Having very recently started getting interested in philosophy, I'm still halfway through my first book (Think: A Compelling Introduction to Philosophy, by Simon Blackburn, as recommended in this question).

Before reading it, it was my belief that we acted according to free will, not bound by any deterministic rules (is this one 'type' of incompatibilism?). If I came upon a t-junction on a road, I could choose whether I went left, right or back. Yes, I had the notion that I would make that decision based upon every knowledge that I had acquired until that moment, thus being bound by that. I understood that my decision was bound to the way every one my my past experiences influenced my thinking and decision making process. But, despite these constraints, I possessed free-will, and I was free to choose down which road I decided to go. I had the impression something like this happened:

A dotted line representing the past comes to a dot marked "Present", and then splits off into multiple directions marked "Future"

Here, the red dot represents a point in time where a decision was/is made. Also, the past was composed of several other ramifications such as this one, but here it can be represented as a straight line, considering that we would ignore all the choices I did not make.

However, after having read the book's chapter on Free Will, such a possibility is not even considered (or if it was, it was not clear to me). I took this easily, as I actually understand that it could be considered, as Blackburn puts it, 'bad philosophy'. This was clear to me, as soon as I read the first quote present on this chapter:

Again, if movement is always connected, new motions coming in from old in order fixed, if atoms never swerve and make beginning of motions that can break the bonds of fate and foil the infinite chain of cause and effect, what is the origin of this free will possessed by living creatures throughout the earth?
Lucretius, De Rerum Natura

Since we are indeed inserted in a physical world that responds to cause and effect, we must consider that we are too bound to this law. This makes it clear to me that determinism is very likely to 'be real' (as in be the 'thing' that really happens), since the past controls the present and the future, and we are not able to control the past, leaving us unable to control both present and future too (leaving us in control of nothing, basically). This would point toward incompatibilism, which I find easier to believe than its alternative.

So, Blackburn spends most of the chapter trying to reconcile free will with determinism, introducing the concept of compatibilism. This is where I have some trouble, for I really cannot see how the two concepts are conciliated.
He asks us to "think of the brain in 'software' terms, as having various 'modules'". Thus, we would have a 'scanner' module (which takes in information), a 'tree producer' module (that, according to what the 'scanner' says, produces options, similarly to the image I put above), an 'evaluator' module (which chooses which 'branch' of the tree is the best option) and a 'producer' module (which carries out the decision made by the 'evaluator'). This works very well, but (in my opinion) only when applied to the thinking I had before having accepted determinism (meaning the image above). I do not understand how the modules are not constrained by the same deterministic laws as everything else. Basically, this cartoon somewhat expresses how I feel:

A Dilbert cartoon. DOGBERT: "Do you think the chemistry of the brain controls what people do?" DILBERT: "Of course." DOGBERT: "Then how can we blame people for their actions?" DILBERT: "Because people have free will to do as they choose." DOGBERT: "Are you saying that "free will" is not part of the brain?" DILBERT: "Of course it is. But it's the part of the brain that's out there just being kind of free." DOGBERT: "So, you're saying the "free will" part of the brain is exempt from the natural laws of physics." DILBERT: "Obviously. Otherwise, we couldn't blame people for anything they do." DOGBERT: "Do you think the "free will" part of the brain is attached or does it just float nearby?" DILBERT: "Shut up."

So, throughout the chapter, the author refines a compatilist definition (that tries to reconcile the modules' free will and determinism), until he reaches this:

The subject acted freely if she could have done otherwise in the right sense. This means that she would have done otherwise if she had chosen differently and, under the impact of other true and available thoughts or considerations, she would have chosen differently. True and available thoughts and considerations are those that represent her situation accurately, and are ones that she could reasonably be expected to have taken into account.

