I've been thinking about a few legal quotes that have initiated my investigation into whether or not choice actually exists:

A "universal and persistent" foundation stone in our system of law, and particularly in our approach to punishment, sentencing, and incarceration, is the "belief in freedom of the human will and a consequent ability and duty of the normal individual to choose between good and evil." - United States v. Grayson, 438 U.S. 41 (1978)

Our concept of blameworthiness rests on assumptions that are older than the Republic: "man is naturally endowed with these two great faculties, understanding and liberty of will." "[H]istorically, our substantive criminal law is based on a theory of punishing the viscious will. It postulates a free agent confronted with a choice between doing right and wrong, and choosing freely to do wrong." Central, therefore, to a verdict of guilty is the concept of responsibility. - United States v. Lyons, 739 F.2d 994 (5th Cir. 1984)

It seems to me from my reading of "The Secret Politics of the Compatibilist Criminal Law" (2007) by Anders Kaye that the current American legal system derives its notion of "choice" under the paradigm of the compatibilist framework.

Despite reading that compatibilism is involved, I am skeptical of that the American legal system relies on compatibilism being true. What does seem legitimate is that there is a reliance on concepts, such as the "will" and "choice."

From what I've read from various sources, my interpretation of what the U.S. Supreme Court was arguing in relation to "will" and "choice" is that if (1) I have a desire and (2) I act on that desire by way of causing (not via libertarian free will) from within myself (by way of using my "will") a volition to seek out satisfaction of that desire, then I have chosen to satisfy that desire.

  1. An agent has a desire
  2. The agent acts upon the desire by way of causing its will into causing voluntarily movement in order to satisfy the desire.
  3. The desire either either becomes satisfied or not satisfied.

This appears to be an argument for agent-causation occurring while working under the philosophy of compatibilism. I'm not sure if that's the proper interpretation of what the legal system is arguing as the philosophical framework for "choice," but it appears that such is the case. I am strongly under the impression that the Supreme Court was implying that an agent causes the will to bring about volition via agent-causation. That's not explicitly argued anywhere, but I can't perceive how it is being argued "access" of the will comes about in order to cause a volition. Maybe its being implicitly argued by the Supreme Court that realization of a desire through it completely neurobiologically manifesting (if at all) causes the will to cause a volition.

My reading of "AGENT CAUSATION AS THE SOLUTION TO ALL THE COMPATIBILIST’S PROBLEMS" by Ned Markosian makes me seriously question how to interpret these matters. It appears that for a person to be held criminally liable that a "type" of agent causality under compatibilist thought must be occurring, namely agent-causation without external influence.

The first thing that is obviously suspect to this notion of having "freely" (without external constraint, as per compatibilist thought) made a choice is that desires exist. I'm not sure if desires exist, and eliminative materialism may be applicable to this situation.

The second thing that is suspect to this notion of having "freely" made a choice is that agent-causation occurs is true (my presumption is that agent causation is necessary for a will to be used). Agent causality appears dependent on an immaterial mind with the ability to cause the volition to satisfy the desire (such as a desire to commit good or evil, moral relativism aside). Whether or not such is actually physical, it is presumed that it is a will (not necessarily above the constraints of the deterministic universe) that brings about causation for a volition to undergo.

It appears to me that to think that choice exists (as per the compatibilist paradigm), then I need to concede that (1) desires can exist and (2) agent causation via a will can occur.

There is not a lot in I've found through Google that covers this matter. I've looked through Google's search engine to resolve my question, but I've not found anything that sufficiently ties the notions of agent causality, compatibilism, voluntariness, and choice together.

As far as I got on the matter was a 2022 article that follows, for which it appears to concede that agent-causation does not exist yet appears to defend that the notion of it be kept around on a definition of what agent-causation is interpreted to be:

The idea of agent causation—that a system such as a living organism can be a cause of things in the world—is often seen as mysterious and deemed to be at odds with the physicalist thesis that is now commonly embraced in science and philosophy. - "Naturalising Agent Causation" (2022) Henry D. Potter and Kevin J. Mitchell

I realize that the way my post's title has been worded this kind of brings this matter into the realm of science, but I think the question of my post can still be constrained to what kind of things I would have to concede to as existing in order to concede to "choice" existing, as the U.S. Supreme Court uses the concept of choice in the quoted material.

For example (to move away from agent causation, as it seems from responses so far that it is not implied to be involved), it appears that if I (1) concede that desires exist, (2) concede that the will exists, and (3) concede that the complete manifestation of a desire in a person can cause the will to cause a volition to satisfy that desire if no external constraints are present, then I should concede that choice (as the term is used by the Supreme Court) exists.

