Chapter 5

KANT’S ETHICS

_______________________

Thomas Aquinas held that there were two avenues whereby men could come to possess knowledge, the way of reason, and the way of faith, of faith in revelation. These were exhaustive but not exclusive ways of attaining to true propositions. Exhaustive, because no other source of knowledge was entertained; but not exclusive, because there were matters on which both reason and faith were equipped to pronounce. One of these issues was the existence of God. It was guaranteed both by five proofs, devisable by reason, and by the promptings and attractions of revelation, capacity to appreciate which is conferred on us by God’s Grace. Certain issues could be settled by reason(-ing) alone: a classification of natural kinds was preeminent among these. And certain things could only be settled by faith. The beginning of the world in time was one of these. That the world has a beginning in time is given in Genesis: reason is impotent to pronounce on the topic—there is no ground either for assuming it or for denying it. It is important that reason does not deny that the world has a beginning in time. For reason cannot teach what faith denies, and vice versa. Reason and faith are always at least compatible, and sometimes they pronounce identically: they can never pronounce oppositely. Credo quia absurdum (“I believe it because it is absurd”) is not a tenet of the Catholic tradition, but “I believe it although I don’t know why it is true” is. Faith is not entitled to contradict reason, but it is entitled to assert that for which no reason can be given.

COMPARISON WITH KANT

Now we know that Kant said explicitly that he “found it necessary to deny knowledge in order to make room for faith.”1 It must also be said that he thought this not only necessary, but possible. He was not deliberately engineering a grand confidence trick. But in saying what I just quoted he was permitting himself an uncharacteristic literary extravagance. For he did not accept, with Aquinas, two distinct faculties, which sometimes pronounced univocally on the same subject matter, but a single faculty, reason, which differentiated itself into two employments, a theoretical one and a practical one. The theoretical use enables you to determine what is true; the practical use enables you to decide what to do. And each use has its pure aspect, guaranteeing the truth of what you say or the validity of what you do on a priori grounds alone, without reference to experience.

Kant assimilates pure reason in its practical employment to faith because it is given the office of pronouncing on matters traditionally within faith’s purview. More specifically, pure practical reason (which I shall often refer to without the qualification “pure”) proves or grounds or necessitates the belief in the objectivity of the moral law, the freedom of the will, the immortality of the soul, and the existence of God. It is with the first two that we shall be concerned in these lectures, as they are treated in the Groundwork, the text for these lectures, the Second Critique having the other two as part of its burden.

So: unlike Aquinas, Kant entertains not two faculties, but a single faculty in two employments. Also unlike Aquinas, the two employments never conspire to yield the same conclusion: there is no overlap. But like Aquinas, reason theoretic never contradicts reason practical: indeed, when the former presumes to discuss what turns out to be the proper subject matter of practical reason, it falls into contradiction with itself. It produces antinomies, discrepant conclusions by equally cogent arguments, and it is by showing that and how it does so, in the transcendental dialectic, that Kant is denying knowledge and certifying faith.

WHY TWO EMPLOYMENTS OF A SINGLE FACULTY?

The question now arises, what were Kant’s motives, and what did he advance as justifications, for treating the sources of knowledge and of moral behavior not as two separate faculties, but as different employments of a single faculty, reason? The question sounds academic, an appeal for a distinction without a difference. It might be argued that faculties are individuated according to what they secure, according to their uses, so that to talk of two different uses of a single faculty is to talk idly, as though one were to say that it’s the same ability, but it does not enable the same thing. But Kant himself placed importance on the identity of the faculty, as is manifest in the section entitled by Paton “The aim of the Groundwork.”2 So we must try to answer the question, why must it be one and the same faculty which guides us to correct conclusions in theory and moral behavior in practice?

FIRST REASON

We may hazard that the first, seemingly superficial motive, was to arrogate to morality the prestige which attaches to valid reasoning in matters of theory. Faced with the conclusions of Aquinas, a man might accept those which he attributed to reason and, in consistency, deny the precepts required for governing his behavior toward man and God which stem from faith. Such a man might be accused of deficiency of religious appreciation: he could not be accused of contradicting himself. But if the content of faith is given not by faith properly so called, but by reason, albeit in a special exercise, then a man who ever exercised and depended on reason, when recognizing its authority in matters of theory, could not consistently disavow its claims in the domain of practice. A man who was theoretically rational, when pressed by a perverse interlocutor who refused to accept his well-grounded conclusions, would eventually have to say to him: you must accept them, since you are a rational being. To such a man Kant could say, in a moral context, you must accept this imperative, since you are a rational being. To put it in another way, theoretical reason provides the grounds or reasons for making assertions; practical reason the grounds or reasons for proper behavior. It is important for Kant that there be no ineradicable equivocation on “grounds” or “reasons” in these statements, for then a man who reasoned theoretically would be under an obligation, as a rational being, to be moral in practice, on pain of inconsistency. And to say that there is no ineradicable equivocation on “reason” here is to say that we have to do with a single faculty in two employments. “Because you are possessed of reason” is the single final answer both to questions like “Why must I assert causes?” and to questions like “Why must I behave morally?”

A PARADOX IN THE ABOVE

Now it must be admitted that there was a ring of peculiarity in the final rejoinder the theoretically reasoning man presented to his recalcitrant opponent. “Because you are a rational being” is an odd-sounding reminder, and I suggest that its oddness is not a trivial one, a consequence of its infrequent incidence in discourse. Perhaps the oddness lies in the following: the perverse reasoner has evinced in his refusal to accept the consequences of valid reasoning that he is not in fact rational, or not willing to be so. And so it is futile at this point to tell him that he is so or has resolved to be so. To use a modern jargon, he may be opting out of the language-game of valid inference. This paradox has its practical counterpart, in the form that it is pointless to point out to a man that he is a rational agent unless he is striving to be one in the first place. This paradox is not solved but only expressed in the closing part of the Grundlegung. For here we learn that we are in reality noumenal beings, and that our observable characteristics, our desires and inclinations, are merely phenomenal, possessed of only apparent reality. We are invited therefore to eschew the latter, since we are fundamentally rational beings. But since we are fundamentally rational beings, as free noumena, there is no point in telling us to overcome desire—the victory of desire is only the outward show of some secret noumenal, and hence moral, even holy, doing.

Kant wished to establish that men are obliged to obey the moral law. The instrument for guaranteeing this is the demonstration that they are in fact noumena: but this proves too much, for it makes it a certain truth that they are doing their duty, and are not merely obliged to be.

A SECOND REASON

This reference to the phenomenon/noumenon distinction leads us to another of Kant’s motives for fusing knowledge and faith and making of them two employments of a single faculty. The noumenal world, or the thing-in-itself, has its existence confirmed by the exercise of theoretical reason, though the latter cannot pronounce on its character. Theoretical reason discovers itself to be possessed of synthetic a priori knowledge, in, for example, the theorems of geometry. It is held that this can only be the case if the objects of such knowledge are constructed in modes which are part of the apparatus of cognition, rather than in modes which appertain inextricably to them as they are in themselves. In the case of geometry, the relevant nonexperientially dependent furniture of the mind is the a priori intuition of space. But since we impose forms, alien to it, on the world, the world as it is in itself must differ from the world as we experience and know it. So the existence of a world possessed of more secure reality, unconditioned by our modes of cognition, is proved by theoretical reason. And though it cannot think about it, because to think is to apply categories, and categories are applicable only to our experience, since they are our categories, it can nevertheless, Kant says, think it; that’s to say, so I opine, it can be rationally assured of its existence.

Theoretical reason not only permits morality, but requires it. Theoretical reason finds itself unable to render a consistent account of the world: to complete its work, reason must issue in another employment, without depending on something extraneous and nonrational, operate in a new manner. For if what complemented theoretical reason were not the same thing used differently, but were rather something like Thomist faith, there would be consistency in human knowledge, but the reliability of reason would be weakened since reason would be required to whirl in antinomies forever. Thomist reason did not contradict itself; it was merely silent. So faith could supplement it. But reason is doomed to antinomies only if it misuses itself, and does not issue in a practical use. In other words, morality is construed as practical reason not only in order to vest it with prestige, but also in order to make good theoretical reason’s antinomial debts. Knowledge is denied in order to make room for faith, and faith is established in order to complete the canvas of reason, which cannot allow that it contradicts itself, without placing its pretensions as a faculty in direct jeopardy, as it would if it required something alien to itself to take in its dirty laundry.

OBLIGATION AND MOTIVATION

Another reason why morality is to be identified with practical reason (and here I focus on this phrase, not on two uses) stems from a general difficulty perennially confronting moral philosophy, the problem of establishing a connection between the recognition of obligation and the disposition to act as one takes oneself to be obliged to.3 The problem is the following: if moral precepts are the product of ordinary intellection, of what Kant would call reason in its theoretical employment, then it is hard to see why a man who became apprised of them should feel motivated to obey them. If we accept naturalistic reductions of moral statements, which make them truth-valuable, we lose the characteristically moral force of the statements: they fail to impinge on the will. Mrs. Foot makes value certain but irrelevant. The nonnaturalists, like Hare, have no difficulty in showing how we are disposed to follow the moral precepts we recognize: for the test of recognizing them is that we follow them. But now when a man “recognizes” a moral precept, he is no longer, as in the first case, coming to appreciate something antecedently valuable: he is valuing something, making it valuable. There is no perception of obligation. The first attaches morality to the intellect, but cannot account for the volitional element in it: the second attaches morality to the will, but cannot satisfy the moral philosopher’s demand for rationality in ethics. Thus Hare’s “arguments” are of a peculiar kind, where reasons have no autonomous force.

Now if it can be shown that the faculty exercise of which constitutes morality is both will and intellect in one, that’s to say, is practical reason, then this problem will be solved. For it is neither mere will, which is blind, nor mere intellect, which is impotent, but a practical employment of intellect, that is, intellectual will. Such an employment is necessary to solve the riddle of obligation and motivation. This, I think, is the teaching of the first part of the Grundlegung. And how such an employment is possible, how pure practical reason (or a categorical imperative) is possible, it is the task of the second half to explain.

The duality of obligation and motivation is present in Kant’s ethics. Practical reason is that which discerns the validity of the moral law; but it is also that which inclines us to follow it. (Mrs. Foot: reason discovers; not will, but desire, moves, and that means, for Kant, heteronomy. Mr. Hare: will moves; there is no discovery of reason. Kant wants reason both to discover and to move.) I can put this point otherwise by stressing the double-barreled character of “determined” in Kant’s use of sentences like: “That action is good which is determined by the moral law.” For the determination of the action by the moral law is both efficient and formal. It is efficient in that consciousness of the law causes the action, makes it happen, and formal in that the law makes the action a good action: it is because it satisfies the moral law that the action is good: the moral law is the criterion of its goodness. (On formal and efficient causes: Why was it a bad film? Because Fellini directed it. Because it was dull, repetitious, unimaginative—last has Janus-force. The formal cause explains why something comes under a certain description.) The problem for Kant is to show how it is possible that the mere entertaining of the moral law should incline the entertainer to follow it, to show how reverence for the law is, as he puts it, self-wrought by a rational concept, not dependent on desire. Because he believes that every action phenomenally viewed is a result of desire, he needs a noumenal world to make the categorical imperative possible. But once he has generated this world, the categorical imperative becomes not merely possible, but forever actualized, and he winds up proving too much.

THE ARGUMENT FROM ARGUMENT

The Grundlegung is an affair of two arguments, the conclusion of the first being the premise of the second. The first argument, which is characterized as analytical, proceeds to establish what the moral law must be like, whether or not there is one, and the answer to this is given in the five principal formulae of the categorical imperative. It also involves establishing the necessary and sufficient condition of the validity of this imperative, which is freedom. The synthetical argument proceeds in the reverse manner, proving the legitimacy of assuming what the analytical argument shows to be required, the validity of asserting the freedom it demands, and deriving the actuality of the moral law from that freedom. Kant believes that to show that man is free is to show that he is under the moral law. To be free just is to be under the law and to be under the law just is to be free. Freedom is not freedom to choose between this or that course of action, where one course may be that which instantiates the moral law. Freedom is not a matter of choice at all, for the choice is between freedom and bondage to desire: if your action is not necessitated by the law, then it is contingent on desire, and therefore unfree. We are not free to choose, but we can choose to be free.4

(For Kant, in a certain sense, “can” entails “ought.” For I can = I am free = I can cause lawfully = I am autonomous = I am moral.)

