The determinism dilemma is that if our actions are predetermined they are not free, and if they are random they are not willed, either way there is no free will. Even if will causation is a mixture of chance and necessity it can be split into the corresponding aspects according to an argument that goes back to Hobbes, reinstating the dilemma: "in order to make the causation somehow indeterministic, a realization of a random variable (completely uncaused and independent) would have to be put amongst the conditions." To put it metaphorically, we are puppets dangling on the strings of causality, or chance, or both part time.

A popular response is the two stage model of free will, where at the first stage actionable alternatives are "freely" generated, and at the second the one that "best" suits our goals and desiresis "willed". Trouble is that the "free" stage is not willed, the "will" stage is not free, and the freedom itself is reduced to imperfection, inability to survey all possible alternatives. Although the model itself is relatively uncontroversial (both Dennett and Kane accept it!) it concedes the Hobbesian split into chance and necessity, and merely shifts the dilemma to the second stage.

But we know of at least one type of processes, where the Hobbesian split can not be accomplished, quantum evolutions. I am not suggesting quantum mechanics as a physical model of volition (but such suggestions were made, by Compton using macroscopic amplification of quantum effects, and Penrose using hypothetical "quantum gravity"), or even saying that it is a "complete" description of reality. But it is a model where the wave function "is all there is" by definition, and it is self-consistent if set theory is.

The wave function is neither deterministic, nor random, nor a "mixture" of the two (see Bell's inequalities). We get a model for events that provably can not be split into a mixture of chance and necessity along the Hobbesian lines envisioned by the dilemma. Free will does not have to operate anything like quantum evolution of course, the point is that combinations of chance and necessity do not exhaust logical possibilities, the two can be entangled in unsplittable ways.

Is the determinism dilemma a false dichotomy? Were non chance/necessity options explored by philosophers as models of how free will might operate? Specifically, how "goals and desires" can be factored into volition without resorting to partial causation?

  • 2
    c.f. philosophy.stackexchange.com/questions/11518/…
    – Dave
    Commented Dec 6, 2014 at 4:08
  • You might also see philosophy.stackexchange.com/a/31834/1547 Commented Jan 30, 2016 at 18:43
  • Randomness and free will can be compatible. Among meanings for randomness is non-computability, or new information / novelty. So random can mean many things and many of those are actually compatible with free will (eg introducing willed novelty). If free will makes any choice, since this choice is not predetermined, it is novelty, that is, coincides with a meaning of randomness
    – Nikos M.
    Commented Nov 6, 2021 at 19:00
  • Partial causation is perfectly compatible with FREE will, as long as it is not full causation (leaving aside approaches to free will compatible with determinism)
    – Nikos M.
    Commented Nov 6, 2021 at 19:41
  • This is a great question; one I've grown to understand is amongst the most intriguing in the 'free will' realm. I'm curious @Conifold, have you come to any answers of your own in the 7 or so years since posting? Commented Feb 6 at 9:13

7 Answers 7


It is not a viable option. Only chance or necessity are implied by the fundamental fact that all thoughts arise impersonally. As William James, among others, noted: "If we could say 'it thinks' the way we say 'it rains' we would be speaking with a minimum of assumption." When added to the other fundamental fact that all thoughts "governing" movement have the immediate effect of movement unless another thought intervenes, the minimum of assumption conclusion is that what we believe to be an act of free will is an automatic response to a stimulus of unascertainable origin. From a psychological perspective, only indeterminism, supportive of either metaphysical option of necessity or chance, is viable. My essay on James's vivid engagement with the question of free will was published in the anthology The Volitional Brain: Toward a Neuroscience of Free Will, and is available on my website.


The failure of the dilemma comes in denying that free will is a primary causal factor in the physical realm. To answer the dilemma, we must posit three possible primary causal factors which can drive physical causal chains:

  1. Deterministic inputs
  2. Random inputs
  3. Free Will inputs

However, there is no direct empirical evidence or proof for free will being a primary casual factor. We posit its existence based on experiential evidence and as a solution to other logical problems required to allow for some epistemologic systems to remain logically consistent (e.g. as a solution to the problem of evil, for example).