  • Googling up "compatibilism" might help, me guess. In this world view, choices do exist. (I personnally think compatibilism is a good world view)
    – Olivier5
    Aug 21, 2023 at 21:10
  • I've done a fair amount of studying on compatibilism, but I have yet to find much information on anyone attacking the notion of "choice" that exists as part of compatibilism. Aug 21, 2023 at 21:11
  • I cannot offer an attack on compatibilism, only a defense of it.
    – Olivier5
    Aug 21, 2023 at 22:09
  • 3
    Compatibilists do not attack the notion of choice because they redefine it. "Free choice" is still fully determined, but through internal mechanisms of agency rather than forces or coercion external to it (no "external constraint"). While causal chains run through the agency, agent causes are regular physical causes (brain function, etc.), not "agent causation". The causes "from within" are only proximate causes, ultimate causes, including those that shape the agency itself, are external, see SEP, Compatibilism.
    – Conifold
    Aug 21, 2023 at 23:34
  • 1
    Unless you start subdividing "people" (e.g. "being of sound mind" vs not) so that you have agency even if those people don't, the only winning move is to act like people have agency, regardless of the truth. If people have agency and you act like it, you win. If people have agency and you don't act like it, you lose. If people don't have agency, there was never a game.
    – g s
    Aug 21, 2023 at 23:57

4 Answers 4


Most legal systems rest on the view, implicitly or explicitly, that humans are largely responsible for their actions. You do not need to invoke a particular definition or test of free will to justify the view. Even if we are all programmed to do what we do, and choice is an illusion, you might take the view- for a number of reasons- that a person who has been programmed to commit a crime should still be punished for it. Indeed, if you take the point to the extreme, you might take the view that if we are all programmed to do what we do, then the decisions taken by judges, juries and law-making assemblies are all programmed too, so the legal systems we have are there by programming rather than by choice.

Where there does seem to me to be a degree of inconsistency in the legal system is in relation to coercion. If you are forced at gunpoint to commit a crime, you will, in most legal systems, be considered not to deserve punishment, and it is hard to draw a clear distinction between such acute and blatant coercion and the more subtle forms to which we are all subjected over the years by our interactions with society at large.

  • 1
    Some people refuse to commit crimes, even if coercion is applied. It opens up the question a bit. Of course, maybe they were just programmed to do that.
    – Scott Rowe
    Oct 28, 2023 at 0:17
  • 1
    @ScottRowe that is an excellent point. Oct 28, 2023 at 5:16

That's a long OP. I'll try to limit my answer to this:

the question of my post can still be constrained to what kind of things I would have to concede to as existing in order to concede to "choice" existing

(emph. added)

My proposition is that you only need to concede to the causal capacity of structures. This is equivalent to saying: structures matter, they have efficacy.

Note that this idea is quite natural. We all know that a car, for instance, is in a very important way different from a pile of spare parts. If one deconstructs a car, dismantles it into its constitutive elements and makes a haphazard pile of these elements (the parts, arranged in a rough pile), one does not expect the car to work anymore. The structure, the way the parts are assembled together into a very particular shape, is important for the good functioning of the car.

So it's natural to think that structures matter. But it does go against another belief, generally called reductionism, which is that only elements matter (to caricature a bit). In reductionism, the ontology of structures is uncertain, as the reductionist sees what happens at the level of any structure (eg a molecule) as dependent and derived from what happens at lower levels (eg the atoms composing the molecule).

In reductionism, elements matter, but structures don't really matter, as they are not directly causal. Any real causality must come from below.

There's plenty of problems with reductionism: 1) it may not be doable, as in practice nobody ever managed to reduce chemistry to physics, or biology to chemistry, etc. 2) it seems to assume the existence of a fondamental level in this reality of ours, and we don't know for sure that reality has such a fundamental level - for all we know, reality could be structures all the way down. 3) It moves causality at a level very very far away from anything like us, living organisms, below quark level -- it's disempowering. 4) It neglects the fact that new structures can do new things, and thus create their new "space" of opportunities and constraints - such as predation for instance.

[The emergence of life led to the emergence of life eating life (predation), which comes with its own risks and rewards. Predation makes very little sense when envisaged at the atomic level. It emerged with life, so it only makes sense at a biological and ecosystem level. If a tiger eats you up, your atoms won't care.]

Now, IF one abandons reductionism as an ontology or metaphysics (one can still use it as a research principle; it's always a good idea to check on them elements as long as you don't neglect the structures), then one goes back to our proposition: structures matter.

This non-reductionist proposition has consequences and advantages: in particular, that now all you need to do to conceive of agent causation, is to see agents as structures.

Aren't we somewhat structurally different from any pile of quarks?

What would explain the fact that our brain has such a complicated structure, if not because it's doing something important, causal, that requires a certain structure to get going: thoughts.