Kant thinks, then, that if he has freedom, he has all that he wants and needs, namely morality. From the fact that he thinks this, we can perceive another incentive for making morality an affair of practical reason. For in the second part of the “Third Section,” entitled “Freedom must be presupposed as property of the will of all rational beings,”⁵ the freedom with which practical reason operates is either derived, or made plain to us by analogy, in fact if not in profession, from the freedom with which theoretical reason operates: Kant says, “We cannot possibly conceive a reason consciously receiving a bias from any other quarter, with respect to its judgements, for then the subject would ascribe the determination of its judgements not to its own reason, but to an impulse” (emphasis added).⁶ If I am engaged in any piece of ratiocination, say in proving the Pythagorean theorem, then I cannot allow that I reached my conclusion as a result of glands or toilet training, but only as a result of the free operation of my intellect, in obeisance to the laws of geometry. My reasoning is most free when these laws necessitate it. The assumption that they do must be made by reason, and this assumption will establish morality, if freedom establishes morality and morality is a matter of practical reason. We have to treat ourselves as free when we are theorizing, thinking about what to say: we have to treat our assertions as governed, formally and efficiently (the latter to avoid mere orthe doxa, right opinion), by the dictates of reason. Similarly, if the deliberation which precedes action can be called practical reason, then we must treat ourselves as free when we are acting by the dictates of the moral law. If there were no laws of Euclidean geometry, or similar ones, then all reasoning about plane figures would be governed by laws of association: reason would suffer from heteronomy.7 If there were no laws of morality, then all action would be governed by desires and impulses: we would not be free. These statements appear strained, and I suggest they are because they presuppose an exclusive and exhaustive distinction between motivation by reason and motivation by desire, this being one of the many dichotomies, whose validity is suspect, which Kant uses.

COMPARISON WITH PETERS

Richard Peters believes in a weaker version of this thesis, in his belief that explanation is either by the rule-following model or causal in character. His is weaker because there is no unique set of rules following which emancipates us from the causal nexus. So if we can cast doubt on his position, we shall a fortiori bring into question the stronger position that action is governed either by desire or by a single principle, the moral law.⁸

Peters’s doctrine is not primarily a theory of the freedom of the will, but rather of the explanation of human behavior. But from it a position on the former can be extracted, insofar as being free involves, as Kant was sure it did, some sort of independence of the causal nexus. But Peters’s conviction that actions are uncaused is not founded on a distinction between phenomenon and noumenon, but on a distinction between cases where causal explanations are in order and cases where they are out of order. He works with a model of man as a rule-following animal. Man’s capacity to guide his behavior according to precepts differentiates him from the animals, whose behavior, to use Kant’s apt language, may operate according to law, that is regularly, but not for the sake of law. If a man is engaged in a given purposive activity, then what he does in the course of that engagement is explained, for Peters, by citing the rules governing the activity which the man is following. Thus the purposive activity of chess playing, the moves within that game, are explained by the rules of chess and the agonistic aims of the chess-player. These do not explain the movement of the castle causally, because here the explanans entails the explanandum, and this, as Hume showed, does not hold in causal explanations. Given the rules and the aims, the action follows logically; that it follows is not guaranteed by past observations, but a priori. Now if a man makes a move in chess which is either contrary to the rules or in manifest contradiction with the project of victory, a causal explanation of his failure is in order, and this explanation may be a neural story. When the rule-following activity is interrupted, the interruption is to be accounted for by generalizations from experience, and psychological theory. If a man adds 7 and 5 and produces 12, there is no demand for a causal explanation; it would serve no function, since the arithmetizing is intelligible on its face, as an instance of rule following. If he gets 11, then a causal explanation of his slip is wanted and sought. But Peters does not say only this. He does not contend merely that in rule-following behavior causal accounts have no point, as many may agree. It is a stronger sense in which they are out of order. It’s not true that they are available but pointless; they are even unavailable. This is not to deny that there is a neural tale behind even the effective performance, where error is absent. But, Peters insists, this tale can only provide necessary, never sufficient, conditions of the action while it provides sufficient conditions in the case of faulty behavior. And the reason why it is not sufficient to explain the action is that it is only sufficient to explain bodily movements, and there is no one-one correlation between actions and bodily movements, enabling you to pass from sufficient accounts of the latter to sufficient accounts of the former.

An example he uses is the action of signing a check. The bodily movement involved in spelling out one’s name can be accounted for causally. But a full description of the bodily movement falls short of a description of the action: simply scrawling one’s name is not signing a contract, unless this is done in the relevant rule-governed context (in the presence of witnesses, etc.). This structure which makes the movement of the hand a case of contract signing, cannot, it is held, be explained causally. What is more, many bodily movements would count as the single action of signing a check, given the appropriate context. And, conversely, the same bodily movement might count as a variety of actions, given a differential in the action-making contexts. From this it is supposed to follow that actions are not liable to causal explanation.

Now something in this is undeniable: if we give a causal explanation of the movement of the check-signer’s hand, we have not explained it as a case of check signing. But must the fund of relevant explanatory causes stop here, in the physiology of the signer? I suggest not. I do not see why a causal explanation could not be offered for the presence of the context, and for the adherence to the rules which are observed within it. The rules do not explain causally, but the force of appealing to them depends on a causal explanation, not on merely assuming, with Peters, that man is a rule-following animal. This would dispose of the conclusions drawn from the absence of one-one correlations. The action can be causally explicable, fully and completely, if the entire set of factors is taken into account.9

Consider, for further clarification of the position and my criticism of it, the following passage in Peters:

We can ask why Jones is mean or why he eats fish. The way it would be answered would depend on the context. It might be answered in terms of a rule-following type of explanation like “because he is a Scotsman” or “because he is a Roman Catholic.” This would assume some established set of norms and a system of training for handing them on. It would be radically different from the explanation “because he is an anal character” or “because he is an oral character.” For these explanations would presuppose that Jones is in some way a deviant from the norm of the circle in which he had been trained. It would state special conditions in his upbringing which occasioned this deviation.¹⁰

For various practical reasons we want to understand the people with whom we have to deal in social life. If I know that all or at least most Scotsmen are parsimonious, I no longer find it an intractable and odd fact that Jones is stingy when I am told that he is a Scotsman. I now know how to deal with him, what roles to play, because I know many things about Scottish behavior. For practical purposes, the explanation can stop here, and I shall be quite satisfied: I shall not crave Humean causes. I think, in general, facts of social life like this tend to insinuate themselves into the so-called logic of language, and it is well to be aware of them to separate depth from superficial grammar, for I think that here we have a typical instance of how an unwillingness to sociologize in an elementary manner leads the philosopher astray, tempts him to think he has discovered a “conceptual truth.”

For I might not only be a practical man; I might have less limited cognitive aspirations. I might, for example, be a social scientist. The question, “But why are Scotsmen parsimonious?” though out of place in most ordinary situations (out of place, irrelevant, but not, surely, without sense) is perfectly posable in a scientific context. And it is impossible to rule out a priori that the explanation for the parsimony Scotsmen manifest in their behavior may lie in a general condition of anality that prevails among them. In short, the causal explanation for the deviation from one rule-following model, an instrument Peters accepts, may also serve to account sufficiently for the institution of and adherence to another. All that is wanted is a removal of the blinkers which lead us to focus on our immediate needs in immediate situations.

To return to Kant. One of his dichotomies, and he has many, as will emerge later in these lectures, is that between caused behavior and law-obeying behavior. And his law is not multiform, like Peters’s, but sole and single, the moral law. There is only one emancipating language game. So if I have succeeded in casting doubt on Peters’s liberal dichotomy, I have a fortiori made suspect Kant’s more rigid one. And he takes himself to need this dichotomy, because it confers freedom, and because without freedom there can be no moral law.

THE FUNCTION OF REASON

This is a very important section, which seems to lie near the beginning of the Grundlegung in sovereign independence of the rest of the work. In reality, however, it is indispensable, providing what may be called the teleological linchpin of the whole argument. For Kant has to do two things: he has to show that there is a course of behavior dictated by reason and by reason alone (this being the way of universalizability), and he has to combat a “so what” response to the exhibition of this course. These are two senses in which he has to show how pure reason may be practical: how it can make a recommendation, and how it can incline the will to follow that recommendation. In the section under consideration, he seeks to show that the whole point and purpose of the existence of reason is to produce moral conduct, so that there would be something incoherent about not using it for that end (like using a knife to bang in a nail and a hammer to spread butter). It will be useful to quote a major portion of the passage:

In the natural constitution of an organic being—that is, of one contrived for the purpose of life—let us take it as a principle that in it no organ is to be found for any end unless it is also the most appropriate to that end and the best fitted for it. Suppose now that for a being possessed of reason and a will the real purpose of nature were his preservation, his welfare, or in a word his happiness. In that case nature would have hit on a very bad arrangement by choosing reason in the creature to carry out this purpose. For all the actions he has to perform with this end in view, and the whole rule of behavior, would have been mapped out for him far more surely by instinct and the end in question could have been maintained far more surely by instinct than it ever can be by reason. If reason should have been imparted to this favoured creature as well it would have had to serve him only for contemplating the happy disposition of his nature, for admiring it, for enjoying it, and for being grateful to its beneficent Cause—not for subjecting his power of appetition to such feeble and defective guidance or for meddling incompetently in the purposes of nature. In a word, nature would have prevented reason from striking out into a practical use and from presuming, with its feeble vision, to think out for itself a plan for happiness and for the means to its attainment. Nature would herself have taken over the choice, not only of ends, but also of means, and would with wise precaution have entrusted both to instinct alone.11

A series of critical questions must now be posed. First of all, quite apart from the metaphysical premises about teleology that we are invited to accept, we are faced with a factual assumption the validity of which is far from clear. I refer to Kant’s conviction that reason is ill-suited to produce happiness. Certainly bad reasoning can lead to misery, but is there not a case for saying that the informed exercise of the rational faculty is the best guide to happiness we have? Now this might be so and it might also be so that reason is to be used for morality; why can’t the proper purpose of the knife be to cut and to spread? For if reason is not the guide to happiness, then what is? And if Kant replies that it is an unwarranted assumption, that something must be the guide to happiness, then why must something be the guide to morality? Anyway he tells us that happiness is a purpose of man.¹² At any rate, Kant doesn’t tell us what the guide to happiness is: he says that instinct could have secured happiness more easily. He puts this in contrary-to-fact terms: could have secured, rather than can secure. Clearly instinct often leads to ruin, and it is reason which can limit it, and at least prevent extremes of unhappiness—for this it seems to be suited. Again, it would seem that any purpose otherwise achieved might be best achieved by an overridingly powerful instinct. Why not morality then?

Well, Kant is probably right when he elsewhere insists that there is a distinction between doing the right thing (e.g., on instinct), and doing the right thing because it is the right thing, the latter alone being conspicuously moral. So instinct cannot be the basis for moral behavior, not because it hasn’t in fact been designed for that end, but because it couldn’t have been. Morality involves consciousness of what you are doing. Therefore, although the teleological argument fails, there is, buried obscurely in it, a valid conceptual point about morality, that it necessarily involves reason in the sense of consciousness, or at least consciousness of a certain kind: and reason includes consciousness.13 But this is not all Kant wants to say: for he wants reason not only to be necessary for morality, but sufficient for it as well. Kant has another reason why instinct cannot lead to moral behavior, a reason which we may feel loath to accept. It is because for him the moral course is necessarily the one which conflicts with inclination, desire, passion, since these are the phenomenal causes of behavior, and morality must be counterphenomenal. This is why Kant believes that he is most morally admirable who has a rotten soul, but acts against its natural dictates. A naturally benevolent man, he feels, lacks all moral worth. This we shall consider later, when we consider his distinctions.