Whether there can ever be direct empirical data which explains the mechanism of free will is unknown. Free will requires additional parameters which make a universe with free will higher order than a universe without it, and thus Occam's Razor would applies in cases where free will is axiomatically assumed to have no utility. Of course, if free will does have utility, Occam's Razor would not apply.

This is why I believe that free will implies dualism, though there are other possible places to eventually "find" and potentially empirically test free will, such as extra dimensions, parallel universes, or alternate realities. All of these potential options could provide a means of satisfying the mechanism for dualism and are not currently empirically verifiable.

  • 3
    There is no direct empirical evidence for determinism either, all we have are infinite precision extrapolations from classical physical laws that we now know do not even apply with infinite precision. Both determinism and free will are arrived at by the same hypothetico-deductive approach, and at present determinism is not exactly suggested by physics. it will certainly take a massive infusion of non-locality and/or hidden variables to work.
    – Conifold
    Commented Jan 1, 2016 at 3:13
  • 3
    I am not sure any additional parameters or dualism are required for free will if we stop thinking of it as based on causes residing somewhere else. Such thinking seems to be an implicit form of Hobbsian split and looking for sufficient cause, which is a persistent habit. We are so used to looking for causes that we subconsciously try to insert them even into something defined not to have them, at least not sufficient ones. This may be the source of expecting dualistic "extras". Free will has to combine influence of the past and lawlessness somehow, but in unsplittable way with no extras.
    – Conifold
    Commented Jan 1, 2016 at 3:20
  • You make good points on both cases, I'm not sure I agree on the dualism point, but that may just be a matter of definition. I'm intrigued by your concept of just not looking for a cause of free will (i.e. impact from some other realm/plane of existence/dimension/etc.). Any links on that particular topic I could follow up with?
    – LightCC
    Commented Jan 1, 2016 at 9:37
  • I find Kane's work particularly insightful, he also challenges other stereotypes in posing questions about free will. This page has a good summary of his arguments with links to original works informationphilosopher.com/solutions/philosophers/kane
    – Conifold
    Commented Jan 4, 2016 at 5:14

Yes. But free will is a sub-option of a wider class of options that are neither random (stochastic), nor deterministic.

Basically, it is possible that the most complete physical theory is neither deterministic, nor based on (fixed) bayesian probability.

There are several mathematical theories (such as the Possibility theory, Evidence theory and Generalized Information Theory) of uncertain probabilities (such uncertainty is known as Knightian uncertainty).

Basically, an information source that is neither probabilistic, nor deterministic can be viewed as "oracle" or "input port" from outside of the physical world or unknown starting conditions of the universe, or as free will.

So, yes, it is possible that physics is fundamentally neither deterministic, nor probabilistic, but it can be explained not only with free will but also by input from the outside (even by God's intervention or by "the player" for whom our universe is a game).

Interestingly, Thomas Breuer had shown that even if there is no input from the outside of the universe, each observer will see the universe as not being probabilistic (of which being deterministic is a case).

This is because the state of the system where he is properly included can not be fully known to him due to self-reference (even probabilistically). So, there always will be unpredictable events (events with unknown probability). This can be interpreted as free will, of course, but also (as he had shown) as in-principle-unknown initial conditions of the universe (the observer cannot know the present state of oneself, so, he cannot know the past or initial state as well). So, such unpredictable events (or unpredictable information) comes either from free will or the force that created our universe or they are basically the same thing.

  • "outside of the physical world" - so, dualism - a common idea behind free will is a soul. "unknown starting conditions of the universe" - don't think this would contradict determinism, as determinism does not imply the past is computable from the present state (there can hypothetically be multiple different universes at moment T_0 that become equivalent at moment T_1- basically how many-one reductions work in computer science).
    – rus9384
    Commented Feb 4 at 21:23

Upon analyzing the determinism dilemma, I find that one part is false (or incomplete). I agree that all actions are either determined, or undetermined. For the determined actions, "free will" is not present. However, for the undetermined actions, there are at least two types, random and selected. Also, for the random actions, "free will"is not present. But for the selected actions, "free will" is present. This results from the application of "free will" by the selector!

