A concise statement of the problem of free will

John C. Wright recently posted on Determinism and Indeterminism and quoted this concise statement of the problem of free will, which was written by a reader of his called Lucky Marty:

Consider the following propositions:

  1. All events are either the deterministic result of prior events or else they are random.
  2. Free choices are not fully determined by prior events.
  3. Free choices are not random.
  4. People make free choices.

All of them seem highly plausible, and in fact it’s not easy to see how any of them could be false. But they can’t all be true

This is the same point I’ve been making, but stated in a clearer way.

So, which of these four propositions is wrong? The first and third are undeniable, so that leaves the other two. My approach is to reject #2. Keeping in mind that “prior events” includes one’s own prior thoughts and desires, I think we could still be free in most important senses of the word even if our choices are fully determined — that, in Daniel Dennett‘s phrase, we could have all “the varieties of free will worth wanting” even in a fully deterministic universe.

But if you insist on #2 — if you define freedom in such a way that it precludes determinism — then I will insist that you reject #4. Free will, thus defined, does not and cannot exist — not because the world happens to be deterministic (which may or may not be the case), but because it is logically impossible.


Filed under Philosophy

9 responses to “A concise statement of the problem of free will

  1. bgc

    My feeling is that this is a process of discovering that mainstream ‘materialism’ is self-refuting, since when you get an unavoidable absurd conclusion from your premises that means at least one of your premises are false.

    The underlying problem is that this statement operates purely at the level of science, assuming the metaphysics of science is complete and the only possible metaphysics.

    In other words the next step is to metaphysics.

    From a metaphysical perspective, all of the four statements are challengeable since they need to be embedded in an explicit metaphysic.

    For example. statement 1 is regarded as wrong by Aristotelian logic, since it leads to an infinite regress (a causes b causes c forever) – it is also wrong because (to jump ahead) it is an example of reasoning, and if reasoning can only be determined or random, then reasoning has no validity.

    Another aspect which strikes me is that statements 2 and 3 are negative statements about what free will is not, statement 4 is a kind of ‘sociological’ statement, or a ‘people believe’ statement – but it is premature since you have not said what free will IS only what it is NOT.

    I don’t see how a fundamental metaphysical thing like free will can be made the subject of syllogistic reasoning until it is know what free will IS – at least what its properties are supposed to be, what work is it doing.

    Reflection then makes it clear that the concept of free will is doing work at a transcendental level – more fundamental than science, more fundamental than these statements.

    Ed Feser is very good on this stuff, from a Thomist perspective- his book on Aquinas, or his excellent anti-atheist book The Last Superstition – http://edwardfeser.blogspot.com/

  2. You stated “The first and third are undeniable”. Not to be contrary, but I don’t see either one as true.

    #1 is a false dichotomy. The logical negation of “the deterministic result of prior events” (commonly called causal) is “containing an element which is not deterministically caused by prior events”. But “Not deterministically caused by prior events” is not equivalent to “random”. For example, Quantum entangling can result in events that are perfectly correlated to concurrent events (non-random) without being correlated to prior events (non-causal). Precognition would also be an explanation for non-random non-causal events.

    If we were to abuse the word “random” to force #1 to be a true dichotomy, then #3 becomes (by double negation) “free choices are the deterministic result of prior events”, which is both far from obviously true and is the exact opposite of #2, and leaving us nowhere.

    #3 is probably false, depending on the definition of “random” used. I most commonly see random defined in one of three ways, two formal and one colloquial. Note that none are the logical negation of “causal”.
    Random 1: “independent of and uncorrelated with measured phenomena”. This is typically assumed true of free will; it’s what we mean by free. This definition leaves #3 false.
    Random 2: “a recurring mathematical input which may be characterized in aggregate but not individually”. This might be true of free will, if choices can be characterized in aggregate. If not, then free will fits into one of two more general mathematical classes: bounded unmodeled input, which requires game theory, a mathematical model that assumes prescience; or unbounded unmodeled input, which is so disruptive no mathematical model can be meaningfully utilized. It is thus not clear if this definition conflicts with #3 or not; an effort to characterize free will in aggregate would be necessary before this could be known.
    Random 3: “without meaning, purpose, cause, …”. This does preclude free will by definition, so by picking it #3 can be true; but it is even less like the negation of causal than are the two formal definitions above, causing #1 to fall even farther into the realm of nonsense.