Apart from all this, there is in the passage an attempt to prove two things, which he scarcely differentiates, but which are clearly separate. The first is that the purpose of reason is morality; the second is that morality is man’s highest purpose. The two are conflated in the following: “another and much more worthy purpose of existence, for which, and not for happiness, reason is quite properly designed.”¹⁴ And while there is a feeble attempt to prove the former, the latter seems merely to be posited, as if it could ride in on the former’s coattails.

That reason cannot provide happiness is an essential conviction for Kant: eventually, in the Critique of Practical Reason, it is a dogma required for the practical proof of God’s existence, as guarantor of the summum bonum, which is a proper proportioning of virtue and joy. And the point is stressed in the section on “How are imperatives possible?,”¹⁵ where he seems to hold with Hobbes that no man can secure to himself happiness by his own efforts.¹⁶

The argument may be reconstructed: you cannot, however much you reason and think, produce happiness for yourself, for you cannot divest yourself of unsettling desires. So why not refuse to heed their proddings anyway, and be moral, for this course is far nobler? And I can show you that it is possible to be so, for you are in truth a noumenal being. This teleological argument is odd in many ways, one being this—that pure practical reason is a noumenal affair beyond our natures, but is said here to be authored by nature.17

THE TWO USES OF REASON DISTINGUISHED

I opened these lectures by accounting for Kant’s insistence on the unity of the rational faculty in its two employments, theoretical and practical. I now want to treat what he considers to be a significant difference between the two uses, apart from the obvious distinction between providing correct knowledge and leading to proper practice.

In sum, the difference is this: whereas in the first critique he is delimiting the capacities and indicating the dangers of the use of human reason, of the cognitive power of flesh-and-blood creatures such as we at least appear to be, he believes that the second stage, the critique of pure practical reason, is concerned not with human reason, but with reason-as-such, and that the moral law it discovers must be valid for all rational beings. It is no part of the doctrine of the first critique that intellection as such requires the use of categories like cause, substance, etc., but only human intellection, connected as it is with the human forms of intuition, space, and time.

Thus he says in the section entitled “Review of conclusions”: “to determine the whole faculty of pure practical reason … we must not make its principles dependent on the particular nature of human reason, though in speculative philosophy [i.e. critique of theoretical reason—GAC] this may be permitted, or may even at times be necessary.”18 Of course, he does not deny that in the application of the moral law, the nature of the being to whom it is applied must be taken into account—but it must not be taken into account in deriving it. In fact, there can be no categorical imperative, unless issued to a being who is not only rational in nature: otherwise there is no need of or sense in commanding him to be rational, since nothing obstructs rationality in him: he is a holy will, of which more below. You can’t order someone to do something he is doing. But note that it does not follow from his not already doing it that he wants not to do it. So it doesn’t follow from the possibility of an imperative that there is a desire that conflicts with it.

The difference between the two uses can be put strikingly in the following way: it is precisely that with which theoretical reason is not equipped to treat that practical reason is, and vice versa. Theoretical reason must always deal with experience, slips into antinomies when it reaches beyond it; pure practical reason must never deal with experience in the first instance, but only with the noumenal which lies beyond it: if it bases its judgments on experience, it is implicated in what Kant calls the heteronomy of the will. Nothing in experience has any intrinsic worth, but only value conditioned on its being desired. So to find what is of intrinsic worth, flight beyond experience is necessary.

To put the point in yet another way. The function of theoretical reason is to render an account of experience, of the world as it is for us. From this it follows that its starting point must be that world, with its peculiar character as conditioned by our forms of intuition. But the function of practical reason is to render imperatives indicating the moral course in this world of experience. So these imperatives cannot be drawn from it, since so to draw them would involve a passage from “is” to “ought,” and such an “ought,” since the passage is impossible, will turn out to be a disguised “is,” and, what is more, the “is” of some or other sort of desire.

Consider the following parallel. Kant would say that we cannot know about the thing-in-itself, but we can know about the thing as it presents itself to us. So we must content ourselves with the latter. Might one not say, analogously, we cannot know whether or not there is a moral law, but we do know what, terrestrially, things are like; we know that men have certain desires, and seek to satisfy them—in the absence of access to realities transcending those of our acquaintance, we might as well content ourselves with building our ethic on these facts of human nature. There may be a good beyond desiring. But so what?

Now this is precisely the parallel Kant refuses to recognize. And it seems to me that he is justified in repudiating it. For there do seem to be admissible questions which adherence to it would forbid to be posed, like, for example, what is the dignity or worth of this desire, this need as opposed to that one, etc.? You say that this is good because you desire it, but is it really good? In the absence of answers to these questions, we may well wonder with Nietzsche, who was prepared to give up on mankind, whether our conduct, our justifications for it, have any basis at all. True, you can construct analogous questions in the theoretical realm: you say that you know that striking caused the match to burst into flame, but is this really knowledge? And in a sense, Kant would say it isn’t, but so what? It is all we require. But can we be similarly satisfied in the practical sphere? Or is there not an irrepressible yearning to do what is right, in an unqualified sense? Moral skepticism has a force which epistemological skepticism lacks.

To put the point otherwise: to the question about the match, one might respond, yes, it is knowledge, in a conditional sense, but knowledge none the less. But if you admit that what you regard as good is only conditionally so, you seem to have given up the claim that it is good, in a way in which you haven’t in the parallel case given up the claim to know. “Conditional” is alienans for good, but not for knowledge.

THE DISTINCTIONS

I now want to list and discuss a series of distinctions, made either explicitly or implicitly in the Grundlegung, on the validity of which its doctrine depends. There are twelve such distinctions, cutting the moral realm off from the realm of desire. Each pair must exhaust the range of actions and no member of any pair may apply to an action to which the other member applies. Also, each pair of distinctions must be equivalent in extension and division with every other. They must all cover every action and slice them off at the same point. All the distinctions refer to actions or to some construct of actions. If we symbolize them as A-B, A¹-B¹, A²-B² … A¹²-B¹², then we can represent the required relations among them as: A ~B1, A¹ > ~B¹ … A¹² > ~B¹², which represents the congruency or similar extension and division of the distinctions. From all this there emerge 90 possible errors in Kant: 24 relating to the exhaustivity and exclusivity of each distinction, 66 to the question of congruency. I think Kant fails on about 40 of the counts (one is enough, because of the equivalences) but I do not propose to exhibit all the failures, since very many of them are so close to one another anyway. What I shall do is explicate each distinction, and challenge only some on one or other of the three counts:

Non- or Immoral		Moral
1	Contingent	Necessary
2	Hypothetical	Categorical
3	Desire	Will
4	Desire	Duty
5	Purpose	Principle
6	Caused	Free
7	Heteronomous	Autonomous
8	Selfish	Selfless
9	Empirical	A priori
10	Empirical	A priori19
11	Phenomenal	Noumenal
12	Particular	Universal

1. Every action is governed either by a necessary or by a contingent law, none by both.

2. Every action is obeying either the categorical imperative or a hypothetical imperative.

3. Every action is either the result of will or the result of desire.

4. Every action is either for duty’s sake or to satisfy desire.

5. Every action is either governed by a principle or governed by a purpose.

6. Every action is either free or caused.

7. Every action is either autonomous or heteronomous.

8. Every action is either selfless or selfish.

9. The spring of every action is known either a priori or empirically.

10. The spring of every action is known either a priori or empirically.

11. Every action can at one time be viewed either noumenally or phenomenally.

12. Every action is either universal in nature or particular in nature.

(1) Contingent/Necessary. In the section “The need for pure ethics”20 Kant tells us that for a law to have moral force it must carry with it absolute necessity. This is to say that it cannot apply to this or that rational being only, but to rational beings as such. So there are laws of action which are valid necessarily for rational beings, and laws which are valid, but contingently, given their particular (12) natures, their purposes (11), their desires (3, 4). The moral law is necessary, indefeasible; laws relating to the satisfaction of our desires or needs are contingent: were our desires or needs to change, they would change. But there is nothing on which the moral law is based, save bare rationality itself. And there cannot be different kinds of bare rationality, or changes of rationality, as there can be different kinds of needs. In sum, all actions are governed either by necessary laws of rational nature, discovered a priori, or by contingent laws which depend on what we are, on our desirous nature, which we discover empirically, but none by both.

(2)21 Hypothetical/Categorical. All actions can be represented either as fulfilling hypothetical imperatives or as fulfilling the categorical imperative, and none can be represented as fulfilling both. Hypothetical imperatives state only what is right for some purpose. The categorical imperative declares an action to be right without reference to any other purpose. If I issue a hypothetical imperative, you may ask of what it enjoins, what is it good for? This question is disallowed in the case of the categorical imperative. Following it is not good for anything. “What’s the Good of following it?” does not apply.

Harking back to the first distinction, the goodness of what it commands does not depend on, is not contingent on, anything else. Now you cannot distinguish between the hypothetical and the categorical by a grammatical criterion alone, by asking whether or not the word “if” has an incidence in the imperative as formulated. The criterion is rather to be sought in how the imperative is backed if challenged, or, strictly, on whether the imperative is backed. Thus, “If something is divine, worship it” is hypothetical in form, whereas “Come here!” is categorical in form. But “If something is divine” gives you no incentive or further reason for worshipping what you should worship; if no backing is offered, then it is categorical, whereas the backing “Otherwise I’ll strike you” can be offered for “Come here!” I say can be, because the imperative may not be backed, although Kant would say that such an imperative unless implicitly backed would never be obeyed. Unless, he believes, you can represent fulfillment of the imperative as satisfying some want of the person to whom it is issued, he will not follow it. The only imperative he will follow which is unbacked by a want is the imperative of duty, the moral law.

So now the criterion looks like being the possible incidence of the phrase “If you want …, then …” But even this is misleading. For consider the case, “Stand on the right.”²² In appearance this is categorical, like “Come here.” But we know that what it intends is, “If you want to stand, stand on the right.” That is, you may walk instead of standing, but if you stand, it must be on the right. But standing on the right isn’t something which fulfills my desire to stand or at any rate no better than standing on the left does. So the incidence of “If you want” fails as well.

The criterion which does the work is the following: an imperative is hypothetical if “If you want …” can have an incidence in it, not as a condition that makes the imperative applicable (as in the subway case) but as a ground for doing it: you do it because of something you want. The hypotheticity resides in the availability of a backing, and the “If you want to stand …” does not back in the way that “If you don’t want to be struck” backs. The former only states when a categorical imperative is in order, when it applies, not why it applies.

Fusing 1 and 2, we get: an action is moral if and only if it exhibits a necessary law, and obeys a categorical imperative, and if and only if it does not exhibit a contingent law and does not obey a hypothetical imperative.

Hare: hypothetical imperatives are derivable from indicative minor premises alone. For example: “This acid burns: If you don’t want to get burned, don’t touch this acid.” Hypothetical imperatives are thus reducible to indicatives: “If you touch this acid, you will get burned.” This schema is inapplicable to the Tube stairway case. What indicative is “If you stand, stand, on the right,” equivalent to? Not clearly, if you don’t place yourself on the right, you won’t stand. There is no indicative which can have the force of an imperative here. And this is what makes it categorical.

To be a nonnaturalist, to hold that you cannot move from “is” to “ought,” is to hold that moral rules are imperatives irreducible to indicatives which have the same force.

(3) Desire/Will.23 All action is either a result of desire or a result of will, of volition, and none is a result of both. This is OK only as a capsule formulation, though sometimes Kant will say it whole-hog. Actually will is present in the desire cases as well, but inefficacious will, will untrue to its own free nature, autonomous, in bondage to desire.

The distinction can be rephrased thus: in answer to the question “Why did he do x?” we will ultimately rest either at the answer because he desired x, or y, or because he willed x, or y. What you want you cannot will, though you can will a way to get it. You cannot will what you want, because when you want you are driven, unfree, and you only will when free. Conversely, what you will you cannot want, since you will freely, and you want in bondage. Skipping ahead to distinction 11, we see that Kant is a “wanting determinist,” given distinction 6 as well. All action, phenomenally viewed, is caused by desire.24

This fusion is perhaps effected by a misconstruction of the propriety of always saying “Because I wanted to” at the end of the chain of explanations of what I did. For this often only means: no further reason, not I was spurred by desire. (In other words, we reach the want that is entailed by the action, and which therefore cannot explain it.) And that for which no further reason is given, no desire, for example, is freely willed, chosen: but many things are so willed, not merely moral behavior. So here we have one crucial breakdown in the series of equivalences. But continuing the evolution of the position, we now see that either what I do is contingent on a desire which pushes me to fulfill a hypothetical imperative, or will adopting a categorical imperative, in actions necessary in themselves, without further basis.

