Groundlaying: Kant's Search for the Highest Principle of Morality

Kant's division of philosophy into three sciences is perhaps a bit misleading to contemporary readers. First of all, readers will probably be puzzled by his labeling of ethics as a science; second, they will probably not be inclined to think of physics as a part of philosophy. These difficulties can be cleared up and readers can avoid being misled if they keep in mind that science in Kant's day and science in the contemporary world are two very different things. The practices that form the core of what we think of as science today were just taking shape in Kant's day and so science was then much less well-defined and broader in scope than it is today. In fact, it might be best to think of a 'science', as Kant uses the word, as referring to any organized body of knowledge. It is then easier to see how ethics could be a science, for we are able at least to distinguish claims to ethical knowledge from claims to, say, historical knowledge so that ethics seems to be organized into its own cluster of concerns. And again, if the science of physics was much broader in scope then than now, so that it included not only what we today call physics, chemistry, and biology, but also included the assumptions that these contemporary sciences take for granted, then it is much easier to see how the physics Kant speaks of can be a part of what we today understand to be philosophy.

With this division into three sciences, Kant is already in the first paragraph starting to make the distinctions that pave the way for what is perhaps the central theme of the book: that genuine morality must have its basis in pure reason alone and not, for instance, in feeling, sentiment, sympathy, compassion, religious faith, or in any mixture of the foregoing. Kant never comes right out and says what the principle of the division is, but the distinctions he makes in the subsequent pages between the material and the formal, and between the empirical and the rational or a priori, give some shape to such a principle. Attached to these distinctions are the notions of universality and necessity which allow Kant eventually to claim that moral judgments must be grounded in the rational and a priori rather than in the material and empirical. All of this starts in this very first paragraph, and it is not to be overlooked. If these divisions and distinctions do not hold up, Kant's moral philosophy collapses. If, for example, the distinctions between reason, which is rational, and feeling, which is empirical, cannot be maintained, then Kant cannot claim that moral judgments have their basis in reason alone, as there would then be no such thing as reason alone. But Kant thinks these distinctions do hold up, and he had already argued for them in his Critique of Pure Reason, the first edition of which was published in 1781, four years before the first edition of the Groundlaying.

Kant continues to make distinctions and divisions in order to set up the framework in which his argument will work itself out. There is little argument here, mostly just unsupported assertions. This is understandable since one has to start somewhere, and Kant has already discussed these topics at length in his other writings such as the already-mentioned Critique of Pure Reason. But of particular interest is Kant's claim that there are laws of freedom. By this phrase, he apparently means laws that govern the choices that rational beings such as humans may make. We do seem to make choices, choosing to take this course of action as opposed to some other course, choosing to decide this way or that way. But whether these are genuine choices is another matter, and philosophers have questioned whether we really are free to make choices. Kant, of course, is aware of these issues regarding freedom, and he tackles them explicitly in the Third Section. But that he recognizes and later deals with the issues does not give him a warrant to claim now, at this stage of the book, that there are laws of freedom. So it is worth asking what bearing this claim has on his subsequent argument: could the claim be removed with no adverse effect on the strength of the book's argument, or is the claim indispensable to such an extent that Kant's project cannot get off the ground without it? The first possibility, that the claim could be removed without adverse effect, might be supported by reading the Preface as merely background material to help the reader find the proper context for the arguments that are to follow in the rest of the book. On this reading, the Preface and any arguments in it may be jettisoned, the arguments in the remaining Sections being able to stand on their own, though the reader may then find it harder to get on track. To assess the merit of the second possibility, that the claim about the existence of laws of freedom is indispensable, we must note that Kant has so far used the claim to support the conclusion that material philosophy is twofold. As we move forward through the paragraphs ahead, then, we need to be on the lookout for either the premise (i.e., laws of freedom exist) or the conclusion (i.e., material philosophy is twofold) being used again, whether overtly or covertly.

The previous paragraph contained the first mention of the concept of a law. Kant there spoke of laws of nature and laws of freedom. But we learned nothing further about these laws, except perhaps that determinate or distinct objects are subject to those laws. In the current paragraph, we now get a bit more information about what is packed into Kant's concept of a law. Kant introduces a third kind of law, namely, laws of thinking (the previous paragraph having mentioned only rules of thinking). About these laws of thinking, we learn, in particular, that they are universal and necessary and that they cannot be grounded in experience. By calling them universal, Kant intimates that the laws hold for all thinking; by calling them necessary, we are to understand that the laws are demonstrable or can be established through formal proof; and by saying they cannot be grounded in experience, Kant means that they are not formulated or arrived at by drawing conclusions from observations of the world around us. So there are three kinds of laws: laws of nature, laws of freedom, and laws of thinking. Additionally, the third kind has been identified as universal, necessary, and not grounded in experience. These observations on the text strongly suggest that the laws of nature and of freedom, since they, too, are laws, are also universal, necessary, and not grounded in experience. Kant, then, would seem to be hinting to the reader that his moral philosophy is going to be rationalist (based on reason and thinking) rather than empiricist (based on experience and feeling) and that moral laws (i.e., laws of freedom) are universal (i.e., apply to everyone) and necessary (e.g., are not dependent on changeable states of mood, emotion, attitude, or feeling, and so on).

Kant has so far already made two broad divisions or distinctions within philosophy: a commonplace division by discipline or familiar area of study (physics, ethics, logic); a higher-level division by abstract subject matter (objects or no objects) into material or formal. We now get a third main division into empirical philosophy and into a priori or pure philosophy. This time the division is defined by the source of the principles on which the philosophy is based. The philosophy is empirical if the principles on which it is based are taken from experience. The philosophy is a priori or pure if the principles on which it is based are a priori or taken from reason alone. These alternate schemes of division overlap and criss-cross each other. For example, physics and ethics both belong to material philosophy, so that the material division overlaps the division by discipline.

With this paragraph, Kant reaches the end of the series of quickly introduced distinctions within philosophy. The distinctions have led up to metaphysics and in particular to the metaphysics of morals. At this point, Kant seems to want the reader to understand at least the following: the metaphysics of morals gets its principles from reason alone and not at all from experience; the objects that the metaphysics of morals deals with are objects of the understanding, not sensuous objects; the objects are governed by laws of freedom that are universal and necessary. In other words, the part of ethics that identifies and justifies moral principles uses reason alone to establish universal and necessary moral laws that govern people (not their bodies but rather their wills or, as it is put later in the book, the real person belonging to the world of the understanding).

An important question which the reader likely has at this point is what exactly the metaphysics of morals is. Unfortunately, Kant never discusses this directly at any length. What we know so far is that ethics has both rational and empirical parts, which Kant concluded after saying that the idea of a twofold metaphysics had arisen (not saying, though, precisely how), and that metaphysics is a pure, a priori part of philosophy. What, then, is the metaphysics of morals? One possibility is that it might just be the pure, a priori, or rational part of ethics. But this possibility leaves out the empirical part of ethics. Another possibility is that the metaphysics of morals is the pure, a priori, rational part of ethics as applied to the empirical part of ethics. This possibility brings in the empirical part of ethics, but not as the source of moral principles; the a priori part of ethics remains the source of moral principles, providing a pure, rational foundation for morals. But then this rational part of ethics is applied to, or in some way interacts with, the empirical part of ethics, giving rise to a metaphysics of morals. This latter possibility, then, takes the metaphysics of morals to be a system of moral knowledge that, though pure in its origins, still speaks to us as humans with physical bodies existing in a sensuous world. Later on in the work, Kant refers to pieces of such knowledge as synthetic a priori propositions.

Up to this point, Kant has made a handful of distinctions within philosophy, culminating with the distinction between the metaphysics of nature and the metaphysics of morals. He now wonders whether the nature of science, ethics already having been singled out as a science, requires that the empirical part be keep separate from the rational part. In the science of physics this means keeping its empirical part separate from and conceptually dependent on its rational part, so that the metaphysics of nature must conceptually precede empirical physics. In the science of ethics, this means keeping its empirical part separate from and conceptually dependent on its rational part, so that the metaphysics of morals must conceptually precede practical anthropology. In either case, the goal would be to see how much pure reason alone can achieve and from what sources reason draws its a priori teaching. All of this is put forth not really as a fact of any kind, but rather more like possibility or conjecture; Kant has made some distinctions, and he is now thinking about their role and significance, pondering what to make of a metaphysics of morals: if there is a metaphysics of morals, what does it do and how does it contribute to ethics, and how are we to treat or handle such a metaphysics of morals?

So far in the Preface, Kant has split up philosophy in various ways, one of these subdivisions being the metaphysics of morals. In the previous paragraph, he began wondering whether each of these parts within philosophy needs customized handling by a specialist or expert. In the current paragraph, he makes a refinement, limiting his concern only to moral philosophy. He asks, in particular, whether it is important to work out the part of moral philosophy that is pure, that is cleansed of anything empirical. He seems to think that it is important to do this, for, he says, it is evident of itself from the common idea of duty and of moral laws that there must be such a pure moral philosophy. By this he seems to mean that by reflecting on nothing more than these ideas of duty and moral laws, we can directly know that there must be such a pure moral philosophy; when we reflect on these ideas we find something in them that clearly shows the existence claim to be true. Kant, unfortunately, does not elaborate; he does not say what exactly these ideas are, how he knows they are common (whatever that might be), or provide any insight into the means by which from these ideas we get to the existence claim. But he does add on a number of statements or examples apparently in favor of the existence claim. We should take note of these statements, which he alleges everyone must admit, and of their constituent elements, reflect on them, and bear them in mind as we go along. To help you get started, I offer the following.
A moral law is a ground of obligation. What makes a law a moral law? Is it that it is a ground of obligation? What is a ground, an obligation, and a ground of obligation? Is there an important difference between saying 'ground of obligation' and 'ground of an obligation'?
A moral law must carry with it absolute necessity. What is this notion of 'carrying with'? What is this necessity, and what more is added by preceding it with the absolute qualifier?
'You ought not lie' is an example of a (moral) command that applies to all rational beings. How does a moral command differ, if it does, from a moral law? What rational beings are there besides humans? What makes a being, whether human or not, rational in the first place?
The ground of obligation for a moral law is to be found in a priori concepts of pure reason. How can concepts alone be a ground of obligation? Even if human nature cannot provide a ground, is there no other possible ground other than concepts? And why must the concepts be a priori? Will only concepts of pure reason, rather than of some other kind of reason, do?
Prescriptions can be practical rules or moral laws. What are the differences and similarities between these?
Universal prescriptions can be based on empirical grounds. How can the empirical be a basis for the universal? Might Kant only mean 'general' here instead of 'universal'? (The German word is 'allgemeine' at viii.20 (4:389.20-1).)
The motive for a universal prescription might be empirically grounded. (Again, is 'universal' the appropriate adjective?) What is it for a motive to be grounded or based on empirical grounds? Are any other grounds available for motives?
A prescription based to any extent on empirical grounds is a practical rule, not a moral law. Why must there be a total separation from the empirical?