So here I see a lot of could and would haves. This is where it really gets itchy, because the only thing I can take from this is that free will (according to these arguments) only exists in retrospective (or as an illusion in the moment when 'decisions' happen). Thus, I can perceive compatibilism as being viable only in retrospective, having my first image only as an illusion to a reality that could be represented as this:

An arrow in a straight line, with the beginning marked "past", the middle "present", and the end "future"

In this possibility (reality?) there is no red dot, since I cannot see how decisions can actually be real.
The thing is that Blackburn then goes on to explain some scenarios that always end up with a lot of could and would haves, and I cannot see them as plausible. The impression I get is that compatibilism is built upon a lot of could and would haves, which cannot be proven (or at least his arguments did not seem, to me, to be able to do it).

Going back to my first example, I cannot see how the road I picked at the t-junction was ever my choice. Sure, my 'modules' are at work, analysing everything in the landscape, etc. so that I can make a decision. Sure it really seems to me that the results provided by such an analysis are the object of my own free will, but how can I ever really prove it, if in the basis of my thoughts I know that the physical world in which I (thus the modules) belong is bound by deterministic laws? What I mean is: upon analysis, I have the impression that I can choose where to go. In retrospective, I too am under the impression that I could have chosen differently. But how can I prove that I really could have chosen differently?

To sum up:
Before I could (or thought I could) see how people could act according to their own free will, unrestrained by any type of determinism, although bound to the conditions 'available' to them (however, this seems to be bad philosophy).
I now can see how we are bound to a deterministic physical world (ruled by the laws of cause and effect), where free will does not exist.
But I cannot understand the middle term between the two views. How can our notion of free choice not be restrained by our physical world, and thus bound by the laws of causation? How can free will and free choice exist in such a world?

  • 1
    On a side note: I would prefer to 'believe' in compatibilism, but after having read that chapter, I really cannot see how it can be true. Also, be nice, since I do not have a lot of knowledge in this field, and sorry if the question is too long (although I do not think i digressed). :D
    – JNat
    Commented Jan 28, 2013 at 16:28
  • 2
    Good question. I myself never understood how compatibilism can be thought to really solve the problem. Just a note: if we consider the view of quantum mechanics, reality isn't deterministic. Of course, being probabilistic does not solve the free will problem, since we might as well simply be bound by randomness. But it turns the issue to a slightly different direction.
    – Koeng
    Commented Jan 29, 2013 at 14:45
  • 1
    @JNat Kinda, it may or may not eliminate the red dot. The first quote is correct in some sense, but we could say that the bonds of "fate and foil" between atoms are probabilistic, i.e., given a point in time, the only thing you can have is the probability of each succession. You don't have a fixed outcome like you do in determinism. That probability, as far as we know, is ingrained in reality, not simply a matter of "lack of knowledge" about the chain of cause and effect.
    – Koeng
    Commented Jan 29, 2013 at 22:40
  • 2
    @JNat I certainly did not want to claim a proof in either direction here. BTW and as you perhaps know, some think that free will is perfectly compatible with determinism -- a lot (too much for my current taste) seems to depend on semantics.
    – Drux
    Commented Feb 1, 2013 at 15:37
  • 1
    Observation: you give one definition of free will. Is it even a satisfying one? Putting your free will into a mysterious black box of choice seems to only multiply the questions. What determines what comes out of the box? (Naively), if there's an answer other than "I just do", then we only end up with parallel determinism. If there isn't, or you keep it magical, then there's something uncomfortably ad hoc going on. For what it's worth, in the modern debates I've read (Derk Pereboom, Carolina Sartorio, Carl Ginet, Harry Frankfurt, others), free will is defined via moral responsibility.
    – commando
    Commented Feb 16, 2017 at 5:00

7 Answers 7


Note: Whether this is actually an answer as opposed to an extended comment is debatable. I write it not as an answer (because I don't actually believe your question has an answer), but as "the closest one can get" to an answer. At any rate, I've turned it into a community wiki so I don't get reputation for it.

The crossroad you have come to is common, and you should be proud you picked up on it: It's a telling sign that you actually understand determinism; a concept which — even after many students of philosophy graduate with their degrees — I can assure you many never fully grasp.

