, which is entirely consistent with the account I adumbrated above:

Leave a Reply

Your email address will not be published. Required fields are marked *