It emerges from this paragraph that even though a part of moral philosophy is pure and a priori, not based on experience but rather on pure reason, there is still a role for experience to play. Kant singles out two areas, both related to implementation, in which experience is still needed: the application of the moral laws, and the acceptance or adoption of the moral laws. So Kant's thinking would seem to be that the pure part of moral philosophy is to identify and establish (or justify or defend) the moral laws in an a priori fashion, using only pure reason and no experience. These tasks of identification and justification having been accomplished, the pure part of moral philosophy can then, so to speak, hand over these moral laws to the non-pure part of moral philosophy (e.g., practical anthropology). It is only at this point, after the hand-off, that knowledge about humans (e.g., their desires, abilities, psychology, etc.) and their circumstances (this knowledge being supplied by practical anthropology, not by morals proper) can be used, for instance, to figure out how to apply the laws in particular situations and how to work the laws into the wills of humans beings. In the Groundlaying, Kant has very little to say about these tasks of non-pure moral philosophy. As he will soon say later in the Preface, his purpose in writing the book is to find and establish the highest principle of morality, and these are the province of pure moral philosophy, province of the preliminary work toward a metaphysics of morals. In a much later work, The Metaphysics of Morals of 1797, he has more to say about the non-pure part of moral philosophy. (Presumably, the reason why this later work is called The Metaphysics of Morals, even though it deals also with the non-pure part of moral philosophy, is that, as Kant says in this paragraph, moral philosophy rests completely on its pure part.)

In this paragraph, Kant finishes his defense of the need for a metaphysics of morals. He has concluded that it is necessary, and for both theoretical and applied reasons. This paragraph contains the first sustained argument in the work, and it is striking in that it is not very good. For example, the basic premise upon which the argument builds itself, that the unmoral ground now and then produces actions in conformity with law but more often produces actions not in conformity with law, can easily be challenged. It is not wholly clear what Kant means by an 'unmoral ground', but if we neutralize this expression so that it means acting for reasons other than for the sake of the moral law, such as acting out of fear of immediate and severe punishment, then one could plausibly claim on firm empirical grounds that unmoral grounds do actually frequently produce actions that are in conformity with (moral and non-moral) law. Fear of timely punishment is quite an effective means of bringing people into line. It might be suggested that this argument occurs in the Preface and so is not to be subjected to the same level of critical scrutiny as the arguments later in the work. This may be so, but even the arguments later in the work often are not carried out with sufficient care. So this argument in the Preface is not atypical. I believe, in fact, that the arguments in the work are often too short, sketchy, and incomplete, and are consequently obscure and exasperating. So the work is not, I think, considered a classic for the quality of its arguments. Rather, it is the ideas and concepts, the point of view and the vision of the work, that set it apart. This combination of being too short and yet being visionary is what makes the work endure, for the vision it projects captures our imaginations and its brevity leaves many holes for us to fill.

Wolff (1679-1754) was a well-known philosopher who was active a generation before Kant, whom Kant respected and whose textbooks Kant actually used in his own classroom teaching. Kant's depiction of Wolff's philosophy is accurate; the noted differences do exist, and Kant's moral philosophy as a whole is indeed something new.

On a curious note, this is perhaps the appropriate place to point out that Wolff died of the gout, and Kant speaks at 12.17 (4:399.17) of a person suffering from gout.

Even if there is no argument in this paragraph, it does contain some valuable information. We learn, for instance, that Kant thinks the best foundation for a metaphysics of morals is a critique of pure practical reason. Given this admission, it is interesting that the Groundlaying has probably been the subject of more scholarly interest than the Critique of Practical Reason. Perhaps Kant's explanation in this very paragraph as to why he has not written a critique — because to bring such a critique to completion would require bringing in considerations that would confuse the reader — also explains this uneven distribution of scholarly interest. In any case, we seem to get a better picture of how Kant's three major works in moral philosophy fit together or were intended to fit together: the Groundlaying is a preliminary study aiming toward a critique of pure practical reason; this critique, actually entitled the Critique of Practical Reason in 1788, really provides the foundation for a metaphysics of morals; and the Metaphysics of Morals, published in 1797, depends on the two previous works to the extent that the moral laws and principles that it uses have their origin in pure practical reason.

Kant offers in this short paragraph a third explanation as to why he has written the Groundlaying. He explains that he did not want to weigh down a metaphysics of morals with the subtleties that are unavoidable in a preliminary work like the Groundlaying; including these subtleties would jeopardize the potential popularity of a metaphysics of morals, and so he has put the subtleties in this separate, preliminary work. I have wondered why Kant put this third explanation in its own paragraph rather than grouping all three explanations in the same paragraph. It cannot be that the previous paragraph would then have been too long; for, judging from the rest of the book, Kant does not seem overly concerned about the length of his paragraphs and the sentences in them. Instead, I think that the separation into multiple paragraphs is due to a slight difference in what Kant is explaining. In the previous paragraph, Kant explained why he has written the Groundlaying instead of a critique of pure practical reason. In the current paragraph, on the other hand, Kant explains why has has written the Groundlaying instead of a metaphysics of morals.

In this paragraph, Kant gives us his clearest statement of what he is up to in the Groundlaying. He says that the Groundlaying is nothing more than the search for and the establishment of the highest principle of morality. The 'nichts mehr' ('nothing more') at xv.4 (4:392.3) apparently refers to the absence of applications of the principle to the actual circumstances of human life; for he has already made it clear that a metaphysics of morals is to be unmixed with anything empirical and that discussing applications would require bringing in empirical facts about humans. In other words, it would be an additional task beyond a metaphysics of morals to figure out how to apply the highest moral principle which such a metaphysics had previously found and established.

There are three more items to note in this statement of purpose. First, we learn that there is only one highest moral principle. At the highest level, then, there is a single principle, not a plurality of principles. Note, however, that this does not rule out there being, at a lower level or levels, multiple principles which might conflict with each other, the single highest principle presumably ultimately resolving any such conflicts.

Second, the phrase 'search for' might mislead. The phrase can suggest an exhaustive, wide-ranging hunt for something so that you might expect Kant to scour the entire moral landscape for clues which then lead to other clues which then allow him finally to home in on the highest principle. But this is not at all what happens in the Groundlaying. It takes Kant scarcely any time at all to zero in on what he's looking for, the principle being something that we, in a way, already know and use so that it needs only to be recalled and clarified rather than to be discovered (see pages 20-21 (4:403-4)).

Third, the 'Festsetzung' at xv.4-5 (4:392.4) or 'establishment' of the principle is vague. This might mean any of several things, among them: (1) proving the principle; (2) justifying the principle; (3) explaining the principle; (4) describing the operation of the principle; (5) securing a prominent place for the principle in the wills of individual humans; (6) setting the principle up as the core of a social norm designed to guide behavior. My own view, which I will not argue for at length here, is that by "establishing" the principle Kant really means something closer to (2) justifying the principle. In particular, my view is that Kant is defending the principle from, or justifying the principle in defiance of, competing principles (such as the principle of benevolence) and the rising threat from empirical science. He is not doing (1), (5), or (6) because Kant believes that the principle is already known and being used so that there is no need to prove it or to work it into the individual and social practices that it already shapes. And he is doing more than just (3) or (4), for he does give arguments, which would not be needed if he were just explaining or describing, and the arguments do not seem to be for, or in the service of, a particular explanation or a particular description. So Kant aims to establish the highest principle in the sense of shoring it up against attack, enabling it to defend itself with rational and reasoned considerations instead of only with considerations drawn from custom, habit, and tradition.

The Preface began with a tripartite division, and Kant wraps up the Preface with a brief note on the methodology he has used in the remainder of the book and announces the three book sections that apparently result from the adoption of that methodology. It is again unfortunate that Kant does not elaborate, for there are a number of questions that immediately come to the fore when reading this paragraph. First, why is this methodology the most appropriate or why was it selected over some other methodology that might have been used instead? Second, what exactly is the methodology? He says that it is first analytic and then synthetic, and there seems to be going-to and returning-from aspects to the methodology. At the end (pages 95-6 (4:444-5) of the Second Section, Kant says that the first two sections have been analytic. So, at that point, perhaps we will be able to look back at what he has done, and how he has done it, in order to figure out what Kant means by the analytic part of the methodology. And presumably, then, the Third Section of the work uses the synthetic part of the methodology so that by the end of the work we will know what this is as well. Finally, why do these specific sections result from choosing the methodology? If some other methodology had been chosen, would there be a different number of sections or differently titled sections?

The second of Kant's possible conclusions in this paragraph is weakly phrased: the good will appears ('scheint' at 2.11 (4:393.24)) to constitute the worthiness to be happy. This is not so with the first conclusion; it is strongly phrased, even bold in its pronouncement. The first conclusion, that only the good will is good without qualification, therefore, is going to be very difficult to establish. There are actually two parts to it: the first is that the good will is good without qualification, and the second part is that there is nothing else that has this kind of goodness. The first part claims that the good will has a certain property or, alternately (if you do not want to talk about properties), exists in a certain way. The second part claims that the good will is unique in having this property or in existing in this way.

There is already much to discuss here in this first paragraph, for instance, what it is to be good without qualification. But what is most striking is Kant's strategy. He does two unexpected things in this opening paragraph. First, he talks about the good will when the reader might expect him to talk about the highest principle of morality, the search for and establishment of this principle being the stated purpose of the work. Second, he seems to want to support the uniqueness claim (i.e., that only the good will is good without qualification) by the enumeration of categorized (e.g., talents, temperaments, etc.) examples.

The explanation for the first unexpected strategy does not become evident until later in the work. Kant discusses the good will first because, as we learn shortly, for Kant morality has more to do with willing, intention, and motive than with the effects or consequences of our actions. When Kant finally gets around to it (which actually does not take very long), the principle of morality turns out to be a principle of willing in a certain kind of way. With this understanding of what is to come, this part of Kant's strategy in the opening paragraph makes more sense.

