1. Introduction
Among the apparently central themes of Kant’s ethical theory is that of universality. Even those with a cursory acquaintance with his ethics will recognize his Formula of Universal Law (FUL): ‘act only according to that maxim through which you can at the same time will that it become a universal law’ (G, 4: 421).Footnote 1 This formula, he claims, gives expression to the categorical imperative, the single principle of morality.
It is, then, unsurprising that much criticism of Kant’s ethics takes the FUL as its principal target. My concern here is with a kind of criticism which alleges that the FUL is unfit to serve its assigned role in the argument of the Groundwork of the Metaphysics of Morals and especially in the Groundwork’s second section (henceforth ‘Groundwork II’). In that text, the FUL stands between Kant’s discussion of the character of categorical imperatives – principles of action insensitive to agents’ given purposes – and his introduction of such other central notions as those of humanity as an end in itself, the autonomy of the rational will, and the kingdom of ends. And it appears to suffer defective argumentative connections both to what precedes it and to what follows it. I shall describe these apparent problems in detail in §2, but here is a sketch of their basic structure. On the one hand, Kant seems to claim that an examination of the sheer notion of a categorical imperative reveals that there is a single categorical imperative and that the FUL is its expression. But it seems that this is false. It seems that FUL differs from the notion of a categorical imperative in both sense and extension. On the other hand, Kant claims that principles introduced later in Groundwork II, formulated in terms of the notions of humanity as an end in itself, the autonomy of the rational will, and the kingdom of ends, are equivalent to the FUL and thus themselves expressions of the single categorical imperative. But this too seems false. It seems that the Formula of Humanity (FH) and Formula of Autonomy (FA) differ from the FUL in, again, both sense and extension.
These two problems are bad enough, but they yield a third: if the FH and FA are not equivalent to the FUL and the FUL is not equivalent to the notion of a categorical imperative, then the FUL cannot serve as the argumentative link between what precedes it and what follows it which Kant appears to intend it to be. Groundwork II, whose central task is apparently the procession of the three formulae from the notion of a categorical imperative, is a comprehensive failure.
Commentators have tended to respond to these apparent problems in one of two ways. First, there are those who believe them to be real and, indeed, insoluble, so that the expository development of Groundwork II must be declared a failure, though, they ordinarily hold, much of philosophical value can be extricated from it.Footnote 2 Second, there are those who believe these apparent problems to be merely so, who believe that the notion of a categorical imperative and the FUL, once properly understood in light of the FH and FA, are seen to be equivalent to each other and to the FH and FA.Footnote 3 But this approach seems also to require interpretative abandonment of the expository structure of Groundwork II, at least insofar as it does not include a demonstration of how the FUL itself and the material preceding its introduction provide for this reading.Footnote 4
My aim in this article is to argue for another approach whose possibility emerges from two things: a recharacterization of the argumentative problems I have described and an application to them of three other important themes from Kant’s practical philosophy. The recharacterization is in terms of three different notions of universality: the first apparently embodied in the sheer notion of a categorical imperative, the second in the FUL, and the third in the later formulae. In these terms, the problems above can be described as that of how these notions are related to one another and of how this relation is expressed by the text of Groundwork II in such a way as to support its argumentative structure. The three themes to be brought to bear are the idea of the Groundwork as proceeding from what Kant calls ‘common moral cognition’, the notion of a ‘typic of pure practical judgement’ introduced in the Critique of Practical Reason, and the notion of a capacity (Vermögen, Fähigkeit) at work in Kant’s Critical philosophy.Footnote 5 In the end, I hope to have shown with the help of these resources that the basic argumentative development of Groundwork II is to be acquitted of the two charges described above, though problems about the FUL itself remain.
I begin by describing in more detail the two problems about the Groundwork’s interpretation introduced above and by recharacterizing them in terms of the three different notions of universality (§2). I then consider what Kant says about the relation between the second and third notions in the second Critique (§3). With this in hand, I turn to Groundwork I (§4) and II (§5) and present an account of the latter’s argumentative structure in light of §§2–3.
2. Two argumentative problems
2.1 The gap in the derivation
The first problem is that of the gap in the derivation.Footnote 6 The gap, if there is one, is thoughtfully marked in Groundwork II by a paragraph break:
[W]hen I think of a categorical imperative I know at once what it contains. For since besides the law the imperative contains only the necessity of the maxim to conform with this law, whereas the law contains no condition to which it was limited, nothing is left but the universality of a law as such, with which the maxim of the action ought to conform, and it is this conformity alone that the imperative actually represents as necessary.
There is therefore only a single categorical imperative, and it is this: act only according to that maxim through which you can at the same time will that it become a universal law. (G, 4: 421)
The notion of the possibility of willing that my maxim become a universal law is then illustrated by the application of Kant’s universalizability procedure to his four examples (G, 4: 421–4). This illustration funds our understanding of the notion: what it means for it to be possible for me to will that my maxim become a universal law is given by the procedure.Footnote 7 Thus, suppose I consider whether to eat this piece of cake, which no one else wants, because it looks delicious. For it to be possible for me to will that my proposed maxim become a universal law is for it to be possible for me to will this and at the same time will a world in which everyone, as a matter of course, eats pieces of cake which no one else wants because they look delicious. The universality of the universalizability test is the universality of cake-eating under circumstances relevantly similar to mine: similar, that is, according to the content of my maxim.Footnote 8 Call this maxim-universality. A maxim is universalizable if I can act on it and at the same time will that it should have maxim-universality.