You should note that there are dozens of different versions of compatibilism, each sometimes as radically different from each other as they are from hard determinism or libertarianism. I reckon some of the early versions were designed as a sort of "easing theory", published in philosophical journals as a "middle ground" to "ease" the academic community into this concept of determinism without seeming too radical. At any rate I have yet to read a convincing compatibilist argument myself, usually they just convolute terms like "freedom" and "determined" by redefining them in slightly different ways that supposedly don't conflict with (hard) determinism. Why do they do this? Because humans are not good at giving up things they are used to, or that they cherish. One of these things is the notion of moral responsibility, and determinism destroys this notion. Determinism also destroys the notion of praise (how can you praise someone for an action they took that was entirely inevitable, that wasn't of their own choosing?). This, in my opinion, is the real source of compatibilism: it is merely an unwillingness to accept that we may very well be automata, peering out the window of a robotic body, so to speak.

I suggest you take a look at the SEP article on compatibilism and see if any arguments are convincing to you, and if not, no fear — simply cast them aside. I also recommend (one of my personal favorites) Daniel Dennet's Elbow Room. He claims to be a compatibilist himself, but his writing all but screams hard determinism. His work helped me formulate my own position, which I write briefly about in my answer to this question.

Lastly, here's another question you might find interesting regarding choice.

  • 1
    Human's aren't just "not good at giving up", they are predetermined not to give up.
    – artm
    Commented Feb 3, 2013 at 9:11
  • 1
    The sentence was about "giving up things" which is a different concept than that which is evoked from the phraseology "giving up".
    – stoicfury
    Commented Feb 3, 2013 at 10:05
  • I do think that people are not good at giving up things they are used to might be the source of compatibilism, but not because they aren't willing to give up praise and moral responsibility. If that were the case libertarianism would be more suitable. It's because they aren't willing to give up... determinism. Perhaps they feel that the alternative is chaos, but it is a rare case of a notion getting very wide spread with literally zero evidence to support it, now that classical mechanics is no longer our fundamental theory.
    – Conifold
    Commented Apr 28, 2016 at 2:03
  • Good point about it being determinism we are reluctant to abandon, and one I hadn't thought of. For a well-developed version of compatibilism there is always the Perennial philosophy.
    – user20253
    Commented Dec 12, 2017 at 13:33
  • Compatibilism makes the most sense. The idea that being an automaton means one is not free, is an error of human intuition. "Free will" is incorrectly defined to require the ability to do actions which are neither caused by reasons, nor random. First, the idea is paradoxical in the first place so we're already as free as any creature could ever be in any hypothetical universe. Second: Even if we were to somehow have it, how would this magical paradoxical ability imply we're any more free/willful than we already are?
    – pete
    Commented Dec 12, 2017 at 19:24

My view is that libertarian free will exists, by virtue of dualism (I believe this is a prerequisite). However, I could be convinced otherwise.

The reason I could be convinced otherwise is the same reason I believe this question has not received a highly recommended answer in over 2 years - there is no way to prove any of the three major theories:

  1. For determinism, based on Wolpert's theory of maximum knowable information it appears impossible for us to prove determinism. Now, this doesn't shut the door completely on getting awfully close. For example, in the future we may be able to sufficiently simulate a real brain enough to answer this question generally to the satisfaction of most people by predicting outcomes, but somehow I think between Godel's limit of proving a system, Heisenberg's uncertainty principle, and Turing demonstration of the Halting problem is a way to show that determinism can not be proven.

  2. Since compatibilism rests fully on determinism (it is compatible with it, after all), that reasoning extends to compatibilism. I am finding compatibilism really to not be about free will anyway - as others have commented, it's really just a redefinition of terms to say that the free will is not really free - it is limited to certain choices. It then goes on to essentially say that we deterministically choose from among those limited choices. This is not a free will at all, from a perspective of being able to choose differently then I ultimately did.

  3. The only other alternative is libertarian free will, and this too cannot be proven. As it really requires dualism (or at least some separate level of existence or dimensions of reality, at least as I see it), it lies in a realm we have not figured out any way to test empirically to date. By the preceding I mean to imply that free will requires an actor outside of determinism - which would mean separate from the physical brain (although I do leave a small bit of room for the brain to create indeterminism on it's own at some quantum level or due to some physical law or dimension which exists that we don't understand yet).