The explanation for the second unexpected strategy is more elusive. Kant's attempt to establish the uniqueness claim by enumeration is doomed to failure because he would have to run through all examples in all categories until only the good will is left standing as the single thing that is good without qualification. Perhaps he was hoping that by using categories his argument could be exhaustive, but he gives no indication that he thinks he has covered all possible categories. So it is puzzling why Kant takes this futile approach. There are much better ways to try to establish a uniqueness claim. One such way, which mathematicians and logicians use with great frequency, would be to derive a contradiction from an assumption that the uniqueness claim is false (i.e., that there is something else, too, that is good without qualification); then, since a contradiction cannot be true and the contradiction depends on the assumption, the assumption must be false; hence the uniqueness claim is true. Kant was also a mathematics instructor and used this proof by contradiction strategy elsewhere in his writings; so it is remarkable that he does not use something like it here. Perhaps he was unable to figure out how to derive the contradiction. Some other possible explanations: he wanted to start the work with a bold statement that would grab the reader's attention; he thought the claim is rather obvious and that nearly everyone would agree with him about it; the claim does not figure importantly in his subsequent argument so that there is no need to argue for it conclusively; the claim is to be taken as an assumption or as otherwise provisional in some way.

In this paragraph and the previous one, we seem to have learned at least all the following about a good will: a good will is good without qualification; a good will is the only thing that is good without qualification; a good will can make corrections and adjustments; it can also be pure; it is the indispensable condition of the worthiness to be happy; it can be helped by qualities; it has an inner unconditioned worth; and it cannot become bad. Kant has given very little supporting argument or even explanation for these claims; so it might be best to view them as his starting points, his assumptions, to get the ball rolling.

The parenthetical text (3.17-9; 394.23-4) in this paragraph is important. In it, Kant makes clear that the willing through which a good will is good cannot be equated with wishing or, at least, ordinary or mere wishing. To be willing of a kind that interests Kant (i.e., willing that is morally relevant), the willing has to call upon all the means that are available, all the resources at one's disposal. It cannot be half-hearted willing, or insincere willing, or willing that uses less than all of one's will power. This parenthetical text, however, is in the middle of a passage about a will that is ineffective in achieving results and yet is a good will. So what we are being told is that beneficial results are not necessary for a will to be a good will. Kant might, in fact, be claiming something more: that results of any kind, beneficial, harmful, or neither, are not necessary for a will to be good. But, with this parenthetical text, Kant seems especially interested in emphasizing that, when the results are otherwise than beneficial, the willing must have been complete. It is less clear whether Kant wants to say that the willing must always be complete. If the results had turned out to be beneficial and the willing was only partial (i.e., not all means available had to be invoked), is that kind of willing sufficient to constitute a good will? Perhaps it is best to read this paragraph as addressing only the question of what is not necessary for a will to be good, leaving to the side questions of what is necessary and of what is sufficient. If we do this, we get the clearer reading that beneficial results are not necessary for a will to be good.

Some readers might wish that Kant had said more about why the idea of the absolute worth of a good will is so very strange. I think, however, that asking for more detail about the strangeness is not at all the question to ask. This allowance that the idea is very strange is, in my view, mainly a front, a pretense. Kant does not really think the idea strange at all. He even says that we all agree with the idea; so it must not be all that strange to us, either. The strangeness claim is really just a ruse that Kant uses to ensnare the empiricist in her own trappings. The better question to ask is why Kant wants to conduct this test of the idea. Conducting this test is the real objective; he just uses the strangeness claim as motivation. But it is not just the test that Kant is after, for he says that he is going to put the idea of a will having absolute worth to the test from a particular point of view. The point of view is apparently the point of view that considers nature to have assigned reason to guide the will. So why is it so important to Kant to conduct the test, and why is it to be done from this point of view rather than from some other point of view? The answer is tied up with Kant's rejection of the principle of happiness as the highest moral principle; for in the following paragraphs Kant contends that if reason guides the will then happiness cannot be the goal of nature. So it is important to Kant that he get reason, in its capacity as a guide for the will, inserted into the discussion early; by doing so, he will be able, without delay, to start undermining the principle of happiness.

Kant is in the middle of his test of the strange idea of the absolute worth of the good will. He is going to conclude that the idea is consistent with the view that holds that nature has assigned reason to be the administrator of the will. This paragraph helps Kant reach that conclusion by arguing that if the happiness of rational beings with a will were the, or at least a, goal of nature, then nature would have chosen instinct to guide the will. It would have chosen to set things up this way because nature always chooses the most appropriate means to achieve its goals and instinct would have been a more appropriate means than reason. Having established this claim that nature would have chosen instinct, Kant can then note that reason, not instinct, has in fact been chosen; accordingly, happiness cannot be the goal of nature. What, then, could be the goal? Kant's answer will be that the goal is the production of a good will, thus showing the consistency of the idea of the absolute worth of a good will with a particular conception of nature.

Kant's strategy here is rather clever. He begins the section with the idea of a good will, develops the idea a bit, allows that the idea of the absolute worth of a good will might be thought to be a strange idea, and then goes on to argue that the idea is consistent with a certain view of the workings of nature. Some readers might think that in discussing the purposes of nature Kant has gone off on a tangent, taking a detour into a side issue. But I think not. What he is doing, admittedly in an indirect way, is bringing nature around to his side of the argument. Nature and the natural world in general are generally considered to be more easily exploited by empiricists who argue for a principle of benevolence or of happiness. So, if he is successful in bringing nature around to his side of the argument, he can undermine one of the major supports of empiricist thinking about morality and in so doing defend or insulate his own rationalist theory from empiricist attack.

Kant is still conducting his test of the idea of the absolute worth of a good will. In the previous paragraph, he contended that reason is a poor choice to be the administrator of the will if nature's purpose is the happiness of rational beings. In this paragraph, Kant gives a couple more arguments for the same contention. The first argument tries to make the point that, if reason selects happiness as an explicit goal to be achieved and pursues that goal directly, then the goal will likely not be achieved. It will actually be better, so the contention goes, to rely on instinct rather than reason. The second argument points out that, as a consequence of the failure of reason to achieve the goal of happiness, many people will come to hate reason. So Kant has now lined up several arguments to establish that reason is a poor choice to be the administrator of the will, if the happiness of rational beings is nature's purpose. The arguments are similar but also are slightly different from each other. All of them involve comparison. The first, from the previous paragraph, argued that instinct prescribes more precisely than reason does the actions to perform in order to secure happiness. So it was an argument for the relative superiority of instinct over reason. The arguments in this paragraph rely on a comparison of the relative amounts of trouble versus happiness generated and on a comparison of different strata of people and their differing levels of adherence to the dictates of reason or to the impulses of instinct. The arguments diverge, however, in that the latter two are, at least as Kant saw them, based in fact, twice using the phrase 'in der That' ('in fact') at 5.21 and 6.7 (4:395.28 and 4:395.37).

In this paragraph, Kant finishes up his test of the strange idea of the absolute worth of the good will, which began two paragraphs ago. He concludes in this paragraph that nature has, with full consistency, made reason the administrator of the will. In particular, he concludes that practical reason (which comes along with reason because, as we learned in the Preface, there is ultimately only one reason) has been given to human beings in order to produce a good will in oneself (and perhaps as well in others) rather than happiness. Kant never explicitly says that the strange idea has passed the test. But it evidently does pass because Kant does not now abandon the idea; rather, he continues to develop it further, arguing that the good will is the highest good even if not the only and whole good. So what test exactly has the idea passed? He said the test was to be conducted from a certain point of view, apparently the point of view that considers nature to have assigned reason to be the administrator of the will. And now he has concluded that nature's purpose in this assignment is to produce good wills; in addition, he has concluded that nature has acted with full consistency in making this assignment even though happiness is not maximized and inclinations are unsatisfied. The idea of the absolute worth of the good will, then, has passed the test in that the production of a good will is consistent with the purposes of nature. Kant's larger goal has been achieved, too. He has managed to bring nature over to his side, the rationalist side, of the argument and turned it against the empiricist side, against the unrestricted pursuit of happiness, against the unrestricted fulfillment of desires, and against the advocates of the principle of benevolence as the highest principle of morality; for the implication is that if nature is consistent in using reason to enable the creation of good wills and this creation involves abandoning the single-minded pursuit of happiness, then it would be inconsistent to align nature with the principle of happiness or benevolence as the empiricists try to do.

By this time, Kant has finished introducing the concept of a good will. He has found so far that we must think of the good will as good without qualification and as good merely through its willing rather than through what it might accomplish; and he has also found that this conception of a good will is consistent with a conception of nature according to which reason is to help produce a good will. He is now ready to begin a more searching investigation of the concept. He hopes to conduct this search indirectly, by investigating the concept of duty. He says the concept of duty contains, under certain limitations and obstacles, the concept of a good will. He seems to think that these limitations and obstacles will actually help the investigation along, and perhaps that is why he decides to continue his investigation of the good will by investigating duty. In this paragraph, we also learn that: the concept of a good will already resides in the natural sound understanding and requires to be clarified more than to be taught; the concept of a good will is always the most important consideration when evaluating the whole worth of our actions; the concept of a good will constitutes the condition of everything else.

As usual in the Groundlaying, Kant says very little by way of explaining what the claims in this paragraph mean. He says, first of all, that the concept of duty contains the concept of a good will. What this means and why Kant thinks it true will become somewhat clearer after later he gives his more formal definition of the concept of duty at 14.14-5 (4:400.18-9). The basic idea seems to be that duty is an obligation to do what one ought to do and that a good will is the kind of will that wills to do what it ought to do, so that when you analyze or look into the concept of duty you find the embedded concept of a will willing what it ought, that is, you find the concept of a good will. But the two concepts are not identical. There is more to the concept of duty than just the concept of a good will. In particular, there is the bit about obligation. Duty also involves an obligation, so that the good will that recognizes its duty feels obligated to will to do what it ought to do. Such a good will that recognizes its duty feels compelled or constrained in some sense —necessitated in Kant's jargon — to will as it does. It is necessitated in this way because such a good will, one that recognizes its duty, is not perfect and so does not automatically will to do what it ought. It is to these imperfections that Kant alludes when he speaks here of certain subjective limitations and obstacles.

Another of Kant's claims in this paragraph is that the concept of a good will already resides in the natural sound understanding and requires to be clarified more than to be taught. This kind of claim is a hallmark of rationalist approaches to moral philosophy and to philosophy in general. The basic idea here is that moral beliefs and standards are already in us in some shape or form and so do not need to be taught; all we really need to do is figure out how to draw them out, give explicit expression to them, so that we can become more assured of what the correct moral standards are. This rationalist line of thought goes all the way back to Socrates and Plato. In fact, there is a Platonic dialogue called the Meno which deals with the topic of whether virtue can be taught. Plato is in general much easier to read than Kant (although his writings still give rise to just as much intellectual confusion and difficulty), and if you have not read the Meno yet you should do so at once: being familiar with it will help you understand where Kant is coming from.