Before the break, Kant seems to say that the categorical imperative expresses the necessity of the conformity of an agent’s maxim to an unconditioned law. In the context, in which Kant has contrasted categorical with hypothetical imperatives (G, 4: 414f.), it is clear that the kind of condition of relevance is that of an agent’s given purposes. A hypothetical imperative expresses the necessity of my doing something, conditional on my having some purpose (4: 414): should it be my purpose to bake a cake, I must preheat the oven. A categorical imperative expresses the necessity of my doing something whatever my given purposes might be: ‘The categorical imperative would be that which represented an action as objectively necessary of itself, without reference to another end’ (4: 414). The law to which I must, according to such an imperative, conform is universal not just in the sense of applying to all rational beings but also in being universal with respect to given purposes, in being indifferent to their content. Let us call this kind of universality purpose-indifference. It is important that purpose-indifference is a property of imperatives and that an imperative is ‘the representation of an objective principle in so far as it is necessitating for a will’ (4: 413). Suppose I tell you to make me some toast and in so doing betray no concern for your given purposes. Though I use the imperative mood, what I express is not an imperative in Kant’s sense: the representation of a principle necessitating for your will. So although it may be apt to describe my instruction as insensitive to your purposes, it is not a case of purpose-indifference in the sense I have introduced.
There appears to be a gap in the derivation because Kant appears to proceed as follows. First, he defines categorical imperatives as purpose-indifferent imperatives. He then – this is meant to be the derivation – claims that the one and only purpose-indifferent imperative is this: act only on maxims with maxim-universality. But the notion of maxim-universality is not the same as that of purpose-indifference, notwithstanding the use of the word ‘universal’ in the definition of each. Thus, he must give some indication of how to confirm or disconfirm a candidate principle’s claim to be a purposive-indifferent imperative and then show that this criterion applies favourably to the FUL and only to the FUL. Consider a competitor to the FUL: the candidate principle ‘do whatever Lytton Strachey would do in your situation’. It may seem ludicrous to suppose that this principle might be an imperative, but at this point in the Groundwork, Kant appears to have said nothing to rule out that possibility.Footnote 9, Footnote 10 I said above that not every unqualified instruction expressed in the imperative mood is a categorical imperative. But Kant has not, at this stage, explained how we are to distinguish those which are not imperatives from those which are.Footnote 11, Footnote 12, Footnote 13
2.2 The procession of the formulae
The problem of the connection of the FUL to later formulations of the categorical imperative concerns a much longer stretch of text. Henry Allison puts it in the following way. He argues – and I agree – that Groundwork II proceeds by way of ever-deeper characterizations of rational agency (Allison Reference Allison2011: 182). He identifies the procession’s beginning as what he calls ‘intra-subjective universalizability’, the compatibility of an agent’s maxim with the same maxim considered as a universal law, and its conclusion as what he calls ‘inter-subjective universalizability’, which requires the endorsability of an agent’s maxim by all rational agents (Allison Reference Allison2011: 196). The latter universalizability is immediately recognizable in the idea of the kingdom of ends: ‘a whole of all ends (of rational beings as ends in themselves, as well as the ends of its own that each of them may set for itself) in systematic connection’ (G, 4: 433). Call this systematic universality. Allison says that there is a ‘great problem’ in understanding whether such a procession is possible.
This is unsurprising. The requirement of the endorsability of one’s maxim by others seems to yield a considerably more confined space of permission than the requirements embodied in the universalizability procedure alone. Consider, for example, Christine Korsgaard’s cases of killing for revenge or out of hatred. She remarks that ‘[t]hese grim kinds of cases are managed without difficulty when using the Formula of Humanity, but it will be difficult to find any contradiction of the sort needed’ in applying the FUL (Korsgaard Reference Korsgaard1996: 100).Footnote 14 This is because the universality of killing for revenge or out of hatred would not make it any more difficult for me to kill out of revenge or out of hatred (as long as killing is all I will to do: as long as, for example, I do not will to kill and live to kill another day, since it is likely that, in the world of the universalized maxim, someone else will kill me before too long) and because I can consistently will the universality of killing for revenge and will never to be killed (as long as I will never to do anything avengeable).
The problem of the procession of the formulae is, therefore, that though some moral content can be identified in the FUL, it appears that later formulations have more. How, then, can they be not just equivalent to the FUL, but deeper expressions of what it expresses?
3. The type of systematic universality
The problem of the procession of the formulae is that of how the requirement to act only on maxims whose universality I can will – in the sense of maxim-universality – can be understood to be equivalent to the requirement that my action has systematic universality. It is natural for the reader of Groundwork II to want to be led from the FUL to systematic universality, since the FH and the FA follow the FUL in the text. But I want, for the moment, to set this matter aside and consider something that occurs in the second Critique which suggests that whatever the Groundwork’s expository order might be, in the proper order of understanding, it is systematic universality that comes before the FUL. Later, I shall explain how I think this material should be brought to bear upon the interpretation of Groundwork II.
3.1 The will of all
Consider a remark Kant makes in the discussion of the third theorem of the second Critique. In this remark, he distinguishes between the sense in which the wish or demand (Verlangen) for happiness is universal and the sense in which an action performed in observance of the moral law is universal. He says that, on account of this distinction, the latter cannot be explained in terms of the former:
It is, therefore, strange that intelligent men could have thought of passing off the desire for happiness as a universal practical law on the ground that the desire, and so too the maxim by which each makes this desire the determining ground of his will, is universal. For whereas elsewhere a universal law of nature makes everything harmonious, here, if one wanted to give the maxim the universality of a law, the most extreme opposite of harmony would follow, the worst conflict, and the complete annihilation of the maxim itself and its purpose. For then the will of all has not one and the same object but each has his own (his own welfare), which can indeed happen to accord with the purposes of others who are likewise pursuing their own but which is far from sufficing for a law because the exceptions that one is warranted in making upon occasion are endless and cannot be determinately embraced in a universal rule. (CPrR, 5: 28)
The wish for happiness is universal: everyone wants it.Footnote 15 But I wish for my happiness, and you wish for yours, and so we do not wish for ‘one and the same object’. The universality of the wish for happiness is maxim-universality. What Kant says in this passage entails that if the wish for happiness is not limited by another maxim, there can be, at most, a merely accidental harmony among wills. (It might happen that the best way for each of us to get what she wants is to allow, and perhaps even to help, others to get what they want.) The maxim whose universal adoption would make for a genuine harmony among wills would be one whose universal adoption would mean that all individual wills have ‘one and the same object’. If it is meaningful – as Kant implies here – to contrast such a maxim with the wish for happiness in the case of accidental harmony, then for all individual wills to have ‘one and the same object’ must be for them to stand in non-accidental practical agreement with one another, in the sense of Kant’s idea of a kingdom of ends:
By a kingdom, however, I understand the systematic union of several rational beings through common laws. Now, since laws determine ends according to their universal validity, it is possible – if one abstracts from the personal differences among rational beings, and likewise from all content of their private ends – to conceive a whole of all ends (of rational beings as ends in themselves, as well as the ends of its own that each of them may set for itself) in systematic connection, i.e. a kingdom of ends, which is possible according to the above principles. (G, 4: 433)
This condition is thus one characterized by what I have called systematic universality. In this condition, because the ends of each individual will are systematically unified with the ends of every other, there is a practical coherence, and thereby a volitional unity, among wills. There is a sense in which the will of all, having one and the same object, is one will. I shall henceforth use Kant’s expression ‘the will of all’ in this sense.
Kant suggests that this notion of universality is the fundamental notion of universality in his ethics. For though the kingdom of ends is ‘of course only an ideal’ (G, 4: 433), nonetheless, ‘[m]orality … consists [besteht] in the relation of all action to the legislation through which alone a kingdom of ends is possible’ (4: 434; emphasis and translation mine).Footnote 16 I want to understand this in the following way. Systematic universality is, on Kant’s account, the perfection of the finite rational will.Footnote 17 ‘Perfection’ here means the perfect exercise of a capacity.Footnote 18 Morality consists in the perfect exercise of the rational will because the latter is a capacity whose principle of exercise is the moral law; our most fundamental understanding of a capacity is our understanding of what it is a capacity to do; and what a capacity is a capacity to do is the same as its perfect exercise. We may formulate the principle of the finite rational will in these terms: ‘constitute a unified will with every other finite rational being’.
3.2 The typic of pure practical reason
Kant thinks, however, that there is a problem with formulations like that: they are not tractable in application to actions – or, equivalently, in application to the adoption of maxims – by an individual in a particular circumstance of action. According to his doctrine in the second Critique, it is on account of this problem that we need the FUL. The FUL is, according to Kant, the solution to a problem of tractability in the application of moral principles and concepts to actions: that is, in practical judgement.
In the section devoted to this topic in the second chapter of the second Critique’s Analytic, the source of the intractability is identified as the distinction between freedom and nature:
[A] practical rule of pure reason … brings with it necessity with respect to the existence of an action and is thus a practical law, not a natural law through empirical grounds of determination but a law of freedom in accordance with which the will is to be determinable independently of anything empirical (merely through the representation of a law in general and its form); however, all cases of possible actions that occur can be only empirical, that is, belong to experience and nature; hence, it seems absurd to want to find in the sensible world a case which, though as such it stands only under the law of nature, yet admits of the application to it of a law of freedom and to which there could be applied the supersensible idea of the morally good, which is to be exhibited in it in concreto. (CPrR, 5: 68–9)
The specific character of the problem depends upon why it should seem ‘absurd’ to find a natural object to which we may have grounds to apply the concept of the morally good. One obvious reading is that to have grounds to apply a concept to a natural object is ordinarily to recognize the applicability to it of other concepts. I may, perhaps, judge that something is a lump of gold on the strength of my recognition of its being yellow, malleable, and soluble in aqua regia. But because the concept of the morally good belongs to an entirely different order than do theoretical concepts, it does not enjoy such connections to such concepts as would underwrite its application.
There seems, however, to be another dimension to the absurdity, because Kant describes the concept of the morally good as an idea. An idea is a representation that surpasses (übersteigt) everything possible in the sensible world (A320/B377). It represents a perfection of which the sensible always falls short. Thus, nothing in the sensible world satisfies the descriptive content of any idea, including that of the morally good. But the satisfaction of a concept’s descriptive content by an object is, at least ordinarily, what makes that concept applicable to that object. There is, therefore, a problem about how the concept of the morally good might be applied to any particular proposed action.
Kant’s solution to this problem is the ‘typic of pure practical judgement’, a canon for the power of practical judgement, the power of the application of practical concepts, like that of the morally good, to proposed actions:
The rule of judgement under laws of pure practical reason is this: ask yourself whether, if the action you propose were to take place by a law of the nature of which you were yourself a part, you could indeed regard it as possible through your will. (CPrR, 5: 69)
Idiomatic differences notwithstanding, this is immediately recognizable as a variation on the FUL of the Groundwork. Indeed, just as in the Groundwork, our understanding of the formula is funded by an illustration that immediately follows its appearance:
Everyone does, in fact, appraise actions as morally good or evil by this rule. Thus one says: if everyone permitted himself to deceive when he believed it to be to his advantage, or considered himself authorized to shorten his life as soon as he was thoroughly weary of it, or looked with complete indifference on the need of others, and if you belonged to such an order of things, would you be in it with the assent of your will? … If the maxim of the action is not so constituted that it can stand the test as to the form of a law of nature in general, then it is morally impossible. (CPrR, 5: 69–70)
And near the end of Groundwork II, he says, in characterizing the FUL in contradistinction to other formulae, that
in moral judging it is better always to proceed by the strict method, and make the foundation the universal formula of the categorical imperative: act according to the maxim that can make itself at the same time a universal law. (G, 4: 436–7)
3.3 The priority of systematic universality to maxim-universality
Kant warns that we should not reckon ‘among concepts themselves that which belongs only to the typic of concepts’ (CPrR, 5: 70). The concept in question is that of the morally good; the typic is the FUL. He thereby suggests that the FUL, taken alone, is not a perspicuous expression of the content of that concept. It is merely the principle by which that concept is to be applied to objects. Systematic universality, on the other hand, as expressed by the idea of the kingdom of ends, is, as we have seen, that in which morality ‘consists’ (besteht). Systematic universality is the content of the concept of the morally good.
We have reason to doubt that the FUL really is a typic for the concept of the morally good. This is the same as the reason, discussed in §2.2, we have to doubt that the FUL is equivalent to the later formulae. Note, however, that what appears to show up in a reading of the Groundwork as a problem about a movement from the FUL to those principles and ideas here shows up as a problem about the movement from them to it. Part of what I am about to propose is that, appearances to the contrary, the latter is also the way in which the problem shows up in the Groundwork. This is because, on the reading I am about to offer, Groundwork II does not proceed from the sheer notion of the categorical imperative to the FUL, then from the FUL to the later formulae, in such a way as would depend upon a dubious equivalence between purpose-indifference, maxim-universality, and systematic universality. Rather, Groundwork II proceeds from a philosophically undeveloped understanding of systematic universality to a philosophically developed one: a procession that may be secure even if the FUL’s status as a typic of such universality is dubious. The difference between the Groundwork and the second Critique, according to this reading, is that in the former the FUL appears against the background of the philosophically undeveloped understanding and in the latter against the background of the philosophically developed one. In both cases, it belongs not ‘among concepts themselves’ but ‘only’ to their typic.
The significance of this claim depends on the specific character of the problem to which the typic is meant to be the solution. Suppose, for example, that when Kant says that without such a typic it is impossible to apply to a natural object a concept belonging to the order of freedom, he means that the terms in which the typic are formulated must not include those of the rational will qua rational (for such terms belong to the order of freedom exclusively). Then there is a sense in which notions figuring in the later formulae cannot be read into the FUL. For in that case, the FUL is intended to supply a criterion, formulable in natural terms alone, for the applicability of a concept belonging to the order of freedom. To read such a concept into it is to understand it as not formulable in natural terms alone.
It seems to me that this is part of what Kant means. But I do not wish to argue this point. I wish, rather, in what follows, to argue that even if this is what he means, and even if we have reason to doubt that the FUL really is a typic for the concept of the morally good, understood in terms of systematic universality, nonetheless, we may judge the exposition of Groundwork II to be in good argumentative standing.
This leaves open, of course, that Kant is even better off than I suggest in this article. Perhaps the most promising plan of rescue for the FUL’s status as a typic involves attention to what he means by ‘nature’ (here and in the Formula of the Law of Nature as expression of the FUL). Stephen Engstrom, for example, suggests that ‘a universal law of nature can be a universal law of our rational nature … and accordingly the consideration of whether a maxim can be willed as a universal law of nature can be understood as the attempt to conceive of it as a law that all rational beings necessarily follow out of their shared recognition of its validity’ (Engstrom Reference Engstrom2009: 161). Sven Nyholm argues for a similar position, appealing to Kant’s distinction between nature considered formaliter and nature considered materialiter (Nyholm Reference Nyholm2015: 290–4).
Now, it may be asked why, if there is a reading on offer according to which Kant does not have a problem, we should care how things stand even if he does. One of the aims of the interpretation of a philosophical text is to achieve precision about different problems an author has been alleged to face. To show that even if she faces one, she does not face another which has been thought to be the same as, or bound up with, the first, itself contributes to such an achievement. Since, however, I have not concealed my own sympathy with a reading on which the FUL is to be understood in natural terms, let me briefly indicate the ground of this sympathy. It is simply that in what is, in my view, the clearest application of the universalizability test, the case of the lying promise, it seems for all the world that what is imagined in the world of the universalized maxim is not ‘shared recognition … of validity’ but rather everyone’s being in the habit of making lying promises to repay a loan whenever in need of ready cash, and that Kant’s point here is that once people have got wind of the practice, the next person who tries will not get away with it. ‘[N]o one would believe he was being promised anything, but would laugh about any such utterance, as vain pretense’ (G, 4: 422). This is not the place to address the many interesting alternatives to this reading. (See, for some of these alternatives, Engstrom Reference Engstrom2009: 196–209, Geiger Reference Geiger2015: 406–7, and Guyer Reference Guyer2019: 26–7.) I acknowledge that I may place too much emphasis on what a single passage appears to suggest against the weight of other evidence. But it is certainly not ridiculous to understand this passage to have the content and significance that I am inclined to attribute to it, as indeed many readers have done; this is enough to secure the interest of asking what else is true on the assumption that this reading is correct.
3.4 The Groundwork and the second Critique
Before I turn to the Groundwork, it is important for me to note, since I am about to apply to its interpretation a notion elaborated in the second Critique, that I do not here defend any particular position on perhaps the most vexed of the questions concerning the relation between these two books: that of whether the second Critique’s Fact of Reason (CPrR, 5: 30–1) constitutes a doctrinal departure from the content of the Groundwork.Footnote 19 Common to my discussion and to discussions of that question is the notion of common moral cognition, but it figures in the two cases in importantly different ways. The fact of reason is, roughly speaking, our appreciation of obligation and thereby of our freedom, such as might be described as an exercise of common moral cognition. That the reader of Kant’s practical works shares in this appreciation is meant to vouchsafe the objective reality of the concepts of moral obligation and of freedom insofar as they figure in those works. There is a question whether, in the Groundwork, and in particular in Groundwork III, Kant attempts to develop a different mode of establishing their reality, one which does not rely on any appreciation that might be described as an exercise of common moral cognition.
I am about to argue that in Groundwork II, no less than in Groundwork I, Kant relies on the reader’s appreciation of what morality requires on the assumption of the objective reality of the concept of moral obligation (and, thereby, of freedom). Groundwork II is, I shall suggest, devoted to a philosophical elucidation of the nature of a capacity first characterized as a capacity the principle of whose operation is the moral law. This is the capacity that in common moral cognition we presuppose we have, though that cognition does not include the philosophically perspicuous elucidation of its nature which it is the task of Groundwork II to provide. Groundwork II is not, however, addressed to the question of whether we human beings really possess this capacity, as, in common moral cognition, we presuppose we do. That is a matter for Groundwork III and, in the second Critique, for the fact of reason. Thus, my topic is not Kant’s appeal in the second Critique to common moral cognition to secure the reality of a concept – that of the capacity he describes in Groundwork II – but his appeal to common moral cognition to supply the raw material needed for a philosophical elucidation of the content of that concept, on the assumption of its reality.
4. Love thy neighbour
Groundwork II, I shall suggest, proceeds from a philosophically undeveloped understanding of systematic universality to a philosophically developed one. But for the principal expression of the first of these understandings in the Groundwork, we must turn to Groundwork I. Having found it in Groundwork I, I shall explain why and how we can take it to be a point of departure in Groundwork II.
Groundwork I begins with a motivational analysis of the notion of a good will, the concept of which ‘already dwells in natural sound understanding and needs not so much to be taught as only to be clarified’ (G, 4: 397).Footnote 20 A finite will is good insofar as it is exercised ‘from duty’ (aus Pflicht) (4: 398). To act from duty is, Kant stresses, not merely to act ‘in conformity with duty’ (4: 398). It is possible to act in conformity with duty but on account of an inclination towards the action in question, whether immediate or mediated. In illustrating these various motivational possibilities, Kant supposes that all will grant that duty demands various things: for example, that ‘a shopkeeper not overcharge an inexperienced customer’ (4: 397), that one ‘preserve one’s life’ (4: 397), that one ‘be beneficent where one can’ (4: 398), and that one ‘assure one’s own happiness’ (4: 399).
But these examples are not merely illustrative. They are, I suggest, essential to the reader’s getting a grip on what Kant is talking about. To see this, set all possible examples aside, and ask: what is the concept of an action from duty? What we can say is that it is an action that is not performed on account of inclination, immediate or mediated. (We might try saying that it is an action performed with a good will, but it is meant to be the concept of duty which clarifies that of the good will, not the other way about (G, 4: 397). And even if it were the other way about, the reader would require examples to get a grip on the notion of the good will.) So one mode in which actions are intelligible – that is, motivation by inclination, immediate or mediated – is unavailable. How else might an action be intelligible? Some philosophers have, of course, thought that this is the only way in which an action can be intelligible. But Kant, his examples at the ready, does not even leave the space for reflection at this level of generality. It is, he says, possible to do such things as to decline to overcharge inexperienced customers and to help others where one can simply because these are the right things to do. The reader’s initial grip on the more generic notion of motivation by something other than inclination depends on her grip on the more specific notion of doing something because it is right. And her grip on that notion is essentially funded by her understanding of these examples.
Kant’s examples are, I think, enough to elicit in the reader the content of common moral cognition which, as I argue below, he will bring to philosophical explicitness in Groundwork II. But it is instructive to consider another kind of remark which he may intend to serve a similar function, and which he makes having developed his four examples:
It is undoubtedly in this way, again, that we are to understand the passages from scripture in which we are commanded to love our neighbour, even our enemy. For, love as an inclination cannot be commanded, but beneficence from duty – even though no inclination impels us to it and, indeed, natural and unconquerable aversion opposes it – is practical and not pathological love, which lies in the will and not in the propensity of feeling, in principles of action and not in melting sympathy; and it alone can be commanded. (G, 4: 399)
This remark may seem to serve as an elaboration on one of the examples already given, that of being beneficent where one can. And beneficence certainly belongs to what Kant has in mind here. But remarks he makes on the scriptural injunction elsewhere suggest that he may comprehend more in it than this. Kant is recorded as saying, for example, that ‘[t]he love for others can be considered in its generality, so to that extent it rests on this, that our ends coincide with those of others in such a way that they are able to co-exist together according to the universal rule of duty’ (Eth-Vigil, 27: 673). Even more strikingly, he is recorded as describing a ‘universal will’ which ‘consists in the universal end of all men, and is called love for others, the principle of well-wishing’ (27: 541). I wish to suggest that the structure of systematic universality is anticipated in this discussion of the scriptural injunction to practical love.
Now, it may seem that though ‘love thy neighbor’ boasts positive content, it boasts too much: that is, it is too specific to be an indication of what Kant means by ‘action from duty’. There are at least two issues here. The first is that, as mentioned above, in the Groundwork passage Kant seems to identify practical love with beneficence from duty, and yet there are, according to Kant, duties to others other than beneficence. To be sure, in the Metaphysik der Sitten Vigilantius passages I have already quoted, it seems that Kant is recorded as understanding love to be more general than beneficence in this narrow sense, since he is happy to identify it with the universal will. And this is also suggested by his discussion of the scriptural injunction in the second Critique. There he calls the commandment ‘Love God above all, and your neighbour as yourself’ the ‘kernel of all laws’ (CPrR, 5: 83). He says that only ‘practical love’ is understood in this commandment and immediately explains what he means as follows: ‘[t]o love God means, in this sense, to do what He commands gladly; to love one’s neighbour means to practice all duties toward him gladly’ (5: 83). Similarly, in the Religion Within the Boundaries of Mere Reason, we are told that the ‘founder of the one true church’
sums up all duties (1) into one universal rule (which includes the internal as well as the external moral relation of human beings), namely, Do your duty from no other incentive except the unmediated appreciation of duty itself, i.e. love God (the Legislator of all duties) above all else; (2) and into a particular rule, one namely that concerns the human being’s external relation to other human beings as universal duty, Love every one as yourself, i.e. promote his welfare from an unmediated goodwill, one not derived from selfish incentives. (R, 6: 160–1)
Since these rules are meant to sum up all duties, we cannot understand Kant’s gloss on the second rule in terms of welfare as restricted to beneficence in the narrow sense.
But can the common moral cognizer of Groundwork I be expected to anticipate this? I think so, but in a somewhat complicated way. To see it, consider that for Kant, even beneficence in the narrow sense is not principally a matter of addressing needs, such as any needy being might have. Rather, it is a matter of an orientation towards others’ ends, such as only a practically rational being can have, an orientation that is part and parcel of treating them as ends in themselves:Footnote 21
Now, humanity could indeed subsist if no one contributed anything to the happiness of others while not intentionally detracting anything from it; but this is still only a negative and not positive agreement with humanity, as an end in itself, if everyone does not also try, as far as he can, to advance the ends of others. For if that representation is to have its full effect in me, the ends of a subject that is an end in itself must, as much as possible, also be my ends. (G, 4: 430)
Duties of beneficence are grounded in this, that others are ends in themselves. But that status is equally what grounds, for example, the duty not to make lying promises (G, 4: 429–30). To acknowledge the immediate normative significance of the ends of others, and to act in accordance with this acknowledgment, is to love thy neighbour in the sense of the second Critique and the Religion. So, according to Kant, once we understand the ground of duties of beneficence, we have everything we need for an understanding of perfect duties too: we have everything we need to understand what, according to Kant, ‘love thy neighbor’ really means. It means to treat others as ends in themselves.
In light of this observation, I want to suggest that even if Kant indeed intends the reader to understand the ‘love thy neighbor’ passage of Groundwork I, unlike passages on the same topic in the second Critique and the Religion, as concerning beneficence in the narrow sense, he nonetheless believes that the common moral cognizer will, upon philosophical reflection – upon the kind of reflection he undertakes in Groundwork II – come to understand that the ground of duties of beneficence in the narrow sense equally grounds perfect duties to others and that these are not just two kinds of duty, one atop the other, but rather expressions of two aspects of one thing: the status of others as ends in themselves. In that case, the common moral cognizer can come to recognize that the formula ‘love thy neighbor’ is better suited to the expression of the command that we acknowledge that one thing than to the expression of the command that we acknowledge just one of its two aspects.
The second issue in connection with ‘love thy neighbor’ is that one has duties to oneself and one is not one’s neighbour. Indeed, the passage from the Religion I have just quoted suggests that duties to oneself (‘the internal … moral relation of human beings’) are not covered by ‘love thy neighbor’ but are covered by the more general ‘love God’. But it is possible that this is not as much of a problem as it might first appear. This is a topic that requires its own sustained treatment, but let me gesture towards what I mean. In the Moralphilosophie Collins, Kant is recorded as saying that ‘self-regarding duties are the supreme condition and principium of all morality’ (Eth-Collins, 27: 344) and that they are the ‘condition under which’ duties to others ‘can be observed’ (27: 341). This idea is not absent from the Critical published texts. Consider Kant’s explanations of the various duties to oneself in terms of one’s status as a practically rational being or in terms of one’s readiness to fulfil obligations. Self-mutilation (MM, 6: 422), gluttony (6: 427), drunkenness (6: 427), and avarice (6: 432) all make it more difficult to act in the fulfilment of one’s duties. Suicide would be ‘withdrawal’ from all obligations (6: 422). Then there are the prohibitions of recreational sex (6: 425), lying (6: 429), and servility (6: 435). Kant claims that these are offences against one’s personality. But one’s personality is precisely what constitutes one’s fitness to enter into moral relationships with others.
My suggestion, in light of this, is the following. It is not that duties to oneself are indirect, as though preserving one’s personality and developing one’s capacities are merely instrumentally directed towards moral service towards others.Footnote 22 Indeed, there would seem to be an incoherence in the idea of preserving one’s personality solely to be of any kind of service to others. Rather, it is the good of the finite rational being to live with other finite rational beings in the kingdom of ends, in which each stands in reciprocal moral relationships with all others. Standing in such relationships is in part a matter of helping each other to realize ends, thereby to be happy. But Kant thinks that there are some things one must do, and some things one must not do, in order to realize and preserve one’s nature (or standing, if you like) as the kind of being who can enter into such relationships: as the kind of being who fulfils duties to others, who is expected and, indeed, counted on to fulfil these duties, and who expects and counts on the fulfilment by others of their duties towards her. These conditions are articulated in the various duties to oneself. (This is to say nothing, of course, about whether Kant has succeeded in identifying these conditions.) This is, to be sure, a hypothesis, but I submit that it fits Kant’s arguments for the particular duties to self in the system of the Metaphysics of Morals, is true to the remarks from the Moralphilosophie Collins, and does not involve the error of thinking these duties indirect.
To return, finally, to the formula ‘love thy neighbor’: if what I have said is on the right track, then we can say that duties to oneself express conditions on the possibility of practically loving relationships (in the general sense of the formula). Not every being capable of entering into such relationships stands in readiness to do so: but because to be in such relationships is the good of such a being, she must fulfil the conditions that constitute this readiness. To fulfil one’s duties to oneself is to make and keep oneself fit for love: fit, that is, to love one’s neighbours and to be loved by them. Like the true ground of beneficence, this is not understood, in all its articulation, in common moral cognition. But again, Kant is confident that philosophical reflection on the formula will expose and clarify it.
And even if none of what I have said about ‘love thy neighbor’ in connection with perfect duties to others and duties to oneself is tenable, still Kant can rely on his four examples, in which duties of all three kinds are accounted for. In what follows, I shall, accordingly, use ‘love thy neighbor’ and similar expressions to capture what Kant elicits from common moral cognition in Groundwork I; however, much of this content is to be comprehended in his paragraph on the scriptural injunction as such.
5. The capacity whose exercise is action from duty
5.1 Categorical imperatives
I have suggested that philosophically undeveloped anticipations of systematic universality fund the reader’s understanding of the concept of duty in Groundwork I. In Groundwork II, Kant wishes to ‘follow and present distinctly the practical faculty of reason, from its general rules of determination to the point where the concept of duty arises from it’ (G, 4: 412). There are two ways in which such a task might be understood. Allison gives expression to the first of these in saying that Groundwork II ‘starts not with our ordinary, pre-philosophical moral convictions, but with the abstract philosophical concept of a rational agent as such’ (Allison Reference Allison2011: 35; emphasis mine). Thus, Allison suggests, the two sections make different and independently intelligible approaches to the concept of duty: Groundwork I begins with our ordinary, pre-philosophical moral convictions and Groundwork II with an abstract philosophical concept of a rational agent as such. Groundwork II makes, in this sense, a ‘fresh start’ (Allison Reference Allison2011: 35).
But Kant does not say, in the quoted characterization of his task, that the ‘general rules of determination’ of practical reason must be, or even seem to be, intelligible independently of the concept of duty which ‘arises from it’. To learn something’s source – to witness it ‘arising’ from this source – does not require that one first forget what the thing is. And not only is there no indication in Groundwork II to the effect that we should abandon the understanding of duty which rests upon common moral cognition: there are explicit indications to the contrary at the end of Groundwork I. There, Kant writes that it is ‘common human reason’ itself – whose cognition has been elucidated in Groundwork I – which must ‘take a step into the field of practical philosophy, in order to receive there intelligence and distinct instruction regarding the source of this principle and its correct determination’ (G, 4: 405; emphasis in the original). Common human reason cannot be denied entry to the classroom in which it is to receive instruction.Footnote 23 This is not to say that Groundwork II is not of a character importantly different from that of Groundwork I. As Jeanine Grenberg puts it, the ‘rational tools’ in the field of practical philosophy are distinctively philosophical: they are, in this sense, not the tools of the common moral cognizer (Grenberg Reference Grenberg2013: 91). But that to which the tools are applied is what, as depicted by Groundwork I, common moral cognition yields.Footnote 24
The other and, I think, better understanding of Kant’s task is that he wishes to supply a description of action from duty as the exercise of a rational capacity and thereby wishes to elucidate the structure of the capacity, which he calls ‘practical reason’, whose exercise it is.Footnote 25 This is in keeping with the general description of the method given at the end of the Groundwork’s Preface:
In this work, I have adopted the method that is, I believe, most fitting if one wants to take one’s route analytically from common cognition to the determination of its supreme principle and in turn synthetically from the examination of this principle and its sources back to common cognition, in which we find it used. (G, 4: 392)
The capacity in question is the source of the principle and is from the beginning identified as this source. (Kant here announces no intention to ground the principle of common moral cognition in a capacity identifiable and characterizable in entirely independent terms.)
In Groundwork I, having funded the reader’s understanding of the notion of duty with his examples and his discussion of practical love, he argues for two ‘propositions’, one of which – the ‘third proposition’ – is that ‘duty is the necessity of an action from respect for law’ (G, 4: 400).Footnote 26 It is with formulations in terms of law or laws that he begins his elucidation of practical reason in Groundwork II:
Only a rational being has the capacity to act in accordance with the representation of laws, that is, in accordance with principles, or has a will. Since reason is required for the derivation of actions from laws, the will is nothing other than practical reason. (G, 4: 412)Footnote 27
He then tells us that for the finite rational will, principles are expressed as imperatives because this kind of will does not inevitably (unausbleiblich) satisfy them (G, 4: 412–13). And, finally, he says that
all imperatives command either hypothetically or categorically. The former represent the practical necessity of a possible action as a means to achieving something else that one wills (or that it is at least possible for one to will). The categorical imperative would be that which represented an action as objectively necessary of itself, without reference to another end. (G, 4: 414)
Thus does the notion of a categorical imperative emerge from the investigation of the concept of a capacity to act from duty.
Note that the notion of a categorical imperative, as it appears here, has something important in common with the notion of action from duty as it appears in Groundwork I. If we set aside all material for understanding made available by common moral cognition, all we can find in this is a negative notion: a categorical imperative is one which is not qualified by an (actual or possible) end. This is what I meant in §§1–2 by ‘the sheer notion of a categorical imperative’: this notion is exhausted by purpose-indifference. If we suppose that the sheer notion of a categorical imperative is all we can understand in Kant’s description of a categorical imperative, then we cannot make sense of why, following a discussion of hypothetical imperatives, he straightaway declares that the categorical imperative ‘may be called the imperative of morality’ (G, 4: 416).
Of course, we might make sense of this by declaring Kant’s exposition sloppy or his argument unsuccessful. But if it is wrong to set aside all material for understanding made available by common moral cognition, then it is wrong to make such declarations (at least on the ground that purpose-indifference is not morality). The notion of a categorical imperative at work here is not just the notion of an imperative that is not qualified by an (actual or possible) end. It is, as we might put it, the notion of an imperative that is not qualified by an (actual or possible) end, like the injunction to practical love: what we lean on in common moral cognition to understand action from duty in Groundwork I we lean on here to understand what a categorical imperative is. This characterization is more philosophically developed than any available in Groundwork I because it is formulated in part in terms of Kant’s account of finite practical reason. But its development is not complete, nor will it be until the end of Groundwork II, when it is explicitly expressed in terms of systematic universality.Footnote 28
5.2 The FUL again
With all of this in hand, I can supply the promised sketch of the expository structure of Groundwork II. My principal claim is, as I have said, that Kant proceeds from a philosophically undeveloped understanding of systematic universality – as expressed, for example, by the words ‘love thy neighbor’ – to a philosophically developed one in the idea of the kingdom of ends. I have also said that the topic of Groundwork II is that of the capacity whose exercise is action from duty, initially understood in terms of the philosophically undeveloped understanding of systematic universality. Earlier, in §3.1, I said that the idea of the kingdom of ends is a philosophically perspicuous representation of the perfect exercise of this capacity and that we understand what a capacity is in terms of its perfect exercise. So much for the great sweep of Groundwork II, as proceeding from a philosophically undeveloped understanding of an act to a philosophically developed understanding of that act as the perfect exercise of a capacity. But this article began with a statement of some serious problems for the FUL, and it is time to turn to them.
The first problem is that of the gap in the derivation. Recall the moment of derivation:
[W]hen I think of a categorical imperative I know at once what it contains. For since besides the law the imperative contains only the necessity of the maxim to conform with this law, whereas the law contains no condition to which it was limited, nothing is left but the universality of a law as such, with which the maxim of the action ought to conform, and it is this conformity alone that the imperative actually represents as necessary.
There is therefore only a single categorical imperative, and it is this: act only according to that maxim through which you can at the same time will that it become a universal law. (G, 4: 421)
The first paragraph seems to express nothing more than purpose-indifference, with no indication of how to tell whether a candidate purpose-indifferent principle really is one. The second, read in light of the procedure for assessing maxims which Kant immediately illustrates, identifies the one and only purpose-indifferent principle as the requirement that one’s maxims have maxim-universality. But if we understand Kant’s use of ‘categorical imperative’ in the way I have suggested, then the first paragraph expresses a partially developed understanding of systematic universality. It thereby expresses an indication of what the one and only purpose-indifferent principle is: constitute a unified will with every other rational being. Given that our understanding of the second paragraph is funded by the illustrative examples that follow, Kant’s claim here is that a requirement formulated in terms of maxim-universality is equivalent to the requirement that one’s action have systematic universality (as partially understood).
As I said in §3.3, we have reason to doubt that this claim is true. But this is not because Kant has not given us any way of telling whether a candidate purpose-indifferent principle really is one. It is because it is difficult to see how maxim-universality could be equivalent to systematic universality. When we read the second Critique, the problem shows up as one about the equivalence of the former to the latter understood in a philosophically developed way: this is the problem of whether the FUL is really a ‘typic’ for systematic universality. When we read Groundwork II, the problem shows up as one about the equivalence of the former to the latter understood in a relatively undeveloped way. In the end, there is only one problem here, common to the two texts, and it is not that of the gap in the derivation. I think, however, that this is a serious problem, and, as I suggested in §3.3, believe that for this reason, it is wrong to read later formulations and principles back into the FUL.Footnote 29, Footnote 30
What of the problem of the procession of the formulae? First, if what I have just said is right, then there is indeed a serious problem about the equivalence of the FUL to the other formulations, a problem which is ultimately about whether the FUL is the principle of practical judgement: the principle according to which the bearer of practical reason self-consciously applies its concepts to proposals for action here and now. But, second, if what I have said since the beginning of §4 is right, this need not vitiate the expository sweep of Groundwork II, from a philosophically undeveloped notion of systematic universality to a philosophically developed one.
It need not: but does it, in fact, not? Can the FUL’s problems be sequestered? An important indication that they can is that, though the FUL appears in the text before the FH and FA, Kant’s exposition does not proceed directly from the FUL to the FH. Here is, in outline, his approach to the FH. First, he asks:
is it a necessary law for all rational beings always to appraise their actions in accordance with such maxims as they themselves could will to serve as universal laws? If there is such a law, then it must already be connected (completely a priori) with the concept of the will of a rational being as such. (G, 4: 426)
Thus, he suggests the task of discovering the a priori connection of the FUL to the concept of the will of a rational being as such. This necessitates a return to that concept, the first moment of whose explicit articulation occurred at G, 4: 412, a moment now repeated, before Kant adds a further reflection:
The will is thought as a capacity to determine itself to acting in conformity with the representation of certain laws. And such a capacity can be found only in rational beings. Now, what serves the will as the objective ground of its self-determination is an end, and this, if it is given by reason alone, must hold equally for all rational beings. (G, 4: 427; emphasis in the original)
He then argues that there is an end given by reason alone: that is, humanity (G, 4: 428). The point of departure for this argument is the same as that for the stretch of text leading up to the FUL. And this is to say that it is not the FUL. Rather, it is the idea of a capacity to act in accordance with, or in conformity with, the representation of laws, where this notion is understood, in accordance with Kant’s discussion in Groundwork I, in terms of action from duty, and where that notion is understood with the aid of common moral cognition. This suggests that, unless it makes a surprise argumentative appearance after this point – and I submit that it does not – its problems can indeed be sequestered.Footnote 31
6. Conclusion
I began this article with the claim that the problems of the gap in the derivation and the procession of the formulae should be understood in terms of the problem of how three notions of universality figure in Groundwork II: purpose-indifference, as expressed by the sheer notion of a categorical imperative, a principle insensitive to an agent’s given purposes; maxim-universality, the content of which notion is funded by Kant’s illustrations of the universalizability procedure; and systematic universality, a practical coherence among wills whose complete actuality is depicted by the idea of a kingdom of ends, in which every individual will shares in the one will of all.
The solution I have offered is this. Purpose-indifference does not figure in Groundwork II in its own right; it is an abstraction from what Kant understands in his notion of a categorical imperative. Systematic universality is the topic of Groundwork II, which proceeds from the philosophically undeveloped understanding of this universality elicited in the reader by Groundwork I – ‘love thy neighbor’ – to a philosophically developed understanding of the complete actuality of such universality as the perfect exercise of practical reason, as depicted in the idea of the kingdom of ends. Maxim-universality is the notion in terms of which Kant formulates the principle of practical judgement: that is, the principle by which, as he would have it, the self-conscious possessor of this capacity exercises it in particular cases, remote from its perfect actualization in the kingdom of ends. This formulation has serious problems: we have reason to doubt that it really is the principle of practical judgement. But, as far as I can see, this leaves systematic universality, and its role in Groundwork II, from beginning to end, untouched.
Universality is not just an apparently central theme of Kant’s practical philosophy. It is the central theme of that philosophy. But in saying this, I do not mean the universality of the FUL. For we are not to reckon ‘among concepts themselves that which belongs only to the typic of concepts’ (CPrR, 5: 70). The true universality, the universality which belongs ‘among concepts themselves’, is, from beginning to end, systematic universality, that of the will of all, whereby we are many wills only insofar as we are one.
Acknowledgements
I extend thanks, for their generous feedback on this article, to David Finkelstein, Chandler Hatch, Jed Lewinsohn, Michael Pendlebury, and Mary Tjiattas. Thanks to Jacob Rosen for early peripatetic conversation in the Allegheny Cemetery and to Jeremy David Fix for having instigated my consideration of these topics and for many rounds of comments and discussion. And thanks to two anonymous referees who provided the opportunity for substantial improvements.