  • 5
    The problem is that the selector selects based on something (history, circumstances, internal imperatives). If that something predetermines it the will is not free, if not then it is partly random and we are back to the dilemma.
    – Conifold
    Commented Feb 9, 2015 at 4:59
  • 1
    How does selected differ from determined? What one selects is determined by one's conscious or unconscious mind, which is determined by development, environment, and circumstances. Commented Dec 9, 2015 at 18:52
  • @Conifold: Although I agree that the selector might be "biased by something," the act of choosing is undetermined, until the choice is made. We are constantly making choices. Some times, the number of choices one has to make in a day, is overwhelming!
    – Guill
    Commented Apr 15, 2017 at 23:36
  • @Conifold "history, circumstances, internal imperatives" don't need to agree on a singular selection. Thus it may well happen that deterministic factors amounts to a zero sum, which could then allow for a free will selection.
    – christo183
    Commented Oct 27, 2018 at 11:14

No. The free will dilemma is not dependent upon how neatly we can separate the deterministic and random parts of a process. The dilemma is in the fact that no matter what creates your will, that will is not created by you prior. This is never even experienced subjectively at a closer look.

  • That's not a problem at all.
    – user71009
    Commented Feb 4 at 14:07

Short Answer Yes, if we have free will, it must be a 3rd option besides chance and necessity.

I accept Hume's split -- free will is incompatible with either chance or necessity.

Slightly longer answer

The efforts to find ways to breach Hume's split using Quantum mechanics, are interesting, but show less than their advocates assert.

Compton's model of using macro scale amplification of QM effects with dynamite, or the poison gas that kills Schrodinger's cat -- both show how the "QM effects are always averaged out at the macro scale" is untrue. Neither Schrodinger nor Compton had the advantage of our contemporary Chaos theory, where tiny inputs can affect macro outputs, but they both came up with Chaotic amplification on their own. Another example would be the cancer caused by a single quantum reaction that produced a cosmic ray. In chaotic systems like life and dynamite, significant macro scale outcomes can be triggered by single probabalistic events.

While these examples show that our world is not determined at the macro scale, they only provide room for free will to act, they do not show that free will DOES act, as the randomness of quantum events are as incompatible with free will as deterministic events are.

What is needed for free will

What we need, is a different conception of causation, where there is a causation source that is free-will compatible.

Our conception of causation has changed significantly over the millennia. Aristotle had 5 types of cause. Descartes allowed for only two -- contact cause, and the will of God. Drop out God, and you only have mechanical pushing, as Princess Elizabeth pointed out.

However, physicist were already studying the phenomena of waves, electricity, and gravity, which all operate at distances, and allow superposition, hence are not limited by contact-based pushing by solids. Newtonian physics, by allowing fields and superposition, does not limit causation, and allows anything that one sees producing coherent behavior to be a cause. The "contact-only causation" mechanistic model of causes is --refuted.

Physicalists presume that causation is limited to "physics", but then what is IN physics -- is undefined and undefinable, per Hempel's Dilemma.

Could willing be in physics? Or should we not care what is in or out of physics, per Hempel, and accept any causation so long as it is coherent? Both are routes to a causal model that includes free will.