Applying determinism fully to a mind is a leap of faith made from our observation of apparent determinism within most physical processes and phenomena (although there are increasingly now counter-examples of those, such as quantum fluxuations). Following from the above logic, this cannot be proven. To go any further, I believe one needs to go beyond empiricism into the realm of meta-physics and/or theology, considering the problem of evil/suffering, body vs. soul, and free will vs. either determinism or providence (different sides of the same coin).

I think this lands squarely in some form of dualism, at least as a means of further studying the implications of free will. For example - Augustine does not try to reconcile free will and grace/providence. He simply shows evidence that both must exist, without going into the how. This is an example from a Christian theological/apologetic perspective.

  • If there is no 'free will' how can one believe in any particular thing?
    – 201044
    Commented Feb 15, 2016 at 5:06
  • @201044 If there is not free will, then one apparently believes things because one was determined to. At least I've not heard any other options proposed...
    – LightCC
    Commented Feb 27, 2017 at 8:40

Compatibilism maintains (roughly):

  1. If you could choose differently, then you would act differently (thus keeping some notion of "free will")
  2. You cannot, however, choose differently (thus keeping with deterministic physics)

Therefore, we can view free will as being "compatible" with "hard" determinism.

To use the same quote of Blackburn's, but highlighting it differently:

This means that she would have done otherwise if she had chosen differently

So your question:

I too am under the impression that I could have chosen differently. But how can I prove that I really could have chosen differently?

I think is not an objection to compatibilism, because the "if...then" statement remains perfectly true even if the antecedent is never met. Indeed, I think most people who call themselves compatibilists would argue that you cannot, in fact, have chosen differently, unless the circumstances were different.

  • 1
    how about: (1) even if you could choose differently, still you wouldn't; (2) you cannot, so the second part of (1) doesn't matter.
    – artm
    Commented Feb 3, 2013 at 20:14
  • @artm Riding the waves of causality while feeling unrestricted is already the ultimate freedom. Whenever we try to imagine anything more "free" or "willful" than what already exists, it literally becomes a logical paradox. So how can we in good faith say we are not already free and willful? Another idea: A sane murderer defends himself by saying he was fated to do them. A driver with manslaughter defends himself by saying he had no time to react. Clearly we hold the former more accountable than the latter. Why? Because he freely made the choices himself. Doesn't matter that it was fated.
    – pete
    Commented Dec 22, 2017 at 17:42
  • 1
    The conclusion (murderer is responsible) does not prove the premise (murderer is a free agent).
    – artm
    Commented Dec 22, 2017 at 18:59

JNat, you've given an excellent summary of the leading incompatibilist argument. I want to highlight a strand of compatibilism that in my view answers your argument. Or I should say, one aspect of that line of thought, the one that best answers the motivating thought of your argument. Carl Hoefer, Barry Loewer, and Jenann Ismael are some important philosophers if you want to explore this further.

The key point is that your physics is wrong, as embodied in the line you drew from past to future. Determinism, as it appears in physical laws such as Schrödinger's Equation and the Einstein field equations, is time-symmetrical determinism. That is, you can start anywhere in history you like and draw inferences about the past (relative to that time) just as easily as the future. At the fundamental physical level of quantum fields and gravitational curvature, the arrow should be double-headed. The past is not in control. Rather, the past, present, and future are all in a pattern together, as far as fundamental physics goes.

But why do we intuitively think of the past as fixed, and the future as open? Because for macroscopic objects with trillions of moving parts, entropy is always increasing, and this leads to wildly different macroscopic evolution of systems into the future vs. the past. There is a well-defined, unidirectional arrow of time at the macroscopic level, where we live. The best definition of causality (https://www.youtube.com/watch?v=3AMCcYnAsdQ) applies only to macroscopic objects and properties described in a coarse-grained way (i.e., ignoring microscopic details). Determinism is not synonymous with universal causality. Causality doesn't go all the way down. At least, not unless you define "causality" in a way that allows that two events might reciprocally cause each other (which won't affect my argument, anyway).