(4) Desire/Duty. All actions are undertaken either in order to fulfill a desire or in order to fulfill a duty and none for both. Note: this does not exclude the possibility of a dutiful action fulfilling a desire or a hedonic action fulfilling a duty: the exclusivity comes in at the level of “in order to,” “for the sake of.” This division is not difficult to appreciate, if you have followed what I have said so far. In the section “the exclusion of interest”²⁵ Kant explicitly says, “In the case of volition from duty all interest [i.e. desire—GAC] is renounced, which is the specific criterion of categorical as distinguished from hypothetical imperatives.”²⁶ In the section “review of the whole argument”²⁷ he says that unless we are conceived as doing our duty, we must be conceived as subject to the physical law of our wants. So we either submit to desire, follow contingently grounded hypothetical imperatives, or will to do our duty, which is expressed in the categorical imperative, which needs no basis.

Tugendlehre: “A duty is a constraint to an end not gladly adopted.”²⁸ Therefore duty cannot coincide with desire.²⁹

(5)³⁰ Purpose/Principle. This seems a very odd distinction, not as readily assimilable as the rest. In fact it is crucial, as Kant’s doctrine of the unique and original goodness of the good will, which I discuss after the distinctions, is connected with it. For he believes that the goodness of the good will is not contingent on the goodness of the purposes it pursues, but on the principle governing it.31

Now in the case of most of Kant’s distinctions, their exclusivity seems to be in doubt: it often seems possible to present a counterexample, an action which is describable as both, e.g., sought out of desire and done for the sake of duty. But one can at least think of clear cases where one is present and the other absent, where dutiful action goes against desire and desire-driven action conflicts with duty. In the present instance, the prima facie difficulty is far more grave than this. Here it seems that not only are there counterexamples, but that there is not a single positive instance of an action done from principle but without purpose or vice versa. So far from it being an exclusive distinction, we may ask whether we have here any distinction at all. For there seems to be no difference between the questions “What was your purpose in Φ-ing?” and “What principle were you following in Φ-ing?” The same answers satisfy each. “My purpose in such situations is to save the women and children first.”

But more careful scrutiny reveals that there is a difference here. First, consider what is meant by calling someone “unprincipled.” The epithet clearly has condemnatory force. But we are not saying of the condemned that he has no purposes. Nor, even, are we saying that his purposes are bad. But we mean that he observes no limits, no forms, no principles in pursuing whatever purposes he has.

Again, when someone makes something “a matter of principle,” quite typically he is sealing it off against all considerations of results or consequences: he is not willing to forbear, regardless of what purposes are frustrated in the adherence to his principle. The extreme of this principledness is the dictum “Fiat justitia, pereat mundus.”

Now you may say that here he simply has one overriding purpose, and this he calls his principle. A principle is the name we give to a very big purpose. But this is not correct. For the difference between purposes and principles is that whereas the former may be fulfilled, attained, achieved, the latter cannot be, but the latter can be observed, adhered to, honored. This is not merely a verbal but a categorial distinction. It’s not just if you call a purpose a principle then instead of fulfilling it you have to observe it. The point is that purposes are particular whereas principles are universal in character (to use distinction 12 to elucidate distinction 5). It can be my purpose, but not my principle, to fly to the moon or to eat a sour apple. We can date the times when these purposes are fulfilled. But if it is my principle never to treat a man unjustly, then the principle is always operative, does not mean to be so where it has been obeyed or adhered to: there can never be a time when it has been fulfilled: the principle, unlike the purpose, is universal in its reference. (Contrast “If anything is a man, I will not treat it unjustly” with “If anything is a moon, I will fly to it.”)

Of course I can give up my principle, but this is not analogous to arriving on the moon, since I can give up that project before getting there. I’ve followed that principle all my life: if I say I’ve fulfilled that purpose all my life I may be asked—why did you repeat yourself so much? Well, I must mean, I’ve been in the process of fulfilling it all my life. But I am not in the process of following a principle all my life. Purposes aim at realizing particular states of affairs, even if these are of a general and temporally extended sort.32 Principles do not aim at realizing particular states of affairs, because they do not aim at states of affairs, because they do not aim at all: rather they are observed when we are confronted by states of affairs.

Now connected with every principle there is a purpose or purposes and vice versa. If it is my purpose to amass wealth, I should make it a principle to cultivate wealthy friends. (Here we see connection between purposes and hypothetical imperatives.) If I am against drinking on principle, I adopt the purpose of furthering the aims of the temperance league. But this connection between principles and purposes doesn’t damage Kant. He is distinguishing between actions which originate in principles, even though purposes are adopted in cleaving to the principles; and actions governed by purposes, even though principles are adopted to realize the purposes. In fact, Kant’s principle is, “So act …,” and from this he distills the purpose of treating particular human beings on particular occasions as ends-in-se.

Why is Kant so committed against purposes and for principles? The reason is that he identifies having purposes with having desires, as is evident from “the formal principle of duty.”³³ This is why he rejects all teleological ethics, and is the deontologist supreme. But the equivalence between purposes and desires is questionable. A temptation to equating them lies in the double use of “I want x,” which can express either a purpose or a desire. But the way the statement would be verified would be different in each case. For in one sense “I can’t want x” without trying to get it: this is the purpose sense. It isn’t genuinely my purpose unless I pursue it. But it can be my desire even though I don’t try to fulfill it. For I can suppress, restrain my desire: but I can’t do this to my purpose—I can only give up my purpose. My purpose is directly within my control; my desire is not. But to say the former is to sin against the whole teaching of Kant’s ethics. But since he thought purposes were just desires, he had to build the ethics he did.

To summarize: actions either pursue purposes, fulfill desires, by obeying hypothetical imperatives, contingently rooted in our empirical natures: or observe the principle of duty, through an act of will, subscribing to the categorical imperative independent of our empirical natures.

As I suggested, Kant extracts the purpose of moral behavior, expressed in the two formulae of the end-in-itself and the kingdom of ends, from the principle of moral behavior expressed in the universalizability thesis. We are thus enjoined to act in a way which would foster the existence of a kingdom of ends. But it is important to realize just what is involved in “fostering” here. The moral action fosters the kingdom of ends because it is the sort of action which would be performed in it: the more people perform the action, the closer the kingdom of ends is to realization, in the sense that the resulting state of affairs approximates more to what a kingdom of ends is like. But Kant is not certifying actions which would bring about a kingdom of ends but are not themselves instances of the sort of actions undertaken in such a kingdom. By killing, lying, breaking promises, etc., I might help bring about the good society, but I am not permitted to do these things anyway, because they offend against the universalizability principle. There is nothing which falls under his kingdom of ends formula which does not fall under the universalizability formula. So the purpose pursued in moral behavior never takes leave of the principle which generates it.

In his essay “Politics as a Vocation,” Max Weber distinguishes between an ethic of ultimate ends and an ethic of responsibility, and I think this division perfectly renders what Kant favors and what he is against.34 What Weber calls an ethic of ultimate ends I call an ethic of principle. The ethic of ultimate ends is expressed in religious terms thus: “The Christian does rightly and leaves the results to the Lord.” Kant explicitly does this, and brings in the Lord to make sure that, in the long haul, good results issue from moral behavior: this is why we must believe in God. By contrast the follower of an ethic of responsibility feels he must give an account of the foreseeable results of his actions. In an ethic of ultimate ends, “If an action of good intent leads to bad results, then, in the actor’s eyes, not he but the world, or the stupidity of other men, or God’s will who made them thus, is responsible for the evil.”³⁵ The believer in an ethic of ultimate ends feels “responsible only for seeing to it that the flame of pure intentions is not quelched: for example the flame of protesting against the injustice of the social order [without, GAC, reckoning on the consequences, possibly negative, of such protest]. To rekindle the flame ever anew is the purpose of his deeds, which are quite irrational when judged in view of their possible success.”36 Precisely this attitude is the fruit of regarding the good will as the only thing which is good without qualification.

(6) Caused/Free.³⁷ It is easy to see how this distinction connects with those which precede it, at any rate on the left-hand side. Causes are contingent, I am caused, by desire to adopt a hypothetical imperative, and purposes are equated with desires. It is less easy to see why caused and free should be thought exclusive. For Kant seems here to be committing the howler identified by Hume and Ayer: the equation of freedom and causelessness, which turns freedom into chance and caprice.

Whether he is making this particular mistake depends on what we think of the passage “The Concept of Freedom is the Key to explain autonomy of the will.” Here we are told that “the will is a kind of causality belonging to living beings in so far as they are rational, and freedom would be this property of such causality that it can be efficient independent of foreign causes determining it.”³⁸ That’s to say, will is being exercised, in truncated form, even when we obey hypothetical imperatives, the truncation stemming from the fact that the exercise of will is determined by something else, a desire.

Kant goes on to argue that causality involves the notion of law. Thus freedom, though uncaused, cannot be lawless, and therefore is not to be identified with caprice. Free will is subject to law, but not to the law of desire, of particular impulses, but to law as such, universal law. The hatred of the particular, manifest in the Platonic tradition, is here given prime metaphysical status. Not any particular causal connection, but the notion of law as such, which is simply universality, is what guides the will here. But this is only the will guiding itself, in independence of particular desires. So freedom is connected with the will’s autonomy, a formulation of the categorical imperative. (So act “that the will can regard itself as at the same time making a universal law by means of its maxim.”)³⁹ Which brings us to the next distinction.

(7) Heteronomous/Autonomous.⁴⁰ In order to explicate this distinction, we can profitably construct an analogy for it, in the realm of intellect. What would autonomy as opposed to heteronomy of the intellect be like? The philosopher who springs to mind as urging the autonomy of the intellect is Descartes. For his skeptical project consisted in disabusing himself of all deliverances which came from without his mind, and resting his thought on a truth which depended on his mind alone, namely the Cogito. The mind was incapable of denying this truth without denying itself, without denying its own existence. And the force of the proof of the Cogito depended on its clarity and distinctness. So this criterion, wrested out of the intellect itself, was made determinative of truth and falsity generally. Had Descartes passively accepted the opinions he subjected to scrutiny, he would have been deep in intellectual heteronomy.

Hume is the epistemological heteronomist par excellence. For with Hume there is no principle which the intellect, just because it is intellect, must accept, on pain of self-denial. All cognition is dependent, contingent, hypothetical, and even putative categories of cause and substance are only habits imposed on the mind by experience, not order imposed on experience by the mind.

This last remark makes us realize that Kant believed in the autonomy of the intellect as well, in its spontaneous generation of categories. If it fails to apply the categories native to it, it becomes a slave to custom and habit. Here is an interesting instantiation of this claim. If I write down a word or a number, and then perform an operation on it, following rules of grammar or arithmetic, my intellect is functioning actively, autonomously. If I write down a number or word and then put down “whatever comes into my head,” if I don’t direct my thinking, but simply let myself have thoughts, I will end up with a screed the subconscious origins of which can be traced. Just when I think I am freeing myself from rules, I am actually selling myself into cognitive bondage. So, to return to willing, if I do not opt for the law, which expresses the highest in me, I will not be free, but my behavior will be guided by desire. My will will be heteronomous, acted on from without, not fulfilling its true nature.

To be autonomous is to be subject to a law, but a law of our own making, since it expresses what is essential in us, our rationality. To act heteronomously is to act subject to laws not of our own making, laws which are contingent, laws which cannot be legislated or discovered a priori (see 9 and 10).

The doctrine of autonomy teaches that we are subject to the moral law because we are authors of the moral law. The theory is a metaphysicization of Rousseau’s theory of the general will.41 The idea in Rousseau is that my willing is directed both to particular objects and to objects valid generally for all who have will. The former exercise is a derogation from the complete nature of will, assimilating it to desire. This is why when the state coerces me to follow laws embodying the general will, it is forcing me to be free. It is keeping me away from the path of heteronomy. This is why punishment is the right of the criminal in Kant: it expresses his true will in opposition to his heteronomous action.