  • 1
    +1 That's a good metaphysical analysis bringing into the lights the metaphysical presuppositions inherent in any discussion such as deciding which property attribution is metaphysical necessary, how bottom-up causation is generally intuitively built into our model of the physical universe, and other merelogical concerns. Thoughts are certainly attributable to minds that have structure. No brains, no minds!
    – J D
    Sep 21, 2023 at 19:42
  • 2
    @JD "No brains, no minds!" Indeed. Vice versa, if the minds produced by brains provided no survival or reproductive advantage to organisms, evolution would not have bothered with brains. In this Darwinian perspective, the function creates the organ. So no minds, no brains.
    – Olivier5
    Sep 21, 2023 at 20:39
  • IMNSHO that's a category mistake. The mental and the physical are distinct domains of discourse wo no causal interaction that is not metaphysical. The mind isn't a biological ontological primitive. You merely can substitute the term neurological computation to denote the property dualistic standin terminology for a more cogent model that occupies a conceptualist middle ground between a realist and instrumentalist position. I think today. Maybe something else tomorrow.
    – J D
    Sep 22, 2023 at 0:22
  • 2
    I don't think so. A piano is a different kind of thing than music, yet one can make the latter with the former. And if all pianos were mute, nobody would build them. So pianos make music, and music justifies and motivates pianos. If no piano --> no music, and (in a less direct way) if no music --> no piano. Bottom up and top down causation at work.
    – Olivier5
    Sep 22, 2023 at 6:11
  • 1
    @JD William James spelled out the evolutionary test case that shows that epiphenomenalism is false -- we have demonstrated variance in thought/action correspondence, yet thought and action remain closely correlated. This is only possible if our thoughts have been evolutionarily selected for correspondence to actions, and selection only applies to actual causal structures. Hence, thoughts are causal. Popper extended this argument against all Identity Theories, not just epiphenomenalism.
    – Dcleve
    Oct 27, 2023 at 19:30

Harry Frankfurt's mesh model, and the thought experiment about someone whose circumstances are preemptively manipulated to guarantee an outcome even if the manipulated agent chooses, from within, to yield that outcome, are taken to be insightful enough (if not, of course, entirely conclusive) to undermine the "choose your own adventure" (same antecedent situation, multiple directly possible outcomes) picture of free will. So any defense of the CYOA-picture comes out as a possible attack on compatibilist reasoning about the concept of choices.

Now, taking Conifold's comment as a point of departure: one way to try to model free action is in terms of attributing a good enough amount of dynamical structure to the will, a parsing of its functions that allows us to make the kind of fine-grained distinctions between outputs of those functions to usefully differentiate a notion of free from unfree action. If we wish to retain the language of choice, here, it is enough that we find nodes in the structure that are sufficiently oriented towards multiple outputs as to evoke the picture of choosing-between-alternatives. The technical specification of this picture, then, is that the concept of choice is understood in terms of the functions having ranges of outputs in the domain/range sense; these are alternative possibilities at least in the way of being generally possible states.

The domain problem: it remains for the incompatibilist concept of the-ability-to-do-otherwise to insist on asking about particular alternative possibilities. Agent-causation as a domain principle can be taken for the claim that inputs given beforehand as constitutive of agents, are the ones that are most pertinent. The machine of the will winds itself up and sets itself in motion (c.f. Diderot's "clock that winds itself"). Some effects of the will's internal processing count as free, others as unfree; the distinction is nontrivial in the theory; however, any given (physical) will appears as itself an element of a range of temporally external functions (ambient as well as occasioning causes of some will's appearance in the world).

Immanuel Kant maintained that since empirical causality strictly holds for our in-world knowledge, we can accurately think that every person's empirical actions can be explained in general, though not necessarily in particular, by tracing those actions back through the sequence of objective determination in time. He simultaneously maintained that at the moment of moral action, it is possible that one will obey the moral law and it is possible that one will not obey, regardless of the whole preceding sequence. Since he brings up the juridical case of evaluating criminal guilt as his vivid example of applying both empirical and moral laws, then his nonstandard compatibilism (or meta-compatibilism, perhaps) seems to be one with agent-causation nonetheless.

The upshot: modern/standard, empirical, compatibilism suffers from a little theoretical weakness in setting aside the question of particular alternative possibilities, but Kant's transcendental compatibilism does not suffer from that weakness (the arguable extravagance of his metaphysics of morals notwithstanding). But I think that we might be able to refine various "models" of compatibilism in light of further reflection on the complexities of e.g. the logic of action or temporal logic. And ultimately, the meaning and epistemology of modality have to come to the fore: the compatibilist doesn't quite have to say, "I know that particular alternative possibilities aren't directly real," their position is compatible(!) with agnosticism about the metaphysics (and is thusly emphasized for the sake of the argument from morally reactive attitudes), but it would be a hit for the position to find out that direct and particular choices are possible on account of some elaborate fact about how possible-worlds relations combine and intersect.