The reasoning that articulates this idea is called "Agent Causation. Agent causation theory is summarized in: https://plato.stanford.edu/entries/incompatibilism-theories/#3

A recent advocate of agent causation is here: https://www.amazon.com/Metaphysics-Freedom-Helen-Steward/dp/0198706464

  • Free will is impossible purely because we cannot create what we will prior. It is not a matter of whether science has “room” for it. You cannot create room for an incoherent concept. Your will would have to exist prior to your will for free will to exist. But that’s nonsense Commented Feb 7 at 13:51
  • @Baby_philosopher -- The "logic" argument against free will presumes a particular logic, basically classical logic, which is also behind the either/or of caused/random. However, classical logic has exceptions -- it does not apply to all of the world. We have an intrinsic logical sense, which is even more flawed than classical logic, which when we apply that sense to itself and correct the obvious errors, we end up with classical logic. But this intuitive logic sense, like our intuitive mathematical sense, does not constrain our world.
    – Dcleve
    Commented Feb 7 at 17:21
  • @Baby_philosopher Mathematicians realized that they can construct infinite math assumption sets, and which one fits the world is an empirical question. Non-Euclidean geometry is the go-to example of our math intuitions being wrong. Logicians have reached the same conclusion. See cambridge.org/core/journals/think/article/… We have not found a logic yet that properly describes libertarian free will, but we will be able to find one. Your declaration of "nonsense" is based on an invalid assumption about logic.
    – Dcleve
    Commented Feb 7 at 17:25
  • The “logic” argument against a square circle presumes a particular logic, basically classical logic, which is also behind the either/or of caused/random. However, classical logic has exceptions -- it does not apply to all of the world. We have an intrinsic logical sense, which is even more flawed than classical logic, which when we apply that sense to itself and correct the obvious errors, we end up with classical logic. But this intuitive logic sense, like our intuitive mathematical sense, does not constrain our world Commented Feb 8 at 12:29
  • @Baby_philosopher. There is no logic argument against square circles. You are presuming that squareness and circles are logically contradictory, but this is not so. For some aspects of our world, non superposition of properties happens to be the case, and that is generally true of macroscopic solids, but not of most of the rest of our world. Quantum particles show particle and wave properties at different times, the two properties superimpose and do not exclude each other. The same is logically possible for circles and squares, our having no actual instances of this is just happenstance.
    – Dcleve
    Commented Feb 8 at 17:47

It is indeed a false dichotomy because determinism does not prevent us from having free will. Suppose our future is already determined. Even so, we still don't know it -- in particular, we don't know what our future actions will be.

Now, what happens when we don't know what we are going to do? We make our best guess based on our experience to date. And that -- guessing our choice -- is free will.

  • 1
    It sounds like you're equating freedom with ignorance of the future. That's to heavily diminish freedom in terms of what most people understand freedom to mean. It's an illusory freedom. It's kind of like saying the puppet is free because he's not aware there's someone pulling the strings. I'm not saying the illusion isn't effective; it is, but there's no real agency in what you describe. Is that a satisfying version of freedom to you? (I'm not claiming free will exists/doesn't exist here) Commented Feb 6 at 8:28
  • @Futilitarian — I think ignorance, as the other extreme, also robs us of freedom. Indeed, freedom comes from knowing our options. Us having the options, however, is due to the limits of our knowledge. Commented Feb 6 at 9:59
  • In other words, ignorance prevents us from knowing our opinions, while omniscience prevents us from having them (aka necessity). And in between we have freedom. Commented Feb 6 at 10:09
  • @Futilitarian while I am sympathetic to your intuitions, I don't think that the compatiblist account given here needs to be seen as a heavy diminishment of freedom. If free-will is about agent-causality then... how are agents implemented? Can a brain not be the physical implementation of an Agent? Can the workings of the brain not be the physical implementation of 'choice'? Why not? If we are agents, and we have freedom, and we have freedom because we have causal power as agents, then... perhaps a brain is exactly the kind of thing that can implment such a thing as 'agency'.
    – TKoL
    Commented Feb 19 at 16:27
  • The brain could be the 'physical implementation of an agent' if the brain originated its processes, but as far as we can tell, it's a physical system which interacts with its environment and is prone to external conditions which prevent anything original. We may one day find some originating mechanism in the brain, but to date we simply haven't been able to locate anything which separates the brain from the conditions in which it exists. (I hope I've addressed your point). Commented Feb 20 at 3:41

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .