by ANDY YU
In this paper, I discuss Kant’s main argument for free will from morality. The aim of this paper is to reconstruct his argument as found mainly in the Grounding for the Metaphysics of Morals and the Critique of Practical Reason. Concisely put, he argues that we can and even must admit free will in order for morality, which we intuitively accept, to be meaningful at all. As preliminaries to the main argument, I begin with a brief introduction to Kant’s metaphysics and epistemology, as well as his conception of free will. Following this, I explore his main argument for free will, which relies on the thesis that morality reciprocally implies free will. I break this argument into two steps. First, I discuss how Kant shows that morality implies rationality. Second, I discuss how Kant shows that rationality, in turn, implies free will. Before concluding, I review Kant’s position on the apparent incompatibility between free will and determinism.
I start with a brief introduction to Kant’s metaphysics and epistemology to establish the kind of knowledge about free will Kant thinks we can maintain. He details this-what we can know and how we can know it-in the Critique of Pure Reason. His metaphysics details what we can know by distinguishing between two worlds, the phenomenal world and the noumenal world. While the phenomenal world is the empirical world in which we experience objects as they appear to us, the noumenal world is the rational world in which we conceive of things-in-themselves. In other words, the phenomenal world is concerned with appearances, while the noumenal world is concerned with things as they actually are. Although Kant does not explicitly state how we are to conceive of these worlds, we can conceive of them as either two ontologically distinct worlds (two world interpretation) or two aspects of the same world (two aspect interpretation). As far as this discussion is concerned, I do not adopt one interpretation or another, as each has its merits. Notwithstanding this, an important consequence of Kant’s metaphysics is that whatever knowledge we can claim about the noumenal world is different not only in degree but in kind to whatever knowledge we can claim about the phenomenal world.
In turn, Kant’s epistemology explains how we can know what we know. For the present purpose, the most relevant way we can acquire knowledge is through reason, which can be either pure or practical. While pure reason is primarily concerned with theoretical or speculative claims, practical reason is primarily concerned with moral claims. However, pure reason and practical reason are not, strictly speaking, different kinds of reason, as they “are differentiated solely in their application” (391). Broadly construed, Kant grounds metaphysics on epistemology: that is, he limits what we can consider real to what we can know about what is real. It follows that although we can claim knowledge of the phenomenal world, any claim to knowledge of the noumenal world oversteps the bounds of pure reason. So as a matter of pure reason, we can neither prove nor disprove claims about the noumenal world. Perhaps in a most restrictive manner, pure reason alone forbids us from resolving the three most pressing issues of free will, the existence of God, and the immortality of the soul. So if we are to claim any knowledge of free will at all, it must be as a matter of practical reason. Indeed, Kant suggests that we posit free will as a postulate of pure practical reason, which is simply practical reason that is concerned with the noumenal world. Specifically, we postulate the idea of freedom as a transcendental idea, a “concept of pure reason” that is representative of, but not ultimately grounded in the phenomenal world. As a transcendental idea, the postulate of free will makes a claim about the noumenal world, but is not itself noumenally known or even knowable as a matter of pure reason.
Having briefly introduced Kant’s metaphysics and epistemology, I now outline his conception of freedom and its relation to the will. In the most general sense, freedom is a property of the will. The will is a causality that is characteristic of, and therefore presupposed, of any rational being. In particular, the will is the capacity to act in accordance with reason alone, independently of external causes from the phenomenal world. Further, the will acts by acting on the basis of maxims, which have the form “Perform action A in circumstance C for the end E.” In contrast then, a non-rational being, which does not have such a will, is only determined by external causes. On this reading of Kant, a rational being is a rational being insofar as it has a will, which is precisely the capacity to act in accordance with reason alone. Since freedom is a property of the will and a rational being is the only kind of being that has a will, it follows that a rational being is the only kind of being that can have freedom. Of course, Kant allows for the will to be unfree. But the will’s lack of freedom is meaningful insofar as it has the capacity to be free and yet is not actually free.
There are, in particular, several distinct but related types of freedom. Transcendental freedom, for one, corresponds well to our intuitive conception of freedom as the will’s capacity to be a “first” cause. As a “first” or “absolutely spontaneous” cause, the will is transcendentally free insofar as it is free to be a first cause in the noumenal world, the effect of which takes place in the phenomenal world. Crucially, such a cause must be itself uncaused and undetermined by any external cause in the phenomenal world. It is as such transcendentally ideal, not transcendentally real, since it is in principle unverifiable in the phenomenal world. Transcendental freedom means that I act in a certain way because I myself want to. To use a more concrete example, it means that I do my logic homework because I want to, rather than because I act merely in response to an external cause. This relates transcendental freedom with practical freedom in both the negative sense and the positive sense. While negative freedom is the will’s freedom from any external cause such as desire and inclination, positive freedom is the will’s freedom to both determine and obey its own laws. Indeed, positive freedom implies negative freedom, since the will is free to determine and obey its own laws only if it is free from any external cause. In fact, so important is practical freedom that Kant identifies the autonomous will as the will that is determined by reason alone in this way. In contrast then, the heteronomous will is the will that is not determined by reason as such. To sum, the free and autonomous will is transcendentally free in that it is itself an effective cause, and practically free in that it determines and obeys its own laws. I return to this conception of freedom later in the discussion to explore them in more detail.
Now that we have at least a general understanding of Kant’s metaphysics and epistemology, as well as his conception of free will, we are ready to tackle Kant’s main argument for free will from morality. To do so, I adopt what Henry Allison calls the Reciprocity Thesis, the thesis that morality reciprocally implies free will. In particular, I adopt the proposition that (1) morality reciprocally implies rationality, and also that (2) rationality reciprocally implies free will. It is this second proposition that I will examine in greater detail. Given these propositions, my reconstruction of Kant’s main argument is as follows:
- (Ax. 1) We accept morality on intuitive grounds.
- (P1) Morality implies rationality.
- (P2) Rationality implies free will.
- (C) From (Ax. 1), (P1), and (P2), we conclude free will.
As this reconstruction of the main argument suggests, the main motivation is to ground morality on a solid foundation. But morality depends on rationality, which in turn depends on free will. So in order to maintain morality, we must maintain free will.
The main argument begins with the axiom of morality, an assumption Kant takes us to intuitively accept as a “fact of… reason” (136). This acceptance of the axiom of morality (Ax. 1) affirms the antecedent of (P1). According to our intuitive conception of morality, morality has several characteristics. First, morality involves a law-a moral law-that commands me to act in a certain way. Second, this moral law is universal in that we conceive of it as binding on everyone without exception. Since the moral law binds everyone, I cannot, or at least ought not to, excuse myself or a friend for any crime. Evidently then, morality as a moral law, and indeed, the moral law, is an imperative in that it demands something of every person, and in particular, each person’s will. But what kind of imperative is the moral law? Based on our reflections on morality, Kant argues that the moral law is a categorical imperative, rather than a hypothetical imperative (106-107). Crucially, our reflections suggest that morality binds the will independently of the will’s desire or inclination. I cannot exempt myself from moral requirements simply because I feel like they do not apply to me. So the moral law cannot be a hypothetical imperative of the form “Do X if Y” (where X is an action and Y is an end that X can help bring about), since such an imperative is dependent upon subjective desire or inclination. Instead, the moral law must be a categorical imperative of the form “Do X” (where X is an action) in that it is absolutely and unconditionally binding on every will, regardless of subjective desire or inclination. But the moral law can only bind the will in such a way because it consists in reason, and the will is precisely the capacity to act in accordance with reason. It is clear then that morality, which consists in the universal moral law expressed as a categorical imperative, and in fact the categorical imperative, depends on reason alone (Preface). This first step of the argument establishes (P1), that morality implies rationality.
Now that Kant has shown that morality implies rationality, he moves onto the second major step of the argument to show that rationality, in turn, implies freedom. In terms of the argument as I have stated it, Kant now turns to (P2), the antecedent of which is affirmed by the consequent of (P1). Rationality, according to Kant, is normative in that it prescribes rules of both reason and morality. That is, thinking reasonably and living morally are the same kind of thing in that they are both prescriptions of rationality. At first, the idea that rationality prescribes morality may be a strange thought. Indeed, some philosophers, such as Hume, argue that morality is based on desire or inclination alone and thus has nothing to do with rationality. Yet let us first consider the relatively uncontroversial claim that rationality prescribes rules of logic as the rules of correct reasoning. Given “P” and “P implies Q,” the logical rule modus ponens persuades me to accept “Q.” Although I may for one reason or another reject the inference of “Q” from the given premises, I would do so in a way that is clearly contrary to rationality. Kant suggests that this violation of rationality means that the will is determined by external causes, whereas reason determines the will internally. Kant’s claim then is that just as rationality prescribes rules of reasoning, it prescribes rules of morality too. So the will that rejects the universal moral law is, in this sense, just as irrational as the will that rejects a valid inference from given premises. For the argument to work, Kant invokes the principle of “ought” implies “can.” He takes it that since the will ought to be rational, the will can be rational. This principle precludes the possibility of having any standard of rationality or morality so high that it is unattainable. In any case, the normative prescriptions of rationality, on both thought and morality, bind the will, which is by definition the capacity to act according to reason and assumed of every rational being.
As I noted earlier, the categorical imperative as an imperative of rationality gives us a command to act in a certain way. More precisely, we can use a priori reason to derive necessary actions or duties, the basis on which we are to act, from one of several formulations of the categorical imperative. Of these formulations, the one that accords best with the conception of freedom is the formula of autonomy. According to this formula, the categorical imperative commands the will to act in a way such that it both legislates laws for itself and at the same time subjects itself to those same laws. But to be sure, not just any law. The laws must conform to reason, which is universal to every rational being. When the will acts according to this formula of autonomy, it is the autonomous will. But the autonomous will is, as I mentioned even before I discussed the main argument, also the free will, in that it is practically free and (presumably) transcendentally free as well. As such, Kant’s conception of free will differs from competing conceptions in that far from being “free” from any constraints, it legislates and subjects itself to certain laws. This then establishes the connection between rationality, and through autonomy, freedom of the will. In any case, the derivation of free will from rationality and rationality from morality is now fairly clear. The good will acts according to reason as expressed in the categorical imperative. In turn, the rational will acts under normative prescriptions of rationality. Since rationality prescribes autonomy of the will, and the autonomous will is identical to the free will, the good will is at the same time both rational and free. Perhaps the only catch here is that strictly speaking, it seems that we can only derive practical freedom from morality, and not quite transcendental freedom. But if we also accept, as Kant insists we should, the postulate of free will as a transcendental idea, then we establish transcendental freedom as well. Accepting free will on this basis means that we have successfully completed the argument. Kant has shown (P2), and so we can validly infer from (Ax. 1), (P1), and (P2) that (C). This concludes Kant’s argument for free will from morality.
As a final word, there is one difficulty I want to mention before concluding. A source of tension lies in that while we are causally determined, we are also a first or spontaneous cause. It is not obvious how we can at once be determined by natural laws, just as rocks and trees are, and at the same time be a first cause whose effect takes place in the phenomenal world. So there seems to a sense in which causal determinism is compatible with freedom, yet Kant explicitly denies this possibility. Kant is an incompatibilist in that he thinks free will and determinism cannot both hold of the same world. Calvin Normore suggests that we can plausibly resolve this tension by postulating free will and determinism as holding at different moments in time in the same world (2008). For example, we can conceivably maintain that free will but not determinism holds from time t1 to t5, and also that determinism but not free will holds from time t6 to t10. Nonetheless, Kant seems to argue for a stronger conclusion that relies crucially on his distinction between the phenomenal world and the noumenal world. Simply put, he argues for a dualistic conception of us as being simultaneously present in both worlds. This way, we can maintain that while we are causally determined in the phenomenal world and subject to the laws of nature, we are also at the same time free in the noumenal world and subject to the laws of reason. Accordingly, the recognition of this dual presence in both worlds solves the tension between free will and causal determinism.
To conclude, I have shown Kant’s argument for free will from morality by appealing to the reciprocity thesis. Specifically, I reconstructed Kant’s argument as showing how the intuitive acceptance of morality implies rationality, and how rationality in turn implies free will. Following this, I mentioned the difficulty in conceiving ourselves as being both an effect of external causes and yet ourselves a first cause. Kant holds that while free will and determinism cannot both be true in the same world, the solution is to understand ourselves as being dually present in both the phenomenal world and the noumenal world.
Allison, Henry E. Kant’s Theory of Freedom. Cambridge: Cambridge University Press, 1990.
Guyer, Paul. Kant and the Experience of Freedom. Cambridge: Cambridge University Press, 1993.
Johnson, Robert. “Kant’s Moral Philosophy.” Stanford Encyclopedia of Philosophy. 2008. http://plato.stanford.edu/entries/kant-moral/ (accessed November 23, 2008).
Kant, Immanuel. Critique of Pure Reason. Edited by Paul Guyer and Allen W. Wood. Translated by Paul Guyer and Allen W. Wood. New York: Cambridge University Press, 2007.
-. Grounding for the Metaphysics of Morals. 3. Translated by James W. Ellington. Indianapolis: Hackett Publishing Company, Inc., 1993.
-. Grounding for the Metaphysics of Morals. Third Edition. Translated by James W. Ellington. Indianapolis: Hackett Publishing Co., Inc., 1993.
-. Kant’s Critique of Practical Reason and Other Works on the Theory of Ethics. 5. Translated by Thomas Kingsmill Abbott. London: Longmans, Green, and Co., 1898.
Normore, Calvin. “PHIL 301.” Montreal: McGill University, Fall 2008.
Pistorius, Hermann Andreas. “Rezension der Kritik der praktischen Vernunft.” In Materialien zu Kants “Kritik der praktischen Vernunft, by Rüdiger Bittner and Konrad Cramer, 175. 1974.
It is helpful to conceive of the postulate of free will as one analogous to the postulate of causality or the postulate of teleology in natural science. Just as the scientist examines the natural world as though causality and teleology were true, rational beings live as though free will were true. In both cases, the rejection of a postulate results in a kind of practical inconceivability: that is, the project in mind (of science or of ethical living) is impossible without first postulating the validity of some law, even if such a law is unknown and even unknowable.
Kant’s conception of free will is thus similar to the modern conception of freedom in political philosophy. A free state is often conceived of as one that is free from some influences but not others. Most importantly, it is free from external causes, but at the same time free to legislate and subject itself to laws in accordance with its constitution.
One critic notes: “I readily confess that this double character of man, these two I’s in the single subject, are for me, in spite of all the explanations which Kant himself and his students have given it, particularly with the well known antinomy of freedom, the most obscure and incomprehensible in the entire critical philosophy” (Pistorius 1974).
Andy Yu (’11) is a Philosophy and Economics major at McGill University
Cover image: “Be Free” by Celsojunior