What else then can freedom of the will be but autonomy, that is, the property of the will to be a law to itself?

Groundwork of the Metaphysics of Morals

So that in a nutshell is autonomy and freedom of the will for Kant.

But the problem is: there's just one way how any correct moral law must turn out - exactly as conceived by Kant, and this in all its details!

Kant had elaborate opinions about many moral issues, like:

  • lying in various circumstances
  • suicide
  • abortion
  • killing in war
  • wasting our talents
  • abandoning children born out of wedlock
  • homosexuality
  • masturbation
  • etc.

Except, of course, that he didn't regard them as his mere opinions and instead as unshakable, ironclad deductions from the categorical imperative.

Every rational being supposedly assents to the categorical imperative and, rationally, also comes to the same conclusion as Kant in all these matters. There's no leeway.

Also, if we don't follow the moral law, Kant doesn't regard this as "free will badly applied". Ironically, the will is unfree then, since our emotions, good or bad, are unfree per se, and they just overpowered our reason:

Now, humanity can be located either in the capacity and the will to share in others’ feelings (humanitas practica) or merely in the receptivity, given by nature itself, to the feelings of joy and sadness in common with others (humanitas aesthetica). The first is free, and is therefore called sympathetic (communio sentiendi liberalis); it is based on practical reason. The second is unfree (communio sentiendi illiberalis, servilis); it can be called communicable (since it is like receptivity to warmth or contagious diseases), and also compassion, since it spreads naturally among human beings living near one another. There is obligation only to the first.

Metaphysics of Morals

Since any deviation from the "Kant-program"(.exe) is irrational and therefore unfree, I just wonder where Kant got his aversion to physical determination.

As long the "Kant-program" flawlessly executed means total freedom realized, one wonders what difference it makes.

OK, let's grant that there's an important philosophical difference between the rational insight that 5 + 7 = 12 and doing the same calculation on a computer.

But why should we call the former autonomous, the latter heteronomous?

Why would we even want freedom in the Kantian sense?

What real freedom is left in his philosophy? I don't see it in the autonomy (the law-giving) and neither in the decision-making.

  • 4
    In the Doctrine of Virtue, and with respect to the notion of imperfect duties already in the Groundwork, Kant says that the moral law is much more "open" than his basic rigorism would seem disposed towards. It might still be that every local practical problem has an exact moral solution, ultimately, or it might be just background problems that have strictly limited solutions, or some combination of both states. And free will, he says, is like the metaphysical substance of ethics (the phrase he uses is ratio essendi), the moral law exists as the law of the pure will. Commented Sep 12, 2023 at 22:45
  • 1
    You might try out a rough analogy with ZFC: as the kernel of "traditional" theories with large cardinals, in all their dazzling intricacy, ZFC is like the foundation of a law, but it is consistent with other extensions of the law, other reformulations of it. Kant does say somewhere (if I might paraphrase) that even the multifaceted formulation of the categorical imperative is a set of shadows cast by the united light of the moral law in itself, so that when we obey the shadows, room is left for us to obey the others, and there is nothing overly strict about such obedience either way. Commented Sep 12, 2023 at 22:50
  • 3
    Because moral laws are not like physical laws. They do not prescribe a single course of action in every situation but rather act as constraints on a variety of possible actions, and whether to respect the constraints is still a free choice, they are not forced by physical compulsion. Deterministic causation completely rules out the only outcome that matters morally - choosing to follow the moral duty. As for not being "free" regarding the laws of reason, "freedom" from reason is no freedom at all. It is only reason that connects intentions and actions to make the freedom meaningful.
    – Conifold
    Commented Sep 12, 2023 at 23:10
  • 1
    Where did Kant claim we have or ought to have "freedom as autonomy" as you put it? Freedom is the power to make choices, not being in a situation where there are no standards about what is the proper choice. Commented Sep 12, 2023 at 23:22
  • 1
    @Conifold whether to respect the constraints is still a free choice for Kant, this is just passions overpowering reason. Not a free choice, imho.
    – viuser
    Commented Sep 12, 2023 at 23:23

3 Answers 3


Sometimes people read the phrase "a law unto oneself" and get existential self-creation out of it; for criticism of this reading, see e.g. Allen Wood, Kantian Ethics (Wood mainly targets a constructivist interpretation of Kant, but if I remember correctly, he also brings up an existentialism-themed (or -named) demi-mirror of constructivism, here). To be fair, latitude in understanding what the (Kantian) moral law is in a given case can become an art of autonomous judgment, but this is not how it always, or often, is meant to work: Kant does not claim, for example, that lying can magically be justified if we have "freely willed" that this be so (c.f. Rawls' commentary on Sidgwick's noumenal scoundrel in sec. 40 of A Theory of Justice).

We should also emphasize Kant's eventual distinction between Wille and Willkür, or (very, very roughly) free will and free choice. At about the start of the Doctrine of Virtue, for example, Kant says that we place ourselves under obligation and have duties-to-ourselves, and he distinguishes the intelligible self from the empirical self, saying that it is our "higher"/"truer" self that gives the law to the latter. The "higher self" is free per the form of the law of the will, the "lower" self has the freedom to choose whether to abide by that law. (This is rather how Rawls addresses Sigdwick's complaint, incidentally; Rawls speaks of "actions expressive of our true nature.")