Kant hovers over the distinction between autonomy and heteronomy.42 But within heteronomy he distinguishes between empirical and rational principles, a distinction which jeopardizes the other sets of distinctions he relies upon. For unless rational principles of heteronomy can be reduced to empirical ones, the structure falls apart. Sometimes he seems to indicate such a reduction: we are attracted, because of our empirical natures, to certain conceptions of reason, not given in experience. But sometimes he insists on a threefold distinction, valid in itself, but out of accord with his other distinctions. For they depend on a rigorous adherence to a disjunction of desire and morality.⁴³

Let us examine closely what he says. If the will passes beyond itself and seeks its law in the character of any of its objects, there always results heteronomy. The will does not give itself the law, but it is given by the object in its relation to the will. This relation, whether it rests on inclination or on conceptions of reason, admits only of hypothetical imperatives: I ought to do something because I wish for something else, whether by immediate inclination or by any satisfaction indirectly gained through reason (here the rational is assimilated to the empirical).

With empirical principles we are already familiar: they are given by desires. Rational principles, we are told, are built either on the rational conception of perfection as a possible effect, or on the will of God. By calling perfection a “rational conception” Kant means that no instance of it can be found in experience: unlike causation, it is a product of reason, not understanding.

Of empirical principles he says that the universality with which the moral law must hold for all rational beings is lost when its foundation is taken from the particular constitution or circumstances of human nature. The ontological conception of perfection is rejected because it is question-begging (does it include moral perfection? which of course we are urged to aim at—but then it is just another name for the moral law). The will of God must either be judged to be good, which again is circular, or, if not, then this is a most heinous principle to follow, mere power-worship. I agree with both of these objections: but I cannot see how adopting either perfection or the divine will as a standard involves assent to a hypothetical imperative.

The important thing to remember is that both autonomy and freedom are connected with universality: Kant calls them reciprocal concepts.44 So, in part 3, where he proves freedom, he takes himself to be proving the moral law.

(8) Selfish/Selfless

This distinction is the least important of the twelve. I throw it in, as it were, as a bonus. It simply adds to the series of conflations I listed when discussing distinction 6. By then it was clear that, as an empirically examinable being, all my behavior was subject to causation by desire, including my purposive behavior, given the assimilation of pursuing purposes and having desires. I am also a selfish being, since the desires which direct me are not yours but mine. Even if I seek to satisfy you, or more grandly, to realize the general welfare, it is only because I desire this end that I strive to bring it about:

Out of love of humanity I am willing to allow that most of our actions may accord with duty; but if we look more closely at our scheming and striving, we everywhere come across the dear self, which is always turning up; and it is on this that the purpose of our actions is based—not on the strict command of duty which would often require self-denial.⁴⁵

So not only is all my behavior governed by desire, but all my desires, viewed carefully, are selfish. This is a new claim, resting perhaps on a failure to distinguish between actions which afford me satisfaction and actions undertaken to afford me satisfaction. I may glow with content after sacrificing myself for you: but I don’t necessarily do so in order to feel rosy. In fact, I may make the sacrifice even if I don’t want to: but even if I do want to (and Kant would insist that I do) what I want is to help you, not to indirectly help myself.

There is another sense in which behavior is selfish. It is so because it follows from particular aspects of my nature, those aspects which make me specific, the specified, person that I am. This is contrasted with moral action, flowing as it does purely from what is universal, and in this sense selfless, about me, my rationality. My true self is that which transcends my self.

Thus there is no contradiction between supporting autonomy and rejecting springs of action which come from my self. It is not my empirical self which matters, but that which I have in common with any rational being, namely my reason, of which will is the practical exercise.

(9) and (10) Empirical/A Priori46

1. These two distinctions have to be treated together. The point is that Kant thinks he has made one distinction here, whereas actually there are two, however much the required equivalence relations may obtain between them.

2. The essential contention, embodied in the distinction, is that you cannot show the actuality of morality empirically, and so it is necessary to show it a priori (if it can be shown at all). But this vague phrase “actuality of morality” sometimes signifies “the objective validity of the moral law” and at other times “the existence of moral behavior.” Clearly the moral law could be objectively valid without there being moral behavior. And conversely, moral behavior might exist, according to the prephilosophic as yet ungrounded but later shown to be sound conception of the moral law,⁴⁷ even if the latter were not shown to be sound. But let me show, textually, how two claims are being made, without being clearly differentiated.⁴⁸

3. First let me draw your attention to the section “The need for pure ethics.”⁴⁹ Here Kant asserts that it is universally admitted that for a law to have moral force it must carry with it absolute necessity: thus, e.g., “Thou shalt not lie” is not valid for men alone, but for all rational beings. From this it follows that the basis of obligation cannot be sought in man’s nature and circumstances but a priori in the conceptions of pure reason. This is the nonempirical character of morality in the first sense. Before indicating the other sense (actuality of moral behavior), two objections are in order. First, what Kant takes to be obvious is hardly accepted by every plain man or even by every philosopher. Second, even if it is true that morality must be valid for all rational beings, it doesn’t follow that its precepts are a function of their rationality, or that they can only be discovered a priori.

4. At any rate, Kant is saying in the above that our notion of duty is not an empirical one. In the section entitled “The use of examples” he says this again, but attaches a totally different significance to it.50 For here he says that we can never be certain in experience whether a given action flows from pure duty. (We shall not labor the point that we can, in experience, always be certain that it doesn’t, since duty must be freely willed, and empirically viewed, action is determined.) To return, he says that we can never, even by the strictest examination, get completely behind the secret springs of action: since when the question is of moral worth, it is not with the actions which we see that we are concerned, but with those inward principles of them which we do not see. Now by “inward principles” Kant either means the phenomena of inner sense, in which case he here contradicts his avowed faith in the subsumability of these under discoverable laws; or he means noumenal activity, in which case, though we indeed discover a priori whether or not duty is being done, the answer is always in the affirmative. In fact I suggest Kant is cheating, confusing the two.

5. Anyhow, in the immediately following section (“Popular philosophy”)⁵¹ he says something in flagrant contradiction with the above. For he tells us that the pure conception of duty, the conception of a moral law, exercises on the human heart an influence so much more powerful than all other springs which may be derived from the field of experience, that in the consciousness of its worth, it despises the latter, and can by degrees become their master. How, one may ask, can such a phenomenological history be backed except by appeal to empirical data about human beings, an appeal which is supposed to be useless according to the previous section? How can you know how strong the sense of duty is in experience without looking at experience? In fact, the confusion stems from illegitimately looking at man as a compound of empirical and a priori parts, in tension, and saying that the latter element is destined to preponderate. But to transform standpoints, ways of looking at the same thing, into two parts of a single thing, is to commit a very juicy category howler indeed.

6. That Kant is assimilating two distinctions, a criticism separate from those I level at the way he treats each separately, can best be seen⁵² by attending to the words of Paton, his chief English apologist. In The Categorical Imperative the wily Scot has this to say: “Kant … insists that we can in no way establish the categorical imperative by an appeal to experience. He knows too well that seemingly categorical imperatives may conceal a motive of personal interest.”53 Now it is hard to see how a categorical imperative could do any such thing. How can a formula have a motive in it? What is easy to see is how action seemingly obeying a categorical imperative is actually motivated by a personal interest, is actually hypothetically inspired. But whether actions follow categorical imperatives is a question separate from whether there is a categorical imperative. And whereas one can at least plausibly maintain that the latter is not an empirical question, it is hard to see how the former can be anything but an empirical question.

Upshot: you can never be sure someone is moral. So the basis of morality must be a priori.

7. The fundamental assimilation at the basis of this confusion is Kant’s conflation of psychological hedonism and philosophical hedonism.⁵⁴ He wants to reject both and thinks that the latter is implicated in the former: in his phenomenal account of human nature he commits the naturalistic fallacy. So he has to deny both that motivation can be established empirically and that value can be established empirically, but he never sees the difference between the two.⁵⁵

APPENDIX

Kant is making several claims, all of which he subsumes under this distinction, and some of which contradict others.

(1) You cannot argue from an “is” (quite generally—the assertion need not be about somebody’s behavior; it might be about human needs, or about conflictless societies) to an “ought.”

(2) You cannot argue from a genuine example of moral behavior to the s.m.p.,⁵⁶ without a petitio.

(3) You can never be sure you are dealing with a genuine example.

(4) All empirical springs are desires—and these can’t generate morality in a secure way (sometimes desire may lead to wrongdoing. But actually Kant has in mind a conceptual claim, not a probabilistic one).

(11) Phenomenal/Noumenal57

There remains little further to say about this distinction, which I have referred to in explicating the previous two, and discussed at length in the opening lectures. Its function is to guarantee the remaining distinctions, by showing us a point of view from which the right-hand items are possible.⁵⁸ But it defeats itself, by showing that the left-hand items are illusory. In order for it not to be illusory, noumenon and phenomenon would have to be not standpoints, but parts, a way in which Kant sometimes illicitly treats them. He also sometimes (e.g., “the two standpoints”)⁵⁹ treats them as a question of double membership: I belong to both worlds. But there are not two worlds but one world, conceived as it is in itself on the one hand and only as it appears on the other. In the section at hand he says: “So far as he belongs to the world of sense he finds himself subject to laws of nature (heteronomy); but as belonging to the intelligible world he finds himself under laws which being independent on nature have their foundation not in experience but in reason alone.”⁶⁰ But he is not allowed to say this, for he makes it look as though I am to opt between sense and reason, whereas the option is not practical, actional, but epistemic.

Descartes, who made body and soul separate substances, found it difficult to account for their apparent interaction. Kant is more Spinozistic, seeing noumenon and phenomenon as modes of existence. It must be asked (1) How can two aspects of a thing interact? (2) Isn’t their interaction necessary for there to be a will-inclination battle?⁶¹ (3) Aren’t the inclinations the phenomenal appearance of noumenal activity? (4) How can man qua noumenon will wrongly? Is he not qua noumenon a holy will? Isn’t it illicit to say he has a moral will and has inclinations? (5) Oddly, Kant says (in second of “the two standpoints”) that man is really a phenomenon in the world of sense as well:⁶² this oddness is necessary to give the inclinations power. (Really, they’re not on a par.)

But let us suppose that positing noumena behind phenomenal men is not a self-defeating move, and question whether the positing itself is legitimate anyway. Here I refer you to the Critique of Pure Reason:

If we entitle certain objects, as appearances, sensible entities (phenomena), then since we thus distinguish the mode in which we intuit them from the nature that belongs to them in themselves, it is implied in this distinction that we place the latter, considered in their own nature, although we do not so intuit them, or that we place other possible things, which are not objects of our senses but are thought as objects merely through the understanding [remember early lecture—we can think the noumenon, though we cannot think about it], in opposition to the former, and that in so doing we entitle them intelligible entities (noumena). The question then arises whether our pure concepts of the understanding have meaning in respect of these latter, and so can be a way of knowing them.63

First, note the reference to “other possible things,” rather than other ways of looking at the same things, or conceiving them. This is the point I’ve been hammering at all too long, and which I now turn away from.

Now take the last sentence. We know that Kant’s answer to this is in the negative, and that proving this answer is what the transcendental dialectic is all about. But since the answer is in the negative, it seems invalid for him to apply to the noumenon the category of plurality, which is the second category of quantity.⁶⁴ That is, he can have no ground for asserting a plurality of noumenal persons, but he must if I am to follow the dictates of my real self. But even if we could permit the use of this category, even if we could allow that the noumenon can be individuated into particulars, it seems unwarranted to make this individuation parasitic on the phenomenal individuation of particular people, which is what he implicitly does. Nor can he say of my noumenal being that it is a substance, as he must if I am to be held responsible for what I do. And finally, though this is an objection to the whole distinction, not just to the use he makes of it in morals, how can noumenal activity cause phenomenal appearances? He only evades this issue by saying, obscurely, that it grounds them. True, he says, we can’t see how it causes. But he is sure that it causes; yet he can’t make sense of this.

I am not suggesting that there is no value in his two standpoints notion. It is beautifully developed by Hampshire, and by D. M. MacKay. And stupidly by Strawson, in Pears.⁶⁵ I am only saying that it is impossible to see how it can lead to saying that I am obliged to act according to the moral law.

(12) Particular/Universal

This last distinction should by now be sufficiently familiar to warrant no further comment. But to understand it better, I refer you to the section “The categorical imperative.”66 Throughout we have stressed Kant’s rejection of particular objects of particular desires. He takes the categorical imperative to be the sole alternative to the former. As the will has been deprived of all particular impulses, there remains nothing but the conformity of its actions to law in general, which alone is to serve the will as a principle, i.e. I am never to act otherwise than so that I could also will that my maxim should become a universal law.

Action is heteronomous when it is an instance of a particular empirical law of desire. The alternative to this is when it is an instance of law-abidingness in general. Distrustful as we are of Kant’s assimilations, we might suppose that not all particular maxims are tied to desires, so that one could be moral, in the sense of not merely pursuing interest, without adopting the categorical imperative. In morality we obey a law just because it is a law, without any interest, but it doesn’t follow from this that the law we obey is the categorical imperative; it may be quite particular, like: help the needy, which we may not feel called upon to derive from the notion of law-as-such. In sum, acting out of respect for law as such is not coterminous with acting out of respect for the law, with acting universalizably, and only universalizably.

SUMMARY

The burden of the distinctions can be put succinctly in this manner: the motivation of duty must be distinct from the motivation of desire; but for action to be motivated by desire (1) is for it to be caused (2), to be infected with particular (4) purpose (3), to be heteronomous (5), to be contingently (6) grounded, selfish (7), and an obeying of a hypothetical imperative (2); so duty must be inspired by universal principle, autonomous, categorical, selfless, and free; in fact men must treat themselves as free; so they must treat themselves as under the law; and they are entitled so to treat themselves because they are not aware of being otherwise, noumenally speaking. My account has focused on the genesis of the doctrine, not, as is more usual, on its applicability (of universalizability, etc.).

The latent pathology is the uncompromising association of freedom and the moral law. This association also poisons existentialism, in precisely the reverse manner. Here it’s not that we’re free because there is a law: we’re free because there isn’t. But “we are free to choose what we value” is ambiguous. It is the basis of the philosophy of total responsibility.

When discussing purpose/principle, I referred to Kant’s doctrine that the only thing good simpliciter is the good will. And I said that this was unlike Mill. The point is: goodness is not derivable from purpose (purpose, being particular, is associated with desire). In sum, what he wills, not what he wants, decides a man’s worth. And if he wills a particular purpose, not principle as such, then he wills heteronomously.67

APPENDIX: SUPPLEMENT⁶⁸

DISTINCTION 7

We are concerned now with Kant’s seventh distinction, according to which, in every action, the will is either autonomous or heteronomous. “Auto” means “self”; “hetero” means “other”; “nomous” comes from “nomos,” which means law. So the difference between autonomy and heteronomy is one between action according to the will’s own law or rule or principle and action according to some principle which is other than the principle governing the will, and is derived from some object external to the will.

I began to try to make sense of this difficult doctrine by recommending that we can understand Descartes’s project in the Meditations as a search for the autonomy of the intellect. The intellect is concerned with discovering truth; the will is concerned with doing what is right. Kant is saying that a will wills rightly only when it wills according to the principle which defines its own nature—what this principle is and why it defines the will’s nature I shall explain in a moment. So in the analogy the intellect would discover truth only when it thinks in accordance with its own principles, and does not accept any principles drawn from a source outside the intellect. Typical such sources for Descartes are sense-perception or traditional lore. These are banished, and then we try to see what the intellect can confidently assert without any recourse to them. And it turns out that the only thing which the intellect can assert without recourse to any external evidence is the Cogito. And what seems to guarantee the Cogito is its clarity and distinctness. So this principle, integral to the functioning of the intellect itself, is the criterion of what can be asserted to be true in all cases: when and only when the intellect alone is functioning, we get clear and distinct ideas.

So to sustain the analogy, we must find a principle which defines the nature of the will in the way that clarity and distinctness define the nature of the intellect, and then, to will rightly, the will must follow this its own principle, just as, to think truly, the intellect must follow its own principle. And just as we determine in the Cartesian case what the principle is by asking what guides the intellect when it is acting without the influence of nonintellectual sources on it, we must ask what guides the will when no sources external to it are acting on it. And then we shall know what it is for the will to act according to its own nature.

Now from distinction 3 we know that an action is guided by and only by the will if and only if the will is not guided by any desire in the action it selects. Thus if the will is guided by desire, it is behaving heteronomously. And if it is not guided by desire, it is acting autonomously. But to be guided by desire, we know from distinctions 1 and 6, is for the subject’s action to be subjected to some contingent law. But Kant thinks that all actions must take place in accordance with some law. So the will’s action must be in accordance with some law but not in accordance with some contingent law, if the will is to be autonomous. But then all that is left is for the will to act in accordance with the notion of law in general; that is to say, the will is autonomous only when it causes an action which could always be willed, by any rational being, by any will, regardless of its desires, in the situation in which the will finds itself. In other words: whenever there is an action, there must be some law which it satisfies. Autonomous action cannot satisfy a particular law, so it must satisfy the law “Act so that … rational beings,” since any other law is particular.

So to act according to the universalizability principle is to act according to a principle which is the only principle that can govern the will when nothing outside the will is moving it, and so to act according to the universalizability principle is to act autonomously. It is for the will to be subject to a law, but to a law of its own making (this has a pedagogically valuable ambiguity), unlike laws which are contingent on facts about the will’s possessor which are independent of the fact that he has a will.

Now the will acts according to principles of heteronomy if its action is spurred not by its own principle but by some object external to it. Now that object may be of two kinds, and so there may be two kinds of principles of heteronomy. The object may be an empirical one, that is, one that we encounter in our experience, and the usual one provided by empirically heteronomous ethics is happiness. Or the object may not be found in experience but may be devised by reason, and here the usual one provided is the idea of perfection, whether as something to be achieved—so that I act in such a way that I shall become more perfect—or as something thought to be actual, namely a perfect being, namely God, to whose dictates I submit my will. But in all three cases, the will is bound to its object by desire, the desire to realize happiness, not necessarily my own, or the desire to become perfect, or the desire to follow God’s will. In all three cases, so Kant claims, we can obey only a hypothetical imperative: if you will happiness, then you must will this; if you will to be perfect, then you must will this; if you will to follow God’s dictates, then you must will this. But the if-clauses of hypothetical imperatives depend on contingent facts about the beings who obey them: thus, in these three heteronomous moralities, we are acting according to contingent, empirical laws, not valid for any rational beings, and hence not moral. The rational principles of heteronomy do not indeed direct us to seek items we have encountered empirically, but they are heteronomous principles because it is only a contingent fact about human beings that by their reason they devise the ideas of perfection and it is a contingent fact about them that they are attracted to these ideas.

Now in addition to rejecting the rational principles of heteronomy, on the grounds that they are heteronomous, Kant has special arguments against any ethic which recommends them. He claims that such ethics must be either circular or inadequate. And their rejection on this basis does not depend on their rejection on the ground that they are heteronomous. The rejection on the ground that they are heteronomous depends on Kant’s complex argument that only an autonomous will can be a moral one. The argument that they are circular or inadequate does not depend on that thesis. They are circular if moral perfection is built into perfecting myself, or into the notion of God. For then I am not in receipt of a satisfactory answer when I ask what it is to be moral, for to say that it is to aim at perfection will involve saying that it is to aim at being moral, and to say that it is to follow the perfect will of God will involve saying that it is to follow a moral will, and each of these answers is inadequate. We asked the question to find out what morality is, and the answers presuppose that we know what it is. But if the answers are not circular, they must be inadequate, for why should I think myself duty-bound to perfect myself where my perfection does not involve moral perfection, but, say, the development of my strength and my talents alone? And why should I think myself duty-bound to obey a God who may not be moral, who may indeed be merely infinitely mighty?

So to build an ethic on seeking happiness, on perfecting oneself, or on obeying God is in each case to adopt heteronomous principles, since they are not derived simply and solely from the nature of the will itself: the will is acting according to a principle outside itself, and this is fatal to acting morally just as thinking according to what sense-experience and tradition suggest is fatal to thinking truly. And in the cases of rational principles, the doctrines are anyway either circular or uninviting when examined carefully.

So if there is such a thing as morality, it is realized when the will is subject to a law of its own making, since it is the only law defining the nature of the will itself, and it is the latter because it is the only law a will can follow when not guided from without. But, Kant will later argue, it is clear that to be autonomous, it is to be free. So in the final section of the book he tries to prove that we must believe ourselves to be free. Then he can say that we must take ourselves to be subject to the moral law.

DISTINCTIONS 9 AND 10

Here I want to claim that Kant is making two distinctions, while he thinks he is making one. He is putting forward two theses about what must be discovered a priori, or nonempirically, and the two things are different. I shall first explain what these two things are, and how they do differ; then I shall bring evidence to show that Kant in fact confuses them. Then I shall offer a comment on each thesis separately.

I shall begin with an ambiguous statement. Its two interpretations are the two theses which Kant is confusing. The statement: Morality cannot be discovered empirically, so if there is such a thing as morality, it must be discovered a priori. But this statement can mean one of two things, and Kant uses it to mean both: one thing it can mean is that “The moral law cannot be discovered empirically” and the other is that “Whether someone is behaving morally cannot be discovered empirically.” Alternatively, the first is, “Assembling observable facts will not enable you to prove the validity of the moral law”—you cannot move from an “is” to an “ought,” and “assembling observable facts will not enable you to decide whether someone is behaving morally.” Now the objective validity of the moral law might have to be shown a priori, but it would not follow that we would not tell empirically whether anyone is following it. Thus we show a priori that 2 + 2 = 4, but we can tell empirically whether someone is respecting this truth. Conversely, we might be able to give some sort of utilitarian and hence empirical basis to morality but still be unable to tell whether someone is being moral.

To put this differently. There obviously could be a moral law without there being moral behavior. And there could be moral behavior in the prephilosophic sense without there being a moral law; that is, we might be able to show that if there is a moral law it is of such and such a character, while failing to show that there is such a law, but succeeding in showing that there is behavior which would be moral if there were such a law.

Now let me point out where Kant asserts each thesis about the a priori character of morality, and then where he manifestly confuses the two theses. First look at the following:

Every one must admit that a law has to carry with it absolute necessity if it is to be valid morally—valid, that is, as a ground of obligation; that the command “Thou shalt not lie” could not hold merely for men, other rational beings having no obligation to abide by it—and similarly with all other genuine moral laws; that here consequently the ground of obligation must be looked for, not in the nature of man … but solely a priori in the concepts of pure reason.69

Here he is claiming that we cannot know that there is a moral law or what it is by consulting experience or facts of observation. Now look at the following:

In actual fact it is absolutely impossible for experience to establish with complete certainty a single case in which the maxim of an action in other respects right has rested solely on moral grounds and on the thought of one’s duty. It is indeed at times the case that after the keenest self-examination we find nothing that without the moral motive of duty could have been strong enough to move us to this or that good action and to so great a sacrifice; but we cannot infer from this with certainty that it is not some secret impulse of self-love which has actually, under the mere show of the Idea of duty, been the cause genuinely determining our will. We are pleased to flatter ourselves with the false claim to a nobler motive, but in fact we can never, even by the most strenuous self-examination, get to the bottom of our secret impulsions; for when moral value is in question, we are concerned, not with the actions which we see, but with their inner principles, which we cannot see.⁷⁰

Here he is saying, not what he said before, that we cannot base the moral law on observable facts, but something quite different: that we cannot by the keenest observation be certain that a moral action has been performed.

Now why do I say that Kant confuses these separate things, both of which he thinks we cannot discover empirically? It is because in a number of places, for instance in the following, he passes from the premise that one can never be sure that someone is behaving morally to the conclusion that if there is a moral law it must be discovered or proved a priori:

Beyond all doubt, the question “How is the imperative of morality possible?” is the only one in need of a solution; for it is in no way hypothetical, and consequently we cannot base the objective necessity which it affirms on any presupposition, as we can with hypothetical imperatives. Only we must never forget here that it is impossible to settle by an example, and so empirically, whether there is any imperative of this kind at all: we must rather suspect that all imperatives which seem to be categorical may none the less be covertly hypothetical.71

Let me quote from Paton, The Categorical Imperative: “Kant … insists that we can in no way establish the categorical imperative by an appeal to experience. He knows too well that seemingly categorical imperatives may conceal a motive of personal interest.”⁷² But my point is that the only thing that can conceal a nonmoral motive is a case where someone assents to or appears to obey a categorical imperative. Someone who seems to be acting morally may be acting out of personal interest. But whether an imperative is categorical has nothing to do with whether a particular person is truly obeying a categorical imperative. The question whether or not there is a categorical imperative is quite independent of the question whether or not anyone ever genuinely obeys one.

Finally, I want to make a point about Kant’s thesis that we cannot through experience establish that someone is behaving morally. He constantly speaks as though experience leaves us uncertain whether anyone ever behaves morally. But in fact his philosophical position commits him to saying something stronger than this. He is committed to saying that insofar as we restrict ourselves to what can be observed in experience, we must always conclude about a piece of putatively moral behavior that it is not in fact moral. For Kant thinks that every item in our experience must be represented by us as having a cause discernible in our experience. This is the teaching of The Critique of Pure Reason, and it is explicitly extended to cover human behavior, including its overt and covert manifestations.

Now as we shall see shortly, Kant thinks that we can adopt another perspective on the world from the observational one, and hence another perspective on human behavior, which is part of the world. From that perspective, he will urge that it is possible to say that we behave morally. (I shall show, however, that, from that perspective, we really always must behave morally, so that Kant proves too much.) But he is not entitled to say that he thereby answers in the affirmative a question which experience is unable to answer. He is rather answering in the affirmative a question which, for Kant, experience must answer in the negative. And this is obviously a less tenable situation than the one he hopes he can be in. In fact, as we shall see, he illicitly transforms two ways of looking at human behavior into two forces operating in human behavior, and he makes the noumenal force occupy the place which he illicitly says we cannot know what occupies it from experience.

DISTINCTION 1173

This eleventh distinction is most important. The burden of it is that there is a nonempirical perspective which we can adopt on human beings, and from the view of that perspective the right-hand items, which exist if and only if morality exists, can be asserted to exist. For every action which we examine empirically must turn out to have all and only left-hand side features, and is hence, thus viewed not moral.

I divide my discussion of this important topic into six parts: (1) How Kant establishes the phenomenon/noumenon distinction; (2) Illegitimate ways in which he uses it; (3) Kant’s practical proof of freedom; (4) Self-defeating character of appeal to noumenon; (5) How Kant avoids the problem that appeal is self-defeating by treating aspects as parts, and some other difficulties; and (6) Conclusion.

(1)⁷⁴

One observation is possible without any need for subtle reflexion and, we may assume, can be made by the most ordinary intelligence … the observation is this—that all ideas coming to us apart from our own volition (as do those of the senses) enable us to know objects only as they affect ourselves: what they may be in themselves remains unknown. Consequently, ideas of this kind, even with the greatest effort of attention and clarification brought to bear by understanding, serve only for knowledge of appearances, never of things in themselves. … it follows that behind appearances we must admit and assume something else which is not appearance—namely, things in themselves—although since we can never be acquainted with these, but only with the way in which they affect us, we must resign ourselves to the fact that we can never get any nearer to them and can never know what they are in themselves. This must yield us a distinction, however rough, between the sensible world and the intelligible world, the first of which can vary a great deal according to differences of sensibility in sundry observers while the second, which is its ground, always remains the same.75

(2) Now in The Critique of Pure Reason, where this doctrine is expounded at much greater length,⁷⁶ Kant strenuously insists that we cannot apply what he calls the pure concepts of the understanding to what is noumenal. These are the concepts which are not gained through experience but are used by us when we organize experience. Three of these concepts are plurality, cause, and substance.⁷⁷ Thus we are not entitled to speak of a number of things in themselves, or to speak of causal transactions between things in themselves, or between things themselves and appearances, nor may we apply the concept substance to the noumenal, that is, we may not speak of enduring particular things within the noumenal. It’s not merely that we can’t answer questions like, how many noumenal things are there? Kant thinks he proves that the question cannot even be asked, that plurality cannot be a feature of the world as it is in itself. He thinks he shows that if you ask such a question you can prove contradictory answers to it (in the Transcendental Dialectic).

So an essential feature of the noumenal is that we cannot apply our categories to it. Yet this is precisely what Kant does in the Grundlegung. This is plain from the rest of the paragraph which I quoted earlier:

Even as regards himself—so far as man is acquainted with himself by inner sensation—he cannot claim to know what he is in himself. For since he does not, so to say, make himself, and since he acquires his concept of self not a priori but empirically, it is natural that even about himself he should get information through sense—that is through inner sense—and consequently only through the mere appearance of his own nature and through the way in which his consciousness is affected. Yet beyond this character of himself as a subject made up, as it is, of mere appearances he must suppose there to be something else which is its ground—namely his Ego as this may be constituted in itself: and thus as regards mere perception and the capacity for receiving sensations he must count himself as belonging to the sensible world, but as regards whatever there may be in him of pure activity (whatever comes into consciousness, not through affection of the senses, but immediately) he must count himself as belonging to the intellectual world, of which, however, he knows nothing further.⁷⁸

Notice here that Kant has identified within the noumenal world a noumenal person to correspond to every phenomenal person. But to do this is to chop the noumenal world up into bits, and that is what we are forbidden to do. And even if we are not forbidden to chop the noumenal world into bits, to individuate particulars within it, what is the justification for making the individuation parasitic upon the individuation of particulars in experience? Of course we know why Kant does this. He must do it, for he’s trying to prove that those beings who seem locked up by desire may in fact be free to behave morally. So they must, each of them separately, be members of the noumenal world. But his doctrine of the noumenal world forbids him even to entertain the idea of its having separate members. And even if that idea could be entertained, it is arbitrary to separate them on the basis of phenomenal divisions, and it is not only arbitrary but contrary to everything Kant says re noumenon, since we thereby know things about the noumenon, e.g. that it has so many human beings in it. Again, if those beings are to be capable of moral responsibility, which is precisely what Kant requires, then they must presumably be substances which endure through time so that past acts of theirs can be assigned to them in the present. But this talk is also quite illicit.

But in the rest of what I have to say, I ignore these criticisms. Let us pretend that Kant can say that each human agent is a noumenal being in reality, and only has desires, etc., in appearance. Let us see what he proposes to do with this conception which we have shown to be insupportable anyway.

(3) You will remember from discussion of autonomy that being moral turned out to be to will autonomously, and willing autonomously was the same as willing freely. Thus if Kant can prove that we are free, he can prove that we are subject to the moral law. Or so he thinks. It is plain that what freedom will prove is not only that we are subject to the moral law, but that we are always obeying it, because of the identities just asserted. This I show next. Then I show how Kant twists the phenomenal/noumenal distinction to avoid this unhappy conclusion. Right now let’s see how he proves freedom. Here is the relevant text (with sentence numbers added for reference):

(1) Now I assert that every being who cannot act except under the Idea of freedom is by this alone—from a practical point of view—really free; that is to say, for him all the laws inseparably bound up with freedom are valid just as much as if his will could be pronounced free in itself on grounds valid for theoretical philosophy. (2) And I maintain that to every rational being possessed of a will we must also lend the Idea of freedom as the only one under which he can act. (3) For in such a being we conceive a reason which is practical—that is, which exercises causality in regard to its objects. (4) But we cannot possibly conceive of a reason as being consciously directed from outside in regard to its judgements; for in that case the subject would attribute the determination of his power of judgement, not to his reason, but to an impulsion. (5) Reason must look upon itself as the author of its own principles independently of alien influences. (6) Therefore as practical reason, or as the will of a rational being, it must be regarded by itself as free; that is, the will of a rational being can be a will of his own only under the Idea of freedom, and such a will must therefore—from a practical point of view—be attributed to all rational beings.79

Sentence 1: this is OK. If to be free is to be under the moral law, then if we must take ourselves to be free, we must take ourselves to have moral obligations.

Analogy: suppose you are a doctor treating a number of patients with a serum, but you can’t know whether your current supplies of serum will be sufficient. If you assume they will be and they aren’t, you have disaster. Then you are practically required to assume it won’t. But then you must also assume all the consequences as much as if you knew them. (This is only partly analogous—for here it is silly to assume otherwise, whereas in Kant’s example, it is inconceivable to assume otherwise.)

2: this is what he is trying to prove.

3: this is OK, this just means we have a will: a will is a causer of actions, even when it is moved by desire. (Notice that here will is practical reason: see early lecture.) So this doesn’t beg the question, as it might seem to.

4 and 5: this I have discussed elsewhere.⁸⁰ I think it is true but with qualifications as I explain.

6. Contrary to what Paton says in his note, this is quite plainly an inference. Else why the “therefore”? And here I again refer you to the earliest part of these lectures on Kant, where I said that Kant wanted the will to be not separate from reason but a special exercise of it, because otherwise his proof of freedom of the will would not work. The proof is based on the freedom of the organ used in willing, and its freedom is established from its use in thinking. I leave you to assess the validity of the inference.

Let us suppose that this succeeds, that Kant has offered a practical proof of freedom. We have to take ourselves to be free. Now he immediately recognizes that this conflicts with what we know ourselves from experience. So he shows that we can consider ourselves from another point of view, and we have already traced how he does this. Now we know nothing about ourselves from that point of view, he says (though we have seen that we must know something). But since we know nothing, we are entitled to make suppositions, if we have reason to. And we have a practical proof of freedom, which experience contradicts, and which therefore we are entitled to suppose fulfilled for what we really are.

(4) Now why do I say that making this distinction defeats Kant’s purpose? His purpose is to show that though we appear to be able to act morally and hence not to be subject to the moral law, we are nevertheless subject to the moral law, we must nevertheless try to be moral, and we nevertheless, despite what experience tends to show, can succeed. And he says we realize we can be moral when we think of ourselves as members of a noumenal world, for as such members we are free. But this is the only world of which we are really members; therefore we must think of ourselves as really free without qualification—but being moral was, through the formula of autonomy, identified with being free; therefore we are moral, and we have nothing to worry about. No matter what we seem to be doing, we are behaving morally, since every phenomenal appearance is an appearance of noumenal activity, and all noumenal activity is free, and therefore moral.

(5) Kant escapes this disastrous consequence of his doctrine by grossest cheating, in two steps. For he gradually transforms these two standpoints we can take on ourselves, from one which is valid and another which is invalid, to two equally valid ones. And finally, he speaks as though we belong to two worlds and then as though we have a noumenal part and a phenomenal part which are in tension, and we behave morally when we favor the noumenal part of ourselves. What he is really doing is making what I appear to be into a part of what I am.

This is manifest from many passages.81

What he’s trying to do is to produce a sophisticated philosophical doctrine which will carry the burden of a Christian spirit/flesh conflict. But there can be no conflict if the flesh is merely apparent and there can be no condemnation of what appears fleshly if what appears fleshly is really spiritual.⁸²

(6) On the identification of freedom and morality. What Kant is saying. What Sartre says.

12TH DISTINCTION: WHAT KANT’S UNIVERSALIZABITY TEST IS NOT

(1) Not rule utilitarianism, for its test (if everybody behaved so, consequences would be bad) presupposes independent notion of what is good. Kant is not saying the universalized state of affairs must not be unwillable.

(2) Not Hareianism, for his test is whether you think it right for others. Kant whether you can will others. Also first way of going wrong (logical one) absent in Hare.

(3) Not Golden Rule (a) The Golden Rule is positive—used to generate maxims, not test them. (b) The Golden Rule also eliminates first way of going wrong.

The two ways of going wrong: (a) Logical—associated with perfect duty—transgressions, datable and victims specifiable (promises, lies, breach of contract, severe injury (exception)). (b) Psychological—associated with imperfect duty—not datable—not specifiable. Latitude of choice is given. (Benevolence, due consideration (exception).)

Problems:

(1) What is the right description? The one under which the man acts: But then sincerity tests are required.

(2) Suppose a debt is onerous. Creditor wants to be paid, debtor to escape payment. If each universalizes, conflict remains. (Debtor proposes to pay, creditor to let him off.) Whose perspective rules? Utilitarian solution possible, but it is not a Kantian one.

(3) How can I will that anyone occupy this space? Maxim must be general.

(4) Billy Budd type case.83

1 Kant, Critique of Pure Reason Bxxx, p. 29.

2 Kant, Groundwork of the Metaphysic of Morals (391), p. 59, p. 57. [Cohen has used H. J. Paton’s translation from 1947, as revised in 1958, in the Hutchinson hardback edition, reprinted in 1963. The pagination differs from the later Hutchinson paperback version. References are given first, in parentheses, to the standard Academy pagination, then to the hardback, then to the paperback.—Ed.]

3 See Frankena, “Obligation and Motivation in Recent Moral Philosophy,” for an account of recent views of the subject.

4 [The text at this point includes the comment “(Sorensonization of Kant).” It could possibly be a reference to Søren Kierkegaard.—Ed.]

5 Kant, Groundwork of the Metaphysic of Morals (447–48), pp. 115–16, pp. 109–10.

6 [Cohen has quoted Abbott’s translation, Kant, Fundamental Principles of the Metaphysics of Ethics, p. 81. Paton’s translation reads: “But we cannot possibly conceive of a reason as being consciously directed from outside in regard to its judgements; for in that case the subject would attribute the determination of his power of judgement, not to his reason, but to an impulsion.” Kant, Groundwork of the Metaphysic of Morals (448), p. 116, p. 109.—Ed.]

7 Compare writing down numbers by association and writing down numbers according to a formula.

8 [As becomes clear from a citation below, Cohen is here discussing Peters’s The Concept of Motivation.—Ed.]

9 [At this point the text contains the comment: “(REPHRASE: TWO POINTS ARE BEING MADE—ONE ABOUT DESCRIPTION, ONE ABOUT EXPLANATION (Does ‘Man is a rule-following animal’ explain?).)”—Ed.]

10 Peters, The Concept of Motivation, pp. 16–17.

11 Kant, Groundwork of the Metaphysic of Morals (395), pp. 62–63, pp. 60–61.

12 In the section “classification of imperatives,” ibid. (414), p. 83, p. 78.

13 [The typescript at this point uses the term “cs-p.” It is not entirely clear what is meant, but it has been assumed that it is an abbreviation for “consciousness.”—Ed.]

14 Kant, Groundwork of the Metaphysic of Morals (396), p. 64, pp. 61–62.

15 Ibid. (417–20), pp. 84–88, pp. 80–83.

16 Hobbes writes:

The Felicity of this life consisteth not in the repose of a mind satisfied. For there is no such Finis ultimus (utmost ayme) nor Summum Bonum (greatest Good,) as is spoken of in the books of the old Morall Philosophers. Nor can a man any more live, whose Desires are at an end, than he, whose Senses and Imagination are at a stand. Felicity is a continuall progresse of the desire, from one object to another; the attaining of the former, being still but the way to the latter. The cause whereof is, That the object of mans desires, is not to enjoy once onely, and for one instant of time; but to assure for ever, the way of his future desire. And therefore the voluntary actions, and inclinations of all men, tend, not only to the procuring, but also to the assuring of a contented life; and differ onely in the way: which ariseth partly from the diversity of passions, in divers men; and partly from the difference of the knowledge, or opinion each one has of the causes, which produce the effect desired.

So that in the first place, I put for a generall inclination of all mankind, a perpetuall and restlesse desire of Power after power, that ceaseth onely on Death. (Leviathan, pp. 160–61)

17 [The text at this point reads: “(first Sentence—purpose of organic being is life-conflicts with B).” This refers to the text quoted above: “In the natural constitution of an organic being—that is, of one contrived for the purpose of life—let us take it as a principle that in it no organ is to be found for any end unless it is also the most appropriate to that end and the best fitted for it.” The suggestion seems to be that this passage contradicts the idea that morality is man’s highest purpose (for the passage suggests that the purpose of organic being is life).—Ed.]

18 Kant, Groundwork of the Metaphysic of Morals (411), p. 79, p. 76. [Note that Cohen is quoting from Abbott’s translation, which differs in minor ways from Paton’s translation. Kant, Fundamental Principles of the Metaphysics of Ethics, p. 34.—Ed.]

19 [As will become apparent, this repetition is deliberate.—Ed.]

20 Kant, Groundwork of the Metaphysic of Morals (388), p. 56. p. 54.

21 [The text at this point contains the remark: “[State what I want to do here—distinguish between them contentlessly, so that any imperative might be either.]”—Ed.]

22 Note that such signs commonly appear on escalators on the London underground, to allow those who want to travel more quickly to walk up or down on the left-hand side.

23 See Kant, Groundwork of the Metaphysic of Morals (448), p. 116, p. 109, for support.

24 Desire is by nature particular; will is by nature universal, not restricted by a particular passion: so 3 and 12 are connected: and cf. 5.

25 Kant, Groundwork of the Metaphysic of Morals (431). (Abbott, p. 59, Paton, pp. 99–100, pp. 93–95.)

26 [Here Cohen quotes the Abbott translation. The corresponding passage in Paton is p. 99, p. 94.—Ed.]

27 Kant, Groundwork of the Metaphysic of Morals (437–40), pp. 104–7, pp. 98–101.

28 Kant, Metaphysical Principles of Virtue (386), p. 43.

29 See also Kant, Groundwork of the Metaphysic of Morals (397–98), pp. 65–66, pp. 63–65.

30 See Sellars, “Thought and Action,” pp. 136–39.

31 [The text at this point contains the remark: “(show how one might plausibly hold former view).”—Ed.]

32 [The next line in the text reads: “Principles are universal: doesn’t follow merely such. (?)” Presumably more work on this thought is indicated here.—Ed.]

33 Kant, Groundwork of the Metaphysic of Morals (399–400), pp. 65–66, pp. 67–68.

34 Weber, “Politics as a Vocation.”

35 Ibid., p. 121.

36 Ibid.

37 [The text at this point contains the remark: “[Reformulate, bringing in will].”—Ed.]

38 [Cohen has quoted Abbott’s translation, Kant, Fundamental Principles of the Metaphysics of Ethics, p. 78.—Ed.] Kant, Groundwork of the Metaphysic of Morals (97), p. 104, p. 107.

39 Ibid. (434) p. 101, p. 96.

40 [The text at this point adds a note: ‘[NB: THIS MATERIAL TO BE INTEGRATED WITH SUPPLEMENT BELOW].’ As no instructions are given about how the integration is to be managed, the supplementary material is added in complete form below.—Ed.]

41 [The text at this point continues: “[as Burde’s theory is a historicization and Hegel’s a metahistoricization of that theory.]” “Burde” is presumably intended to be “Burke.”—Ed.]

42 Kant, Groundwork of the Metaphysic of Morals (441–45) pp. 108–13, pp. 102–6.

43 [The text at this point continues: “(Perhaps criticism above can be averted: see my remark, p. 100).” This refers to a remark written in Cohen’s personal copy of the 1947 edition of Paton’s The Moral Law, which reads: “Note that heteronomous principles are those proposed by other moral philosophers, not ones actually acted on. This may solve the problem how there can be rational heteronomy (without desire)—there can be, in a false doctrine.”—Ed.]

44 Kant, Groundwork of the Metaphysic of Morals (450), p. 118, p. 111.

45 Ibid. (407), p. 75, p. 72.

46 [At this point in the text contains the remark “N.B. THIS MATERIAL TO BE INTEGRATED WITH SUPPLEMENT. … BELOW’. Once again there are no instructions on how to integrate the material, and so the additional material is simply added as an appendix to this chapter.—Ed.]

47 [At this point the text contains the note: “Expand.”—Ed.]

48 The ambiguity treated here is like that involved in “He had no motive for this action”: (a) There was, in fact, no (good) reason to do it; (b) He did it for no reason. A skeptic about reasons could say these two things, just as Kant, as skeptic, supposes these two things about apparently moral reasons.

49 Kant, Groundwork of the Metaphysic of Morals (388), pp. 54–56, pp. 56–58.

50 Ibid. (406–9), pp. 73–75, pp. 71–73.

51 Ibid. (409–11), pp. 76–79, pp. 73–75.

52 [Here Cohen includes the following footnote: “No, look at Kant himself, p. 87.” Assuming that Cohen is referring to his edition of Paton’s The Moral Law, there are two marked passages on that page, but the one that corresponds most closely to Paton’s remark reads: “[W]e must rather suspect that all imperatives which seem to be categorical may none the less be covertly hypothetical.” Kant, Groundwork of the Metaphysic of Morals (419), p. 87, p. 82.—Ed.]

53 Paton, The Categorical Imperative, p. 127.

54 [The text at this point contains the remark: “Expound each, show how he holds the former.”—Ed.]

55 [The text continues here with a short appendix to this section. This is distinct from the supplementary material, referred to above, and included as appendixes below.—Ed.]

56 [Presumably this stands for “the supreme moral principle.”—Ed.]

57 [The text here adds a comment “To be integrated with Supplement.” Once again there are no instructions for how the text is to be integrated, so it is included as an appendix.—Ed.]

58 [The text at this point contains the note: “(Remaining distinctions: analytical part; this: synthetical).”—Ed.]

59 Kant, Groundwork of the Metaphysic of Morals (450–453), pp. 118–21, pp. 111–13.

60 [Here Cohen quotes from the Abbott translation. The corresponding Paton translation is ibid. (108–9), p. 120, pp. 112–13.—Ed.]

61 But see ibid. (451–63), pp. 118–31, pp. 116–23: the doctrine is much more complex than is allowed here.

62 Ibid. (115–19), pp. 124–26, pp. 116–18.

63 Kant, Critique of Pure Reason B306, pp. 266–67.

64 Ibid. B106, p. 113.

65 [Cohen does not give further references. Strawson is clearly his contribution to a symposium with G. J. Warnock and J. F. Thomson, entitled “Determinism,” which is a transcription of a BBC radio discussion. MacKay is most likely to be “On the Logical Indeterminacy of a Free Choice.” The Hampshire is harder to pin down. Cohen is most likely to have in mind one or more of Hampshire’s books Thought and Action, Spinoza, and Freedom of the Individual.—Ed.]

66 Kant, Groundwork of the Metaphysic of Morals (402–3), pp. 69–71, pp. 67–68.

67 [At this point the text continues with the following rough notes: “Connected view that evil man more praiseworthy. (The ‘love’ of the Gospels is practical, not pathological). A deep antinomy in our moral thinking (contrast with Nietzsche). Contrast rising above environment, above fiendish passion.”—Ed.]

68 [It seems likely that this supplement was added when Cohen typed up his lectures in 1999.—Ed.]

69 Kant, Groundwork of the Metaphysic of Morals (389), p. 57, p. 55.

70 Ibid. (407), pp. 74–75, pp. 71–72.

71 Ibid. (419), pp. 86–87, p. 82.

72 Paton, The Categorical Imperative, p. 127.

73 [It is apparent from the manuscript that these final pages are more of a draft than the previous discussion, which underwent several revisions. It is included for completeness, as well as for the intrinsic interest of the material.—Ed.]

74 [Here the text contains the remark “Begin by reading pp. 118–9 sidelined. (And explain it.)” The passage is set out above, though no explanation is added.—Ed.]

75 Kant, Groundwork of the Metaphysic of Morals (106), pp. 118–19, p. 111.

76 Kant, Critique of Pure Reason B306ff., pp. 266 ff.

77 Ibid. B106, p. 113, for the whole list.

78 Kant, Groundwork of the Metaphysic of Morals (106–7), pp. 118–19, 111–12.

79 Ibid. (448–49), pp. 115–16, pp. 108–9.

80 Cohen, “Beliefs and Roles,” pp. 30–31.

81 See Kant, Groundwork of the Metaphysic of Morals (452–53), pp. 120–21, pp. 112–13.

82 See Kant, Metaphysical Principles of Virtue (423), p. 84.

83 [Presumably a reference to Peter Winch’s famous discussion in “The Universalizability of Moral Judgements.”—Ed.]