Thus, we need two lines: Determinism at macro and micro scales

The macroscopic line is erratic because when we coarse-grain our descriptions (e.g., "the baseball broke the window") there simply isn't enough detail in those propositions to determine the future course. The macroscopic world, from fluid flows to brain activity, has enough chaotic dynamics that microscopic details ramify into macroscopic differences.

The microscopic/fundamental line is straight because at this level of description (note, the diagram is not limited to microscopic objects) there is enough detail to determine the future and the past. Intuitively, we think of the past as fixed and the future as open. When we learn the scientific facts about determinism, we see that this is wrong, because the past is not fixed. It depends on the present, including what we do now. The features in the past that depend on what we do now are mostly microscopic, and entirely unpredictable in practice. That's why we are blissfully unaware of them. But our best deterministic theories imply that they must be there.

Edit: if you prefer branching diagrams to represent indeterminacy, then the first diagram, for macroscopic descriptions, should branch out in both the past-going and future-going directions.

  • Thank you for your thoughts. There had already been a bit of discussion about how quantum mechanics would play into this in the comments here. As I understand it, your answer talks more about how I'm representing the flow of time, and whether or not events are fixed and caused by the past unidirectionally. Can you elaborate on how that proves there is an actual moment where a decision can be proved to have been made? Your graphs conspicuously omit the red dot ;)
    – JNat
    Commented Apr 17 at 11:45
  • In "empirical analysis" link we don't prove propositions involving subtle or sophisticated concepts. Instead we lay out the observed patterns and then reflect on whether our pre-theoretical concepts can be useful in communicating about these patterns. As for the red dot, I doubt there is anything ontologically special about the present. Since I wanted my graphs to show both ontic features like laws of physics, and epistemic ones like inferable info, I prefer to omit it.
    – Paul Torek
    Commented Apr 18 at 16:44

I've just released a discussion paper that attempts to resolve the outstanding issues between compatibilists and incompatibilists. In it, I argue that in common discourse, 'free will' is not taken to refer to 'contra-causal free will'. In fact, I question the hard determinist's and libertarian's contention that we intuit that we have such contra-causal free will. Using an ordinary-language analysis, I attempt to show that a 'free will' is an unencumbered will and that free will is restricted in four types of situations: coercion, manipulation, addiction and mental illness. Examining these situations, I distill four requirements that must be met for an act to be considered as resulting from a free will. These constitute my 4C theory of free will and are:

  1. absence of Compulsion;
  2. absence of Control by third party;
  3. consonant with agent's Character; and
  4. Cognitive capacity to reason.

I argue that, in fact, these four criteria underpin jurisprudence, forensic psychology and our ordinary moral intuitions and our practice of praise and blame. I also go on to provide a credible counterfactual conditional analysis of 'could have chosen otherwise' along the lines of 'given the agent's character, the agent would have chosen otherwise in the given situation if the circumstances were different'. I welcome your feedback. You can read the paper at Free Will and Compatibilism

  • I have not read the linked papers, but this seems essentially the same answer as the one that Daniel Dennett developed - very convincingly - in Elbow Room. What we care about in 'free will' is not the philosopher's abstract notion of arbitrarily free will, but control; I care about the fact that the choices are my choices. And they can only be my choices, in so far they are coarsely determined (consistent with my past or my character). There is no contradiction between being in control, making the deliberate choices I make (that is, realizing a particular choice), and being predictable.
    – mudskipper
    Commented Jun 28 at 20:42
  • People thinking about 'free will' tend to forget that being in a choice situation (where I actually need to make a choice) is almost always perplexing, painful and hard. In such a situation, I just don't know what to do, which course of action is right - I'm not able to predict my own actions. (But this doesn't need to imply that others cannot either.) An arbitrarily free will would be like the medieval pons asinorum - poised between equally attractive options, there is no reason to choose either and the only way to break the impasse is to toss a coin...
    – mudskipper
    Commented Jun 28 at 20:54
  • In other words, +1 on this answer.
    – mudskipper
    Commented Jun 28 at 20:57

Aristotle in his book, Metaphysics, stated that in his time all philosophers agreed that principles of things are contraries.