  3. Thank you both for your comments.


    I don’t see this as a refutation of materialism, because none the premises all seem to be common-sense axioms, the truth of which does not depend on materialistic assumptions. They seem equally self-evident even if you assume the existence of God and spirits.

    The idea that reason has no validity if it is deterministic is something I’ve heard before, from Lewis and others (Victor Reppert bills it as “C. S. Lewis’s Dangerous Idea” in his book of that title), but I don’t really understand it. Reasoning, and deductive reasoning in particular, is essentially a matter of following an algorithm, and it’s not immediately obvious why it would require free will in order to be valid. Certainly a computer can make valid mathematical calculations by following a deterministic set of instructions.

    I agree with your main point, though, that the premises here are all (not only #4) non-rigorous sociological statements about what people generally believe, and that a clear definition of free will itself is needed. The premises do represent what people generally believe, though. When people just “know” by common sense that they have free will, it generally means that they subscribe to this inconsistent set of propositions. Free will as intuitively understood by non-philosophers is a self-contradictory concept. Any attempt to define free will in a non-contradictory way will involve rejecting this popular conception.


    You are of course correct that terms such as “random” need to be defined more rigorously in order for this argument to be really useful.

  4. Bruce Charlton

    Sorry if I misunderstood – as a sociological observation of the incoherence of current public discourse, this is true enough – although perhaps it would be more accurate to introduce an assymetry, in that people often believe they themselves have free will, while regarding the behaviour of others (or some others) as dictated by prior events.

    Most famously this is seen in Marxism, where everyone is regarded as a passive subject of economic forces, merely a representative of their class – except for Marxists, who are the only free agents, and therefore the only legitimate rulers.

    The self-serving error is obvious, but that didn’t and doesn’t stop the idea being believed.

  5. i was just the other day thinking about gravity and quantum mechanics ( very lightly )
    And it occurred to me that the principle problem with ‘spatializing’ how gravity works or how electrons can absorb photons and such; Is that we invariably try to mash their activity into how we perceive things happening in the macroworld.
    Specifically; How can Exchanging Particles possibly explain in any way how electricity works or how ‘curved space’ would explain gravity, You still need an understanding of something pulling or pushing any given ( x ) along this curved surface.
    At the Quantum Level ( Teeny Tiny ) of things; i think that we have to make a leap of ‘Incomprehensibility’ in that things on this level have NO Correspondence with how things work at The Table Top Level.
    The same thing applies to Free Will or Autonomy.
    The Rules change at A Tangential Level when trying to understand these things.
    It may seem like a cheap trick to simply say that Another ‘Incomprehensible’ layer of reality is responsible for Gravity, Electricity & Freewill, But given that these things are all impossible when you try to make them fit into The World of Everything that We have Labels for, The Contradiction or Paradox forces our hand.
    Freewill is necessary for Consciousness, which is apparent, Assuming that Robots may behave exactly like people, but not be “Aware” of what they are doing.
    How can a Pattern of ElectroChemical Disparate Neurons produce Consciousness?
    Does every Pattern, of Anything, Generate, by this same mechanism, produce an instance of Consciousness?
    This seems crazy, But this rule of ‘Incomprehensibility’ kicks in again.
    We simply don’t have the proper vocabulary of labeled things or ideas to discuss these things.

  6. i have a new thought on this:
    Determinism; Which implies that everything that is currently happening, was decided, or set into motion at some point (x) when the universe started—
    Which; given the current state of order & structure in the universe, particularly in regards to humanimal existence, seems very improbable.
    Granted; In a universe that is infinite, anything ( ? ) is possible,
    But setting that aside; The current orderly condition of the universe strongly implies that there was ‘something’ shaping or modeling the universe as it proceeded along.

  7. Free will is incompatible in both a deterministic universe as well as an indeterministic universe (Hard Incompatibilism).

    1, 2, and 3 all hold. It is the 4th proposition that is wrong. Dennett’s “varieties of free will worth wanting” are not free in the selse of allowing responsibilty (even though he suggests they are).


  8. Pingback: The necessity of agency | Bugs to fearen babes withall

Leave a Reply to Luther Cancel reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s