It's worth recalling that John Rawls tried to find one of the internal justifiers for his theory of justice, in Kantian ethics. He was, at the same time, in tune with the moral culture of the United States. Both the SCOTUS decisions you quote at the outset are from later than 1971, the year of publication for A Theory of Justice (in its first edition). 1978 particularly was still in the moral shadow of the Vietnam War and the development of American international legal consciousness attendant upon that war. It had just been in the 1960s when the mathematics community, starting with Paul Cohen, discovered the pluralist themes of set-theoretic forcing, and so the metaphysical pluralism of strong alternative possibilities was a fire that would have been fanned around that time for another reason. How responsive to these cultural currents the SCOTUS is, I don't know, though I do recognize that, from the outside, they seem more amenable to public opinion than they supposedly "should be," perhaps.

A corollary is that law students often cross-pollinate their degrees with philosophy courses, so the philosophical frontrunner (for the time being) in the conceptual analysis of free will, which frontrunner is in fact, at this time, some species of compatibilism, will spawn its children in the thoughts of future lawyers and judges and the like, including (or especially) in the US.

Consider Kant's distinction between Wille and Willkur in the architecture of transcendental freedom. This is more specific, and thence intricate, a matter than having a will that exercises itself merely like the flipping of a lightswitch (at best). Take, for example, the use of the will in aiming for either general or future effects: there is something integrated about the will-to-build-a-building, say, such that when we finish building, we have finished an action whose purposefulness was projected from the past, onto the future. And so roughly as Kant says in the Groundwork, practical reason "deduces" actions from maxims, or specific actions from the Form of Action (if you will...), so we have good reasons to think that the will can mediate both general and particular representations (in either direction, no less).


Re: the (so far) final phrasing of the OP question:

... [should I] concede that choice (as the term is used by the Supreme Court) exists.

If you mean, "Should I concede that the Supreme Court is correctly using the term 'choice' or not?" then if you distinguish (A) agreeing with specific SCOTUS decisions from (B) believing more (if not much more) in the Supreme Court as an institution, you can go on to ask if the Court has been, in all its rulings, consistent in its use of the terminology of choice. Moreover, though, if you're not of a mind to want to use words like "choice" so much (while you reserve the right to use words like "desire" and "will"), just note that your antecedents don't mention the term "choice." By contrast, the Supreme Court decisions as quoted do use such terms explicitly, so you can then either (1) take them to be stipulatively defining their terminology on the basis of antecedents (or precedents, as they say in legal reasoning) which you do accept, which stipulative definition does not carry with it a demand that you, as an outsider to the argument in Court, need apply in your own writing; or (2) you can take the Court to be adding a more robust notion of choice to a definition you are otherwise inclined to accept, in which case you can disagree with the premises of the Court just in case you feel the need to avoid using choice terminology over and above the other terms of your compatibilism.

If you mean, "Should I concede that the concept of choice has a solid place in compatibilism?" then that depends locally on your definition of compatibilism modulo the antecedents in your if-question, or globally on the analysis of the varieties of compatibilism.

  • I don't feel this answered my question. The Supreme Court must have been using a paradigm of choice to argue that blameworthiness was a possibility. What paradigm that was seems indecipherable. Aug 22, 2023 at 0:47
  • @DennisFrancisBlewett Kant's influence on theoretical and practical reasoning has been fairly vast, e.g. concepts like the League of Nations and the UN find much purchase in his moral/political discourse. I would venture that he's had a more direct impact on European than US legal theory, but I've heard that SCOTUS decisions can involve hundreds of pages of material, so I would be surprised if no SCJ was at least peripherally influenced by Kantian compatibilism at some time or another. Aug 22, 2023 at 0:52
  • 1
    Very informative! I'm going to stick with The Yoga Sutras of Patanjali 1:33 and say that if someone says they had no choice in doing harm I will say I have no choice but to confine them, and if they act virtuously by choice, I will choose to appreciate them. The stupid argument that we lack free will works both ways, and I will use it both ways.
    – Scott Rowe
    Oct 28, 2023 at 12:06

Does choice exist? Yes, we are constantly faced with alternative possible courses of action of which we must choose one to be implemented.

Can we choose freely? Yes, choices are always made alone, for the individual's own reasons and purposes. There is no-one else messing with our choices.

Can we choose without external influences? No, external influences are knowledge about external reality. We cannot make any choices with our preferences alone, without any knowledge about reality.

Do desires exist? Yes, desires are knowledge about a mismatch between reality and our preferences. Something feels not right, we need to do something about it. In other words, desires are problems that need a solution. Our choices are the solutions by which we try to satisfy the desires.

  • I chose to upvote.
    – Scott Rowe
    Oct 28, 2023 at 0:18

You must log in to answer this question.

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