Autonomy (in the Kantian sense), then, stands at a midpoint between an internal totalitarianism inspired by compulsory laws of physics on the one hand, and the sparkling chaos of random decision-making on the other. It might not really pan out all too well in the end (or at least it's an incomplete picture, with much room for improvement), but autonomy, here, at least doesn't seem necessarily more inchoate or incoherent than various other proposed foundations of moral laws as such.

  • As Timmermann shows in one of his (German) books, Kant has two kinds of Wille: Wille1 as the faculty of desire as such, which includes Willkür, and the other kind of Wille you describe also being part of the former, which forms the principles of willing. He is pretty explicit about only the Willkür (freedom of choice) with Wille2 together allowing for freedom. Since infinite rational beings have the principle-building Wille2 but lack Willkür, they cannot be considered having a free Wille1. Kant is consistent about that from Grundlegung up to Anthropologie.
    – Philip Klöcking
    Commented Sep 14, 2023 at 6:11
  • 1
    Sorry for being pedantic here but that was an integral part of the argument of my bachelor thesis hehe
    – Philip Klöcking
    Commented Sep 14, 2023 at 6:16
  • @PhilipKlöcking what would be a good way to edit my answer accordingly? Or, does the will/choice distinction help address the OP? I'm torn about the question, here: (A) I want to defend Kantian autonomy even when it is a rigid thing, but (B) I want to defend the non-rigid passages too. Commented Sep 14, 2023 at 16:19

Freedom from the laws of physics is important because otherwise, we'd have the "freedom of a turnspit" as he puts it in the second critique: we would have two (or more) possible "sides" (choices) but ultimately, it is external forces that decide which one is on the upside (which choice is taken). How is that freedom in a meaningful way?

Now, the freedom of will is exactly to decide between following one's inclinations (be a turnspit) and let your will be determined by things outside of your own willing or positing one's willing outside of the world and follow only the law the will gives itself (autonomy - governed by its own laws). Finite rational beings do not have to follow their rational side, they can still choose to follow their inclinations - that is what makes them free.

Ultimately, both paths are determined: one by the laws of physics, the other one by the law of freedom (the Categorical Imperative). The freedom is mainly a negative one: It is not to choose one's inclinations. But to positively determine one's will, there has to be a principle of willing if one chooses that path (given it exists in the first place). And that principle is the principle of the free will (and nothing else).

Kant tries to achieve both showing that this path exists and that the determination of a will free of inclinations can only happen categorically, hence the Categorical Imperative.

Now, why is it important that the outcome is willingly chosen via moral choice? Because morality is a matter of responsibility and owning one's actions - autonomy. If you program a robot to necessarily do the right thing, you'd not really revere it's moral behaviour, would you? You'd rather say its programmers did a really nice job there.

Edit after heavy editing of the question

You only have half the picture here. You describe positions he develops in the Groundwork, the Metaphysics of Morals, and a small essay here. Most of them (except, indeed, the lying to a murderer) are snippets of casuistic reasoning. Kant wrote in those works and elsewhere how much he despises casuistic reasoning, which tells us he used it only when he needed to to make a point different from the case itself ("to bring the Categorical Imperative closer to intuition").

Thus, we should rather consider his other writings, where he makes clear that the highest form of morality is a moral character, ie. being able to intuitively feel and know the right thing.

Another important aspect that is not apparent in his more formal writings on morals is that he always acknowledged the importance of having a variety of experience and how this shapes our ability to form practical rules. You know, ultimately, it is wrong to think the Categorical Imperative comes in a thousand shapes that fully determine in every case how moral behavior looks like. That's not how it works. Rather, if you lack moral character, you can take the subjective principle of your will (the maxim) and test whether it is moral. The moral law itself is purely negative and can only help finding out whether a given content is moral or not: since being noumenal, it cannot be positively determined.

  • The forces in a robot aren't external in a strict sense. Data is handled internally. Of course it only executes its program and we wouldn't revere its moral behavior, which was part of the analogy. But humans (according Kant) don't seem different. We are endowed by reason through some "creator". Reason will act according to its laws and compel us to a certain action. So it's just that the "creator" of man did a really nice job, at least in humans coming close to a perfectly rational ideal.
    – viuser
    Commented Sep 15, 2023 at 23:01
  • We all know that we cannot un-accept Pythagoras' theorem once we understand the proof. In this sense we have NO "freedom from reason", but this sense is unproblematic (even desirable, perhaps). But IMHO we HAVE freedom from reason regarding our actions. If we DO NOT, like Kant seems to claim, how is it not "freedom of a turnspit" again? It seems like a version of compatibilism...
    – viuser
    Commented Sep 15, 2023 at 23:03
  • @viuser As said, the freedom in Kant comes from a) being able to choose one's subjective principles of the will (Willkür choosing maxims), b) the possibility of them being in conflict with the moral law, and c) our ability to decide which maxim is to determine our will. You make is look as if the moral law described one concrete willing. That's not the case. It merely allows us to distinguish moral willing (maxims) from immoral ones. What is "freedom from reason" supposed to mean here, anyways? If not governed by rules it becomes arbitrary, anarchist.
    – Philip Klöcking
    Commented Sep 16, 2023 at 6:13

The question really brings out the crux of morality (as alluded to, autonomy). Juxtapose that with Kant and we come face to face with an either-or dilemma. Someone as astute as Kant would surely have been aware of this, his word, antinomy. Yet, from the OP's question, he seems not to have addressed it, even partly, let alone wholly. Perhaps, it is possible, may be, he felt it was too obvious to state what he was getting at. We would need to do something akin to a crime scene reconstruction ... of his reasoning (sensu amplissimo).

You must log in to answer this question.

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