In this view, one might suppose that the free-will of man and the determinism of nature is a contrary that is a principle. That the essence of spirit - freedom - is opposed to the essence of matter - neccesity.

If one instead is a pragmatist, and so asserts both the determinism of physics and the freedom of the human will, then one can say that they are compatible but we simply do not understand nor describe precisely the mechanism of how this is the case. This sort of proof is reminiscent of an existence proof in mathematics where one asserts the existence of some mathematical object but it does not construct it, in some cases it is even shown that one cannot construct it.


I switched from incompatibilism to compatibilism a while ago. I think the driving factor was the realization that anyone who actually fully embraces incompatibilism and determinism should be okay with doing nothing or making really bad decisions, since they have zero control over anything. But they don't do that and continue to make an effort to make good decisions, which I believe is due to a part of their brains which accepts that they are still "in control" and can will things, in spite of determinism being true. According to incompatibilism nothing can really be "in control" of anything, and the words "in control" and "decision/choice" lose meaning which strikes me as bizarre. More compelling arguments below:

If we don't have free will, it implies we are constrained in some way or lacking some sort of ability. What is the thing we are missing, which if we had, we'd be free? There is no such thing. It's not even a superpower like teleportation or mind-reading which could hypothetically be true with enough technology/magic. It is literally a paradox. It seems pointless to lament something that is missing, if that missing thing could never have existed anyway. We are already as free and willful as any creature in any hypothetical universe could possibly be. Riding the waves of causality without feeling restricted is already the ultimate freedom.

Another approach: Incompatibilism basically says "free will" involves some sort of "magical decision" which is both impossible to predict, and not random. Both compatibilists and incompatibilists agree that a magical decision which is impossible to predict and non-random is a paradox and can't exist. But the compatibilists ask, "what exactly is it about being impossible to predict that makes someone more free or willful than they would've anyway?" The mind-reading machine will predict which button I will press every single time, but I still pressed them based on my wants (will); I wasn't forced to press them.

  • An incompatibilist would argue that you only pressed the button because each event directly influences the next one, since the dawn of time — that's the extent to which you were "forced" to press it.
    – JNat
    Commented Dec 12, 2017 at 8:23
  • I'm aware of that argument, and I agree with the first half of the statement, but to say it is "forced" would be a stretch in my opinion. Under this logic all actions would be "forced". The word "forced" is only meaningful if it is possible for something to be "not forced". We recognize its meaning because it is in contrast to behaving freely. If there is literally no way for any creature in any universe to ever behave in a contrary manner how can we say it is "forced"?
    – pete
    Commented Dec 12, 2017 at 18:22
  • This is a tangential discussion, it seems to me, that delves into the semantics field. Furthermore, you were the one who introduced the word "forced," presumably because I mention "lack of will." Where I'm trying to get is that there is no way you could have done something different from what you did, even if it appears as if you could. The predictability, or lack thereof, of the actions being examined is also not relevant — I'm trying to say a completely predictable action and a completely unpredictable one are both completely devoid of "will," despite of, again, what it seems like.
    – JNat
    Commented Dec 12, 2017 at 18:33
  • IMO it goes more than just semantics; it's about which criteria need to be satisfied in order for us to be "truly free". If those criteria are by definition paradoxical, then the position that free will does not exist is tautological/meaningless. When you say the actions are devoid of "will", can you define "will"?
    – pete
    Commented Dec 12, 2017 at 19:13
  • Re: defining "will:" see the original question, as that's pretty much explained there. As for the rest: you're not answering my real question ("How can our notion of free choice not be restrained by our physical world, and thus bound by the laws of causation? How can free will and free choice exist in such a world?") and are, instead, going around the issue. The arguments about social responsibility (which the book goes to great lengths to explain and make the focal point of the existence of compatibilism) are also beside the point, in my view.
    – JNat
    Commented Dec 12, 2017 at 19:21

You must log in to answer this question.

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