In this paragraph, Kant begins his investigation of the concept of duty. To conduct the investigation correctly, he first of all wants to make sure that he has at his disposal some clear-cut cases of duty. To ensure this, he passes over and sets aside some cases that are not cases of acting from duty. In particular, he distinguishes in this paragraph the following cases:
1. actions that are contrary to duty;
2. actions in conformity with duty but that are performed because of an inclination that is not immediate;
3. actions in conformity with duty and that are performed when an immediate inclination to perform the action is also present.
Of these three types, Kant dismisses or sets aside the first two and will not discuss them since they are not actions from duty. He thinks that it is in the third type of case that it is difficult to distinguish between conforming actions done from duty and conforming actions done for a self-interested purpose. Even though it immediately follows the description of the third type of case rather than the second, the shopkeeper example is of the second type of case. The example is apparently intended to show that cases of the second type are, relative to those of the third type, easy to spot and so can be set aside.

There are two examples in this paragraph. Both are examples of the third type of case (i.e., conforming to duty and with an immediate inclination) described in the previous paragraph. In both, the action is to preserve one's life, and this action is in accord with duty; additionally, because everyone has it (9.22 (4:397.33-4)), there is in both an immediate inclination to preserve one's life. In the first example, people preserve their lives just because they have an immediate inclination to do so; accordingly, the maxim of their action has no moral content. In the second example, an unhappy person preserves her life; and, although she, as does everyone, has an immediate inclination to preserve her life, she does not act from that inclination. She acts instead from duty. It is this, that she acts from duty rather than from the inclination, that allows the maxim of her action to have moral content.

It might be wondered whether the person in the second example really has an immediate inclination to preserve her life. From Kant's description of her psychological state, it might be thought that her immediate inclination is to kill herself. But perhaps it is possible to have multiple simultaneous immediate inclinations and to act only on a subset of those that are present or, as in the second example, to act on none of them. In the second example, then, Kant might contend that she has both an immediate inclination to preserve her life and an immediate inclination to kill herself and that she acts from neither of those inclinations but rather from duty alone.

As a final comment, 'maxim' figures importantly, and appears for the first time, in this paragraph, but Kant does not say what it means. He rectifies this oversight later in the footnote on page 15 (4:400). It should also be noted that Kant speaks here of maxims having, or being able to have, moral content; actions, though, can have moral worth; meanwhile, psychological states such as anxious care can lack inner worth.

Kant is still working through his distinction between actions done from duty and actions done from self-interest (or at least done not from duty). This distinction has an impact on the moral content of maxims and then on the moral worth of the action or actions specified by the maxims. The three examples that he gives in this paragraph help to establish the connections between the distinction, the moral content of maxims, and the moral worth of actions. In all three examples, the action is an action of beneficence; so the action is in conformity with duty. The differences between the examples have to do with whether the action is done from duty or done from inclination and under what conditions.

In the first example, the person acts beneficently because she is naturally disposed to do so, because she is by nature a friend of the human being; her action has no true moral worth because she acts from inclination attached to inner pleasure and delight. It is, however, not clear from this example whether Kant wants to go so far as to say that the action has no true moral worth only because the action is done only from inclination. Kant only rules out motives of vanity or self-interest; he does not design the example to rule out a motive of fulfilling one's duty. So it is possible that a person in this example might act from both duty and inclination. In such a case, the example would then suggest that an action has no true moral worth if it is done from inclination no matter how little the inclination contributes to the motive.

In the second example, we again have a friend of nature, but this time her natural inclination to help others is blocked or neutralized in some way by her own concerns; so she performs the beneficent action only from duty, and this action has genuine moral worth. From this example, it is clear that Kant wants to say that for an action to have true moral worth it must be done only from duty; it cannot be done from both inclination and duty at the same time.

The third example presents a case in which the person is not a friend of the human being; she has no natural disposition to be helpful to others. So the difference between this example and the second example is this: in the second example, the person, having a natural disposition to be helpful to others, is a friend of the human being, but the inclination to help others is temporarily disabled by her own concerns; in the third example, there is no disabling of the inclination because, lacking the natural disposition, she is no friend of the human being, and the inclination was never present in her heart. Though there is this difference between the second and third examples, the actions described in them have moral worth because they are done only from duty.

Toward the end of the paragraph, there is the slightest hint of the possibility that Kant thought that the action in the third example has a higher moral worth than the action in the second example. This possibility is suggested by the question Kant asks and by the emphatic affirmative answer that he gives in reply to the question. The question suggests this possibility because it speaks of a worth higher than that of a person with a good-natured temperament, and the person in the second example might be such a person. If this is the way to interpret the question, then the emphatic affirmative answer declares that the worth is higher. But it remains possible that the question should not be interpreted in this way and that the emphatic answer merely signals that Kant thinks this third example provides only an even clearer illustration of acting from duty than does the second example.

Kant gives us yet another example of action that accords (in a sense) with duty and that, only if done from duty as well, has moral worth. The example of the gout sufferer is not as straightforward as the previous examples, and Kant might have nevertheless included it in order to further undermine the principle of happiness. The example is not as straightforward because it is not a settled matter whether the gout sufferer does act according to duty, in which case the example would not fit the pattern of an action that accords with duty and is done only from duty. Some advocates of the happiness principle might argue that the gout sufferer, by consuming the food now, is taking the correct action to promote her happiness because she enjoys the moment only at the expense of the mere possibility of future benefit. But other advocates might claim that she actually acts to diminish her happiness because by succumbing to her present desires she very likely loses a much better, healthier, future. So the indeterminate nature of the concept of happiness (tied perhaps to incomplete knowledge of the future) makes it impossible to settle whether the gout sufferer is really acting in favor of her own happiness and thus whether she is acting according to duty. Although the example is not straightforward because of this inadequacy in the concept of happiness, Kant still manages to make his main point: if the gout sufferer is acting in favor of her happiness (whatever that might be) and thus according to duty, then her action has moral worth only if it is done from duty.

In this paragraph, Kant gives us his final example of action that is in conformity with duty and that has moral worth only if done exclusively from duty. There is a new element in this example of loving one's neighbors and even enemies, though. Kant says that beneficence from duty can be commanded not only when there is no inclination for such action but also even when there is disinclination for such action. In none of the previous examples was the possibility of disinclination mentioned, at least not explicitly.

There are a number of puzzles about this paragraph. First, the paragraph's opening announcement of the second proposition makes the reader wonder what the first proposition was. Kant never made an announcement about a first proposition (but he will again with regard to a third proposition). So the reader has to figure out for herself what the first proposition was. Reflecting on the previous paragraphs, the most likely candidate for the first proposition is something like the following: an action has moral worth only if it is done from duty and not at all from inclination. But, again, it is not clear whether Kant wants to assert a biconditional so that being done from duty is both necessary and sufficient for the action's having moral worth and whether he wants to exclude all traces of inclination from the motivation for the action.

Another puzzle is what exactly the second proposition is. The sentence in which the announcement of the second proposition occurs actually contains several assertions. These assertions are: an action from duty does not have its worth in the purpose that is to be achieved by the action; an action from duty has its moral worth in the maxim according to which the action is decided; an action from duty does not depend on the actuality of the object of the action; an action from duty depends merely on the principle of willing according to which the action is done. Which of these assertions, or which combination of them, is the second proposition? Complicating matters further, there is an 'also' ('therefore') at 13.17 (4:399.37) in the announcement sentence, suggesting the presence of an inference; so we might wonder whether the "second proposition" is a single proposition, several propositions, or an argument.

A third puzzle is why Kant goes from speaking of the 'Princip des Wollens' ('principle of willing') to speaking of the 'Princip des Willens' ('principle of the will') and then back again to the former.

Finally, there is a puzzle about the arguments in this paragraph. In particular, why does Kant give two separate arguments for the same conclusion? The first argument takes a more direct approach to concluding that the moral worth of an action from duty lies in the maxim of the action; the second argument is less direct, making use of what determines the will. Are both of these approaches equally good in terms of strength of argument? Is one more informative than the other? Or, since the first argument is supposed to be clear from what has gone before, perhaps Kant added the second argument in order to be a bit more specific about what parts of what has gone before should be used in arguing for the second proposition.

Kant starts the paragraph by saying that the third proposition, that duty is the necessity of an action from respect for the law, is a consequence of the previous two propositions. The two previous were: (1) an action has moral worth only if done from duty; (2) the moral worth of an action lies in the maxim by which the action is decided. The third proposition clearly does not follow as a direct logical consequence of the first two propositions, for it contains some concepts, such as necessity and respect, that are not even present in the two previous propositions. So it would seem that either Kant does not intend the third proposition to be a strict logical consequence of the previous two propositions, or there are some additional premises at work behind the scenes. Perhaps Kant makes some of these premises explicit in the present paragraph. But the brief analysis (above) of the argumentative structure of the paragraph reveals that quite a lot is going on and that there is going to be no quick and simple way to piece it all together. So I will not attempt that here. What I will do is point out a few connections between the claims made in the paragraph, and these connections might provide some clues as to how the argument might be understood and assembled.

One of these connections is between the activity of a will and the law. Kant's first reason for claiming that one cannot have respect for an object as an effect of one's proposed action is that the object is an effect and not an activity of a will. Later in the argument, Kant reaches the intermediate conclusion that only the law can be an object of respect. So it looks like that, if an object of respect has to be an activity of the will and the law is the only object of respect, then the law must be an activity of the will. The existence of this connection is borne out by Kant's later discussion of the autonomy of the will, the will's activity of legislating for all rational beings.

The second connection I will point out is between the concepts of exclusion and separation. One of Kant's premises is that only what excludes inclination can be an object of respect. Another premise states that an action from duty is to be separated from the influence of inclination. We already know that the law is the only object of respect; so it must be what excludes inclination. Perhaps, then, the law is also what separates an action done from duty from the influence of inclination.

The third and final connection I will mention is between the concepts of necessity and determination. Necessity is an important component of Kant's characterization of duty in the third proposition. And one of Kant's conclusions in this paragraph is that all that remains that can determine the will is the law. We already know (from the previous paragraph) that the will must be determined by something. Putting these together, we have that the will must (i.e., necessarily) be determined by the law when an action from duty is done.

These are at least some of the connections that need to be exploited in order to assemble Kant's argument for the third proposition. It would, however, take many pages to fit the pieces together more tightly — a good project for an undergraduate senior thesis.

As usual, the pieces of the argument in this paragraph do not quite fit together. In particular, the biggest gap in the argument here seems to be in the move from the highest and unconditional good to the moral worth of an action. How, exactly, does Kant bridge this gap? We already know (from pages 1 and 7), and Kant reminds us here, that the highest and unconditional good is to be found only in a good will, in the will of a person who is acting from duty. And we are told here that a will is not needed in order to produce the effects of actions. Putting these together suggests that the highest and unconditional good is not needed in order to produce the effects. So there is a kind of separation between the highest and unconditional good and the effects. How does this separation lead to a corresponding separation between moral worth and the (expected) effects? To fill this gap, Kant seems to need some claim establishing the highest and unconditional good as the sole source of moral worth. With such a claim in place, Kant can then claim that if effects are separated from the highest and unconditional good then they are also necessarily separated from moral worth because the moral worth can come from no other source than the highest and unconditional good. Such a claim, or something close to it, is in fact already available to Kant; it is the second proposition: the moral worth of an action from duty is only in the maxim of the action, only in the principle of action used by a good will when deciding to perform the action. The second proposition, then, might be what Kant is implicitly using here to plug the main gap in the argument in this paragraph.

This paragraph, which continues to develop the previous paragraph's introduction of a representation of the law, begins with a question that asks for information about a particular kind of law. The question itself helps to narrow down the possible answers, for the question itself specifies that the law is of a kind such that the representation or thought of the law must determine the will if the will is to be called good absolutely and without limitation. Although Kant does not give a simple and direct answer to the question, the kind of law that emerges as the probable answer is a universal law, a law that applies to all rational beings everywhere and at all times. So it is the representation or thought of this kind of law, a universal law, that must determine the will if the will is to be called good absolutely and without limitation. That this is the answer to the question emerges in the argument by elimination that immediately follows the question. In previous paragraphs, Kant has eliminated inclination as a moral incentive. Here he goes a step further and eliminates incentives that might arise from following a particular law. All non-moral incentives have thus been eliminated. Kant's reasoning in this argument then is that, because all incentives that might determine the will have been taken away and because either such incentives or a principle of conformity to law in general must determine the will, only the principle remains to determine the will. This principle of conformity to law is the following: I ought never proceed otherwise than in this way, that I can also will my maxim should become a universal law. So the principle, in specifying the kind of law, answers Kant's opening question.

Having the answer to the question is by itself not all that helpful. It would be more helpful to know why this is the answer: what considerations prompt Kant to give this particular answer, and why does Kant ask, in the way that he does, the opening question in the first place? To achieve this fuller understanding of what is going on in this paragraph, we need to remember and make use of what Kant has already taught us. In particular, Kant has already (in the previous paragraph) argued that the representation of the law constitutes the moral good and that (and this is the second proposition) moral worth of actions is only to be found in the maxims of a good will, the willing of these maxims being what makes the good will good. So it seems that these maxims are the way in which humans represent the law to themselves. But, as Kant has also been arguing, these maxims cannot be based on inclinations or any other empirical incentives. All that remains, then, that can determine to will is the law itself. So these maxims must be willed as universal laws if the will is to be a good will. What I have just outlined is perhaps Kant's train of thought in this paragraph. He is basically trying to figure out what makes the good will tick, so to speak, and he has found that it is this principle of willing maxims as universal laws.

There are a few other items of note in this paragraph. First, Kant claims that the principle he specifies is in fact used by ordinary human reason, which always has the principle before its eyes. Second, ordinary human reason completely agrees with the principle. Third, the principle, in its expression of universal conformity to law, not just does, but also must, serve the good will as its principle; the principle is necessary if morality is real. This last is very important to note, for it reveals that Kant is not yet committed to the claim that morality is real rather than just a fantasy. All he is claiming so far is that if morality is real, then this principle is what the core of it must look like.

In this long paragraph, Kant gives an example of making a promise that one does not intend to keep. The example is apparently, judging by how the surrounding paragraphs end and begin, supposed to illustrate several things: the application of the principle of willing a maxim as a universal law, which Kant introduced in the previous paragraph; that ordinary people actually use something like this principle in their moral thinking; that no special training is required in order to use the principle.

Kant starts wrapping up the initial phase of his search, much in this paragraph succinctly repeating what he has already said. The paragraph opens with a conclusion that no special ability is required in order to take the steps required in order to ensure that one's willing is good. This conclusion immediately follows paragraphs in which Kant has claimed that ordinary human reason agrees with and makes use of the principle of willing maxims as universal laws and in which he has given an example of the application of that principle. He makes many of the same points again in this paragraph: all we have to do is ask (as if it were always a simple matter to arrive at the answer) the question whether we can also will our maxim as a universal law; if we cannot so will it, then the maxim is to be rejected — and rejected not because of any disadvantage to us but rather precisely because the maxim cannot be willed as a universal law so that it cannot fit into a possible universal lawgiving. This last bit, about a possible universal lawgiving is new, and Kant will not have more to say about it until deep into the second section of the work. Most of the rest of the paragraph is a rehash: by noting that he so far has no insight into the ground for the respect that a good will has with regard to the practical law, he hints again at the possibility that morality may be a fantasy; he again says that anything that inclination praises is of much less worth; he repeats the third proposition which identifies duty as a kind of necessity that is a requirement for a good will; and he repeats the claim that the worth of a good will exceeds the worth of anything else.

In this paragraph, Kant finishes his wrap-up (which he began in the previous paragraph) of the search for the principle of morality. He announces that we have now arrived at this principle (or at least a version of it), and he insists again that ordinary human reason actually uses the principle, though in a different form, as a norm for moral judgment. Kant even goes on to claim that it would be easy to show that ordinary human reason, by using this principle, very well knows how to make moral distinctions. Though Kant says this would be easy to show, one might wonder how it might actually be shown. The claim is about how humans actually think rationally. If such ordinary use of reason cannot be studied a priori or non-empirically, then the evidence for the claim would presumably come from empirical studies of human rationality or from what Kant would call practical anthropology. It is doubtful that there were any such studies that Kant could have used for evidence; so Kant is either relying only on his own limited experience of humanity in the Königsberg area, or on anecdotal accounts from others, or he has entirely something else in mind. This last possibility is the most intriguing. Perhaps Kant was thinking along these lines: I have shown that if genuine moral judgments are made at all, then they must be made by using the principle of willing maxims as universal laws; so, if humans really are making such genuine moral judgments, then they must be using the principle that I have uncovered in the course of my search for the highest moral principle. If this is Kant's line of thinking, then he does not need any direct empirical evidence (and maybe this is why he thought it easy); for his general argumentative strategy here is to argue that for X to be even possible Y must be the case. But Kant does not go on in the rest of the paragraph to argue anything like this. Instead, he argues that science and philosophy are not needed in order know what to do in order to be morally good. And he ends by wondering whether philosophy should just leave morality well enough alone.

Kant has claimed that ordinary human reason already makes use of the principle of willing maxims as universal laws, and he has realized that this use suggests that it might be better if philosophy kept its hands off of morality. In this paragraph, Kant begins to reveal what role philosophy still has to play in morality. He argues that wisdom or wise conduct still needs science so that what wisdom prescribes can get a foothold in the will and eventually secure a lasting presence in it. To support this conclusion, Kant essentially says that innocence (i.e., the character of a will that has not been tutored in philosophy) is a fragile thing that can easily be shattered by the natural dialectic that arises when the commands of duty are pulled down by the counterweight of needs and inclinations. So it seems that without science, perhaps without those three parts of science that make up philosophy, the wisdom that ordinary human reason exhibits in prescribing actions in accord with the principle of willing maxims as universal laws might fail to prevail against the trappings of happiness.

In this and the immediately preceding paragraphs, Kant diagnoses a problem and then proposes a philosophical solution to the problem. In this particular case, the problem is that common human reason can lose its grasp on genuine moral principles. This loss can occur because of ambiguity into which common human reason can fall. This ambiguity, in turn, is apparently precipitated by the counterweight of needs, inclinations, and wishes that beset reason and its commands of duty. It is not so hard to see how philosophy might be used to clear up ambiguities, for philosophy is well-equipped to deal with concepts, ideas, definitions, and arguments. So Kant proposes a critique of reason, which is to reveal the limits of reason's powers, what it can and cannot do, where or on what it can and cannot operate. Perhaps such a critique will remove the ambiguity into which reason falls.

But the ambiguity is not the root of the problem that Kant has diagnosed. The root is the opposition between reason and inclination or, in general, happiness. Can philosophy do anything about this opposition, get at the root of the problem? Can philosophy remove the opposition, or even reduce it, or can philosophy merely manage the effects of the opposition by constantly battling against the resulting ambiguity? If philosophy can only do the latter, perhaps it would be easier to take a different approach to handling the problem. Perhaps it would be better to use psychology and biology (e.g., genetic manipulation), not philosophy, to change human needs and inclinations in such a way that the opposition between reason and happiness disappears. We may now, or soon, have the technological means to make such changes. These means, of course, were not available, perhaps not even imaginable, in Kant's time. So Kant must hope that his proposed critique of reason is so sufficiently transformative that it changes the way reason conceives of itself, thus changing the dynamic between reason and happiness so that their opposition does not result in ambiguity.

Are there any other options? If the problem is reducible to an opposition between reason and happiness, then, in order to eliminate the opposition, either reason must change (Kant's apparent solution) or happiness must change (the psychological solution suggested above) or both must change. Or, instead of trying directly to change one or the other or both, might it be possible to change the setting, the host environment, the circumstances of their coexistence, in such a way that the opposition between them never arises?

Though there is no argument in this paragraph, there are some important points to emphasize. First, the concept of duty is not a concept of experience. Kant emphasizes that though the concept of duty is used in the moral thinking of common human reason, he has not treated the concept as if it were a concept based on experience. He has not, that is, tried to base the concept on our experience of human behavior that is in accordance with duty; he has not done this because it always remains possible that such behavior is not done from duty.

Second, Kant says that for an action to have moral worth it must be done from pure duty. The use of 'pure' ('reiner' at 25.13 (4:406.10)) here can be read as emphasizing that the motives behind an action that has moral worth must only be of one kind; they must all be based on duty with no mixture of empirical motives such as desires and inclinations. This insistence on purity of motive does not rule out the possibility that other motives might be present when an action from duty occurs. What it does rule out is that these other motives are active in producing the action from duty. For the action from duty to remain intact as such, any other motives, incentives, or impulses which might be present must play no role in actually producing the action.

Kant continues in this paragraph to make the case that the concept of duty is not a concept of experience. The argument he gives seems to depend on an implicit premise to the effect that some regions of our inner life are, or at least might be, inaccessible to us. No matter how thoroughly we search ourselves inwardly, we cannot be sure that we have surveyed every region that is there. It might always be the case that there are some operations of the mind, some features of our motivational structure, that we cannot reach through our self-examination. It is hard to know whether this implicit premise is true. There have been philosophers, perhaps Descartes for example, who have claimed that our inner lives can be completely transparent to us; if we put forth the effort, we can clearly and distinctly perceive our inner mental workings and be sure of the nature of the mental operations we are observing. At the other end of the spectrum, there are theorists such as Freud who propose that some of our inner life, the subconscious, is completely opaque to us and we cannot get at it, at least not without special therapy. In this argument, Kant seems to be siding with the theorists of the opaque. But perhaps he does not need to go so far. Perhaps all he really needs to claim is that there are parts of our inner life of which we cannot be sure; they are not totally hidden from us, not secrets that we cannot access. He might claim instead that these parts of our inner life are accessible, but not clearly and distinctly accessible or not accessible long enough for us to understand them completely. So, for example, to get his conclusion to go through, Kant might only need to say that we have a complex, mixed set of motives, some moral, some not; and, although we are aware of all of them, we cannot ever know which specific motive is operative or decisive in producing a given good action. And, since we cannot know which specific motive is operative, we cannot know for sure if the action is done from duty.

In this paragraph, Kant continues with his objections to the thought of those who make morality out to be a mere fantasy because it has no empirical or experiential support. Kant continues to agree that morality has no such support, examples, for instance, not providing such support. But Kant also continues to insist that this lack of support does not entail that morality is a mere fantasy, for the support might lie elsewhere, namely, in a reason which gives a priori direction to the will. Because this direction occurs in an a priori fashion, independently of experience, it should not be expected that examples from experience would provide any support for the reality of morality. So the mistake that these people are making is not in their belief that experience provides no support for morality; that belief is actually correct. The mistake, instead, is their belief that only experience could provide such support. Kant is complaining against these people that they are not casting their net wide enough or, rather, that they are not casting it in the right direction. They are looking for support in the wrong place, and they are too narrow-minded to look elsewhere, to a priori reason, for philosophical support for morality.

There are several items to note and emphasize. First, notice that Kant is still not committing himself to the view that morality is real. His first premise in the paragraph is hedged with the condition that truth and reference to a possible object are not completely denied of morality. So the assertion that morality is real is not part of this premise. Rather, the premise asserts something more like this hypothetical: if morality has at least some truth to it, then it applies to all rational beings with absolute necessity. Kant is committing himself only to the view that morality must have these features, universality and absolute necessity, if there really is such a thing as morality.

Second, moral laws apply to all rational beings, not just to human beings. Kant is never explicit about whom, besides human beings, these rational beings might be or explicit about what precisely it is that makes these beings rational. We can surmise with regard to the first that Kant might have had in mind beings such as angels and God and lesser beings such as extra-terrestrials living on other planets. Even at this early stage there is a bit more to go on with regard to the second, for Kant has already made it clear that maxims play an important role in acting from duty. So it looks like the ability to formulate and consider maxims is at least part of what it is that makes a being a rational being as opposed to a non-rational being.

Third, we get a slightly clearer picture of what Kant means when he says that moral laws hold with absolute necessity. Here Kant contrasts this absolute necessity with holding merely under contingent conditions and with exceptions. So moral laws hold only under conditions that are not contingent. What conditions are not contingent? That is a tough question, but a full answer would likely mention, by way of contrast, conditions that vary from one rational being to another: physical appearance, talents and abilities (e.g., musical, athletic), wealth, native spoken language, and so on. Characteristics such as these are contingent conditions because the rational beings who have them do not have to have them. The characteristics might have been absent or missing. Or the characteristics could have been different, either from the start or after a period of similarity subsequent to which the characteristics change or morph into some other characteristic or are lost altogether. So, when Kant says that moral laws hold only under conditions that are not contingent, he means at least in part that moral laws cannot be based on characteristics that can vary from individual to individual. Kant also mentions that moral laws do not allow for exceptions. What might this mean? Perhaps what Kant has in mind is that specific individual (or groups of) rational beings cannot be exempted from a genuine moral law. That is, there cannot be a moral law that applies to everyone except rational being so-and-so. If it applies at all, then it applies to everyone.

In this paragraph, Kant specifies how examples are to be used in morality. He says that they are only to be used for encouragement by showing the feasibility of what is depicted in the example, and they are to be used to make intuitive what the practical rule expresses more generally. It might be an instructive exercise to examine Kant's own examples in light of this paragraph and see if they serve as encouragement and make his thought more intuitive.

It is likewise instructive to reflect on Kant's answer to his question (29.22-3 (4:408.37-409.1)) about where we get the concept of God as the highest good. Back at 7.12-14 (4:396.24-5), Kant said that the good will is the highest good. Does that jibe with what he says here in response to the question?

In this brief paragraph, Kant continues his attack on popular moral philosophy which has the bad habit of mixing empirical elements in with the pure a priori core of morality. Because of this bad habit, Kant thinks it especially important to move beyond popular moral philosophy and instead move toward a metaphysics of morals. He suggests that making a case for this move toward a metaphysics of morals would not even be necessary if this bad habit of mixing in the empirical were not so prevalent; for, as he says in the first sentence, if there is no genuine highest principle of morality that can rest even a little bit on experience, then the question does not even arise as to whether it would be good to give such a general and abstract presentation of the a priori concepts. It is, then, these pernicious popular moral philosophies that cause the question about goodness to be raised, and so these philosophies cannot be ignored, but must be combatted, to ensure that the obviously correct answer to the question is recognized.

In this paragraph, Kant lodges some complaints against popular moral philosophy and its use of folk concepts. This reference to folk concepts is particularly interesting and needs further examination. What are these folk concepts, and what distinguishes them from non-folk concepts? How do they differ, if they do, from the concepts of a good will and of duty, which Kant uses and seems to think are in fact used by ordinary human reason? Are these concepts of a good will and of duty not folk concepts even though they are widely used by ordinary folk? Perhaps the answers to questions like these lie more in how the concepts are assembled and then used than in what the concepts are. If the concepts are assembled in an a priori fashion, from pure reason and not at all from examples and experience, perhaps then they are of a quality that can be used to support foundational thinking about morality.

In this paragraph, Kant tells us that popular moral philosophy presents morality as an impure mixture of elements from both experience and reason. He includes in his list of elements the make-up of human nature and the idea of a rational nature in general, happiness and perfection, moral feeling and the fear of God. Kant also suggests that popular moral philosophy fails even to consider the possibility that the principles of morality are not to be found in human nature and fails to realize that a metaphysics of morals must be completed before any attempt at popularization is attempted. Though Kant here gives neither the names of any popular moral philosophers nor the titles of any works of popular moral philosophy, he might have had philosophers such as Christian Garve, Moses Mendelssohn, Christoph Meiners, and Christoph Friedrich Nicolai in mind.

In this paragraph, we get a glimpse of Kant's elusive moral psychology. Kant tells us that the representation of duty can through reason influence the human heart more decisively than can anything from the empirical aspects of our make-up. The thought of duty, which we have through our reason, can master any of the empirical incentives that we feel. We do not have to be slaves of our passions, as Hume thought. We can act from principles identified by reason and do not have to succumb to our empirical desires and impulses. Kant, unfortunately, never worked out a complete moral psychology. If he had, it probably would not have been a simplistic account, for in this paragraph alone — in just one sentence, in fact — we get a smattering of terms which are sometimes hard to differentiate but which are important for a full understanding of human psychology. Kant speaks of impulses or stimuli ('Anreizen'), of incentives ('Triebfedern'), of inclinations ('Neigungen'), and of motives ('Bewegursachen').

In this paragraph of conclusions, Kant effectively summarizes much of what he has argued for up to this point. There are some surprises, though. He seems to conclude, for instance, that it is impossible, without a metaphysics of morals, to create pure moral dispositions. Up to this point, Kant has said very little about moral psychology, part of which deals with how humans actually acquire moral beliefs and attitudes. That he has said almost nothing about this is in itself not surprising, for these topics fall under what Kant calls practical anthropology which is not to be a part of the grounding of morality. What is surprising, though, is that he then draws a conclusion that requires for its support evidence from practical anthropology. Perhaps Kant got carried away when he was writing this particular paragraph. Or perhaps he was just trying to impress upon the reader that the purity of the moral dispositions requires metaphysics, the creation of those dispositions in beings such as ourselves being merely the setting for emphasizing this particular point about purity and metaphysics.

Having in the last paragraph summarized the results of this investigation into the highest principle of morality, Kant looks ahead in this paragraph, deciding how to proceed in the investigation. He says the transition from ordinary moral judgment to the philosophical has already been made; that was presumably accomplished in the First Section. He now wants to make a transition from popular philosophy, which he has just found inadequate in part because of its reliance on examples drawn from experience, to metaphysics. To make this proposed transition by natural steps, Kant says we must follow and clearly present the practical faculty of reason from its universal rules of determination to where the concept of duty arises from the faculty.

The student might wonder why Kant starts yet another investigation, this time into practical reason, here. He announced in the Preface that the ultimate purpose of the work is to find and establish the highest principle of morality. But then he started an investigation into the good will, followed by an investigation into duty, and now he is starting another investigation into practical reason. The student might wonder what all of these investigations have to do with that stated purpose of finding and establishing the highest principle of morality. The answer, I think, lies especially in appreciating the second part of the stated purpose: the establishment of the highest principle of morality. In order to establish the principle, Kant needs a theory that can be used to justify or defend the discovered highest principle. To get this theory, he needs to embed the highest principle in a web of concepts that reveal the principle's connections to other aspects of morality and then to morality's place in human experience in general. By investigating the good will, duty, and practical reason, Kant is in effect showing how the highest principle (to do only those actions whose maxims can be willed as universal laws) fits in with other morally relevant concepts, thus generating the theory he needs in order to establish the principle. So these investigations are part of the overall process of assembling the theory by natural steps, starting with what is familiar to anyone acquainted with morality, a good will and duty, and then progressing into areas less immediately familiar such as practical reason.

If we take him at his word (from the previous paragraph), in this paragraph Kant begins his investigation into the faculty of practical reason. Though it is not obvious how he does it, Kant very quickly reaches the conclusion that the will is nothing other than practical reason (see also 89.4-5 (4:441.17)). But then Kant distinguishes two kinds of will; or, if not two kinds, then they at least behave differently in relation to practical reason. The first is the kind that actual humans do not have. It is a will that is fully determined or controlled by reason to do those actions that practical reason has identified as good or practically necessary actions. In the case of such a will, the actions identified by practical reason as good are said to be both objectively and subjectively necessary. The other kind of will, the kind that humans do have, is not fully determined by reason to do what practical reason has identified as good; this human kind of will can, for instance, be swayed by inclinations that are not aligned or in sync with practical reason. In the case of this human will, good actions are said to be objectively necessary but subjectively contingent. Because of this subjective contingency, only the human will is necessitated by the objective law. This necessitation is what humans experience as moral obligation or as moral duty. Kant's cryptic third proposition, stated at 14.13 (4:400.18), that duty is the necessity of an action from respect for the law, should now make more sense.

The term 'imperative' appears in this paragraph for the first time in the work. Kant will go on to explain in more detail what an imperative is, and he says here only that an imperative is the formula of a command of reason. In the subsequent pages, it becomes clear that a command of reason can be expressed in multiple ways, each of these ways being a distinct formula of the one command. So a formula, in this context, is basically just one way of expressing, of formulating, in words a command of reason. We also learn in this paragraph that a command of reason is a representation of an objective principle so far as the principle is necessitating for a will. In plainer language, this means that a command of reason is the thought or general conception of a prescription for action which reason has identified and which a human will is obliged to follow.

Delving deeper into the operation of practical reason, into how willing works, Kant has found the imperative. In this paragraph, Kant begins a lengthy description and explanation of what imperatives are and how they work. He first notices that all imperatives are expressed by an ought. This ought that is expressed in all imperatives is supposed to reveal that some wills (e.g., human wills) do not necessarily follow an objective law that reason has identified (e.g., humans do not always do what they should do). Because of this always-present possibility of a breakdown between what reason prescribes and what is actually willed, the relation between these wills and the law identified by reason is characterized by an ought or necessitation. (Recall again the third proposition regarding duty at 14.13 (4:400.18).) Kant emphasizes that when practical reason identifies an objective law or specifies a practical good, then practical reason has found something that does or, in the case of humans, ought to determine the will of all rational beings just because they are rational beings. What practical reason specifies as a practical good, then, markedly differs from what is merely pleasant because what is pleasant can, and frequently does, differ from one being to another since the beings have different sensitivities.

Kant is in the early stages of examining practical reason. He has found, by recognizing the existence of the necessitation of some wills, that imperatives figure prominently in the operation of practical reason. So far, Kant seems most concerned to make the point that imperatives are formulas. In particular, they are formulas of commands of reason, and these commands of reason are necessitating for an imperfect will that does not always do what reason has identified as practically good to do. Actual humans, of course, have imperfect wills. Their wills can be influenced by non-rational elements such as inclinations and desires. Because of these non-rational influences that often pull in a direction opposite to that of reason, the human will sometimes fails to do what reason has identified as practically good to do. To overcome these non-rational influences, then, reason must issue commands that tell the will, that necessitate the will, to do what is practically good. Humans receive these commands as formulas, in the shape of imperatives, that tell them that they ought to do something.

In this paragraph, Kant gives his first division of imperatives into two kinds: hypothetical and categorical. It is important to note that both kinds of imperative involve the representation of the necessity of an action. In the case of hypothetical imperatives, the necessity involved is the necessity of a possible action as a means to achieve something else that a rational being wants or might want. The categorical imperative, on the other hand, represents the action as necessary in itself, not as necessary or required in order to get something else. In short, hypothetical imperatives represent an action as necessary as a means while categorical imperatives represent an action as necessary in itself.

In this paragraph, Kant gives his second division of imperatives into the two kinds. It might be wondered why he presents the division twice in consecutive paragraphs. In the first division, in the immediately preceding paragraph, he characterized imperatives as commanding either hypothetically or categorically. So he was there perhaps trying to emphasize the different ways in which imperatives function as commands; they function either hypothetically (as means to something else) or categorically (aiming at nothing else beyond the action specified in the command). In the second division, in this current paragraph, on the other hand, Kant focuses on the goodness of the action specified by the imperative, specifically on whether the action represented in the imperative is viewed as good as a means to something else or as good in itself. If the action is viewed as good as a means, then Kant names the imperative hypothetical; if viewed as good in itself, then Kant names the imperative categorical.

Up to this point, Kant has been describing imperatives. In this paragraph, he draws two conclusions about imperatives in general. He also gives at least two, perhaps three, explanations as to why some rational beings do not always follow an imperative, that is, do not do the good action specified in the imperative.

The first explanation is that the rational being does not know that the action commanded by the imperative is a good action and so does not do the action. Depending on how much weight the concept of knowing is to carry, this case might better be taken as a case of not being cognizant of the imperative; for, if the rational being is cognizant of the imperative, then it knows what action would be good to do insofar as the imperative, as asserted in Kant's first conclusion, says precisely what action would be good. In short, the rational being does not follow the imperative because the rational being is not aware of the imperative. It might be interesting to investigate the ways in which this lack of awareness can occur. The investigation of this subject, however, is a task for practical anthropology.

Kant's second explanation of why some rational beings do not follow an imperative is that the being might have maxims that are contrary to the objective principles of a practical reason, the ignored or unfollowed imperative being a product of the operation of practical reason. In this kind of case, the subjective principles (i.e., the maxims) have more sway over the will of the rational being than do the objective principles recognized by practical reason and formulated as imperatives. It might also be interesting to investigate the ways in which one set of principles can exert more influence over the will than another set of principles. This, too, however, is an investigation to be conducted by practical anthropology.

The possible third explanation of failure to follow an imperative is a combination of the first two. Kant might want to say that a rational being can lack the knowledge that an action is good and at the same time can have maxims that are at odds with the objective principles of practical reason. This combination of factors, then, leads to a failure to do the good action that an imperative commands.

Summing up, the three possible explanations of failure to do a good action are:
1. ignorance of the imperative (and hence that the action is good);
2. having subjective maxims contrary to objective principles;
3. a combination of the previous two explanations.

In this paragraph, we learn more about the relationships between imperatives and principles and between purposes and ends. Kant says that an hypothetical imperative is either a problematic practical principle or an assertoric practical principle. In the case of the categorical imperative, however, he says not that it is but that it holds as an apodictic practical principle. There might be no real difference here: an imperative might be a practical principle if and only if it holds as a practical principle. To find out if this equivalence is true, we need to look back at what Kant has said about principles, commands, representations, formulas, and imperatives to see if they all fit together in the appropriate way to make the equivalence true. A good place to start this investigation is at the short middle paragraph on page 37 (4:413).

We also learn in this paragraph that there is a close connection of some kind between purposes and ends. Kant says that the categorical imperative designates an action as objectively necessary for itself, and it makes this designation without reference to any purpose. Kant then explains what he means by 'without reference to any purpose' by saying that this also excludes any other end, aim, or goal. So Kant might be implying that purposes require ends. The connection might even be more subtle than this. It might be that having a purpose requires having some other end. In the context of a categorical imperative, then, the situation would be this: the action designated by the imperative is, since the action is objectively necessary, itself an end; and the imperative commands the action without referencing any purpose, without, that is, specifying any (other) end other than the action itself.

Kant gives two examples in this paragraph. It is not entirely clear what exactly they are supposed to illustrate. The first example, of the doctor and the poisoner, seems intended to illustrate Kant's point that imperatives of skill say nothing about the moral worth of the ends to be achieved by performing the actions commanded in the imperative. That is why Kant says that the prescriptions that the doctor uses and those that the poisoner uses are of equal worth. With imperatives of skill, what matters is whether the action commanded is effective in achieving the purpose. In short, imperatives of skill are not imperatives of morality. But it also seems possible that the example is supposed to illustrate that imperatives of skill can be either problematic (dealing with merely possible purposes) or assertoric (dealing with actual purposes), for the purposes of the doctor and poisoner might be actual for some people but only possible for others.

The second example, about parents wanting to have their children learn many different skills which might serve the grown-up children's purposes, is similarly ambiguous about what it is supposed to illustrate. At first, the example seems intended to illustrate that imperatives of skill hold as problematic practical principles, for the children will be learning many skills which do not serve their current actual purposes but only their possible later purposes. But Kant also builds into the example that parents often fail to educate their children to make appropriate judgments about the worth of the ends to which children might put their newly acquired skills. So again in this example, Kant also seems to be trying to make the point that imperatives of skill are not imperatives of morality.

In this paragraph, Kant uses the hypothetical imperative commanding the furtherance of one's happiness as an example of an assertoric practical principle. The imperative is hypothetical because it commands the means to achieve something else and only hypothetical imperatives reference means. The reasoning Kant gives in support of the assertoric status of the imperative is more interesting. It is assertoric because the purpose (i.e., furthering happiness) specified in the imperative is actual for all rational beings with imperfect wills; and it is actual because, as Kant argues, it can safely and a priori be presupposed for all such rational beings insofar as it is part of their essence. So Kant's chain of reasoning goes from essence to a priori presupposition to actual occurrence. Lots of questions should be popping up now. In general, how are we to determine the essence of something? How does Kant do it in this particular case? Are all parts of something's essence such that they can safely and a priori be presupposed? Or do only some parts of something's essence have this standing? Is there any significance to the 'safely'? Could there be unsafe a priori presuppositions? What is the nature of the inference from a priori presupposition to actual occurrence? If a property P can be a priori ascribed to object O, does it then necessarily follow that if something is an object of type O then it has property P? What if no Ps actually exist; what if no Os actually exist? Would the existence of an O guarantee the existence of a P?

Though there is no argument in this paragraph, it does give us more information about the categorical imperative. First, the categorical imperative does not lay down for its ground, as a condition, any other purpose to be obtained through a certain conduct. Second, the categorical imperative immediately or directly commands conduct. Third, the categorical imperative does not deal with the matter or content of action or with what is to result from the action. Fourth, the categorical imperative deals with the form and the principle from which it follows. Fifth, the essential good of the categorical imperative consists in the disposition. Sixth, the categorical imperative is the imperative of morality.
The fourth item, that categorical imperatives follow out of form and principle, in the list is most interesting, for it gives us a glimpse into the genesis of categorical imperatives. This fourth item suggests that a categorical imperative might come to be in the following way: reason looks into itself and identifies a law (see page ix (4:389)); the law has a form (see 16, 20 (4:401, 403)); reason expresses this form in an objective principle (again see pages 16, 20); a command of reason represents the principle (see page 37 (4:413)); the imperative is the formula of this command (again see page 37).

With this paragraph, Kant finishes up his classification of imperatives. He has enumerated them and described their similarities and differences. He has found that in practical reason there are two main kinds of imperative: hypothetical and categorical. But the hypothetical imperative comes in two varieties: technical and pragmatic, or alternately, rules of skill and counsels of prudence. Only the categorical imperative is a law or command, and it alone is the imperative of morality because it alone carries with it an unconditioned, objective kind of necessity.

What all of these imperatives have in common is that they say that an action is necessary and along with this that the will is necessitated. For hypothetical imperatives, however, that necessity is conditioned, contingent, and subjective; the action is necessary only as a means to something else. This something else is the condition upon which the necessity expressed in the imperative is based. Because this something else, the condition, is contingent and subjective, it can be taken away or removed, effectively disabling the necessity that was expressed in the hypothetical imperative. In this paragraph, Kant's example to illustrate this weakness in hypothetical imperatives is happiness. Kant has already mentioned (see 11-12 (4:398-399)) that the concept of happiness is built up out of empirical components and so is contingent and subjective: one person's concept of happiness might differ from another person's concept of happiness because each person has built up the concept from a (at least partially) different set of empirical components. So an hypothetical imperative that tells us to do a certain action in order to further happiness has only a conditioned and subjective necessity, for the imperative's condition, to further happiness, is itself contingent and subjective. The necessity in the categorical imperative, however, is not conditioned or subjective; it is absolutely necessary and objective. In the case of the categorical imperative, there is no something else as a condition, and so there is nothing that can be taken away or removed that would weaken the necessity expressed in the imperative. So the necessity expressed in the categorical imperative is unconditioned, absolute, and objective, and it is precisely because the necessity has these features that the categorical imperative alone can and must serve as the imperative of morality.

At this point, Kant has finished his descriptive account of imperatives. He turns now to the question of how the imperatives are possible. With this question, Kant means to ask merely how the necessitation of the will can be thought and not to ask how the carrying out of the action commanded by the imperative can be thought. In this paragraph, Kant answers the question only with regard to imperatives of skill; subsequent paragraphs deal with the remaining imperatives. Kant's main point here is that imperatives of skill are analytic and so the question of how they are possible can be answered rather quickly and briefly.

Kant's use of 'analytic' is here different from his use of it at the end of the Preface; here it characterizes a kind of proposition while there it characterized a methodology. As applied to propositions, 'analytic' means that the concepts expressed in the proposition are necessarily connected in thought because one concept contains the other; that is, the thought of the one concept includes, without having to add any further information, the thought of the other concept. In this paragraph, Kant also refers to 'synthetic' propositions; in such propositions, the thought of one concept does not include the thought of the other, and the concepts are connected only by adding additional information. So, to take Kant's example of equally dividing a line: only by adding additional information from mathematics are the thought of an equally divided line and the thought of intersecting arcs brought together to form the synthetic proposition that intersecting arcs must be used in order to divide a line equally into two parts; but, having already learned that this synthetic proposition is true (i.e., that I have to use intersecting arcs in order to divide the line equally), it is an analytic proposition that if I will to divide the line equally, then I will to use intersecting arcs; and Kant thinks this analytic proposition is true (and keeping in mind all the while that the synthetic proposition is true) because the thought (or representation) of the line as having been equally divided by me is the same as, and so includes, the thought of my using the intersecting arcs to divide the line.

So what is the answer to Kant's question of how imperatives of skill are possible, of how the necessitation that they impose on the will is to be thought? Kant's answer is that the necessitation is to be thought as analytic, as a necessity in thought. In willing to do one action (e.g., the end) the will is also willing to do another (e.g., the means). These two willings are really the same willing. They appear different only when their representations (i.e., how we think of them) differ (e.g., thinking of something as my effect versus thinking of myself as a cause of the effect). But when we examine the concepts involved in the representations more closely, we find that one concept contains the other so that the associated imperative expressing necessitation is analytic.

In this paragraph, Kant continues to answer the question of how the three kinds of imperatives (of skill, of prudence, of morality) are possible, that is, how the necessitation of the will, which the imperative expresses, can be thought. His answer with regard to imperatives of skill was that the necessitation can be thought as analytic because the proposition that whoever wills the end wills the means is (with the stated qualifications) analytic. Kant wants to give the same answer in the case of imperatives of prudence. But there is a complication: the concept of happiness is too indeterminate to be used in a principled way and therefore unsuitable as a law and so also unsuitable as an imperative, a formula of a law. Strictly speaking, then, imperatives of prudence do not command and thus do not necessitate the will. So the question of the possibility of this kind of imperative is meaningless; the question can be meaningfully asked only of imperatives of prudence if they are viewed as if they were genuine imperatives.

It is worth considering what exactly Kant means by saying that the concept of happiness is indeterminate and why this indeterminacy makes the concept unsuitable to be the basis, ultimately, for an imperative. To consider this, we should look more closely at what Kant says about indeterminacy and at the examples he gives.

Kant says the indeterminacy is at a level that does not prevent human beings from wishing to attain happiness but that does prevent them from ever saying determinately and with self-consistency what the human being really wishes and wants. So there is enough determinacy in the concept for us to wish for happiness but not enough for us to say determinately and consistently what we are wishing to attain. This is not much to go on. He might mean that the concept is ambiguous or vague. If it is ambiguous, then we grasp it clearly but in multiple ways; sometimes we might clearly see it having features X, Y, and Z, other times with features W, Y, and Z. If it is vague, then we never see it clearly, but we do grasp that it is something that has some set of features though we cannot say specifically what those features are. Or the concept could be both ambiguous and vague, in the following way: we might on occasion grasp the concept only vaguely so that no part of it is in focus; but then it slowly comes into focus and we clearly see its features; but then it goes out of focus again and when later it comes back into focus we clearly see it as having a different set of features; and the concept continues to fluctuate in this way, from vague to clear but ambiguous, never becoming fixed, stable, and persistently distinct. This tendency of the concept to fluctuate would also help to explain Kant's claim that we cannot speak of it with consistency; the concept is too changeable for us to say definitely that it always and only has a certain set of features.

What Kant says next about the cause of the indeterminacy might also help us figure out what Kant has in mind. He says that the cause of the indeterminacy is that all the elements of the concept are empirical in the sense that they must be borrowed from experience. At the same time, he says that the concept of happiness requires a maximum of well-being in one's present and future circumstances. And then he gives a number of examples of willing for something that we think will make us happy; but in each case, the thing we will to have brings with it the possibility of unintended and troublesome consequences. Because we are not omniscient about what we will experience in the future, we cannot know in advance what these consequences will be. So it is impossible to design principles that will give us a safe and secure route to happiness; the best we can do is come up with counsels of prudence that work out pretty well on average. Kant, then, might be claiming that the cause of the concept's ambiguity and vagueness, and so its indeterminacy, is the impossibility of having complete access to the full range of our possible experiences. We might now guess, based on what Kant has already said, especially in the Preface, about the relative merit of concepts based on empirical and rational grounds, that Kant will not find this indeterminacy in concepts whose elements come solely from reason rather than from experience. Because reason is fixed and stable, the same for all rational beings, and because we have (or will have after reason is subjected to a thorough critique that reveals its powers and limits) sufficient access to reason, concepts based on reason alone can serve as the basis for principles, laws, and (genuine) imperatives that command with objective practical necessity.

Kant leaves us hanging a bit in this paragraph. He does not, for instance, help us understand why not relying on a presupposition makes the possibility of categorical imperatives more problematic than the possibility of hypothetical imperatives. It is also not yet clear why Kant repeats at this point his objection to using examples. He seems worried that readers might lump categorical imperatives together with hypothetical imperatives and accordingly try to base categorical imperatives on the same empirical grounds. But this is odd because he has just spent several paragraphs classifying the imperatives and showing how they differ.
We do, though, learn in this paragraph at least this, which is noteworthy because most of the time Kant speaks of what is good, not of what is bad, in itself: that if an action must be considered as bad in itself, then the imperative prohibiting the action is categorical. Judging from Kant's example of a deceiving promise, an action bad in itself is one which is bad even apart from any bad consequences that it might bring about. And, going by what we have already learned from Kant, such a bad action must be bad because its maxim cannot be willed as a universal law.

After reading this paragraph, we get a better idea of why Kant in the previous paragraph did what he did, in particular why he stressed that hypothetical imperatives rely on presuppositions and why he repeated the argument against the use of examples as support for the foundation of morality. It is now revealed to us that he did what he did in order to pave the way for his claim in this paragraph that the investigation of the categorical imperative will have to be conducted in an a priori manner. Because categorical imperatives are not hypothetical imperatives, the necessity in them does not rely on any presuppositions and so does not rely on any empirical presuppositions that are rooted in the wants, desires, needs, and inclinations which we experience. So, because we cannot get at categorical imperatives by using experience and consequently not by using examples drawn from experience, we have to use a priori methods to investigate the possibility of categorical imperatives.

Kant has been wondering how imperatives are possible, more precisely, how the necessity expressed in imperatives can be thought. He has found that there is no difficulty associated with the possibility of hypothetical imperatives; they (or the necessity expressed in them) can simply be thought as analytic propositions. But Kant does find difficulties with the categorical imperatives. The first difficulty, apparently relayed in the previous paragraph, with categorical imperatives is that experience cannot be used; the investigation of the categorical imperative has to be conducted a priori, without appeal to experience. In this paragraph (and the accompanying footnote), he gives an account of the second difficulty: categorical imperatives are synthetic a priori practical propositions. Two questions jump to the forefront: what makes synthetic a priori practical propositions so difficult, and why are categorical imperatives synthetic? The very short answers to these questions are: the difficulty with synthetic a priori practical propositions is that they assert a necessary connection between concepts which can be thought separately; categorical imperatives, unlike hypothetical imperatives in which willing the end entails willing the means, are synthetic because the concept of the will does not contain the concept of the action specified in the imperative.

In this paragraph, Kant tells us what approach he is going to try first and why. His approach is to examine the concept of a categorical imperative to see if the just that concept alone might provide the formula of a categorical imperative; this might be possible (i.e., the formula might be the formula of a categorical imperative) if the provided formula contains a proposition which alone can be a categorical imperative. He explains that he is taking this approach because investigating the possibility of the absolute command that a categorical imperative issues will require a strenuous effort and will be postponed until the Third Section. Kant might also have added that he is taking this approach through the concept of a categorical imperative because the approach, as he has just finished arguing in the previous paragraphs, must be a priori. He might also have been hoping that this prior look into the concept in search of a formula might help clear some things up so that the later task of answering how a categorical imperative is possible will be somewhat less difficult.

In this paragraph, Kant gives us some help in understanding what imperatives contain. We learn that imperatives, or at least categorical imperatives, contain only two things: the law and the necessity of the maxim to be in conformity with that law. This is especially helpful because Kant has previously said that imperatives express the necessity of an action (39.15-22 (4:414.12-17)) and that they express the necessitation of the will (43-44 (4:416-17)); this paragraph strongly suggests that these two expressions boil down to the necessity of the maxim of a categorical imperative to be in conformity with the law. Kant does not draw attention to this suggestion, much less explain it, but the suggestion is reasonable in light of the meaning Kant has attached to 'maxim' as a principle of willing (15 (4:400)) and as a principle of action (52 (4:421)). These principles, whether subjective or objective, are used by the will when it acts. So, in the concept of a maxim, we have a connection between willing and acting, between the will and actions. And when the principle is objective, this connection between willing and acting is infused with the necessity of the maxim's conformity to the law so that a duty to act arises from the willing.

Kant concludes in this paragraph that there is only one categorical imperative. The formulation he gives here is typically called the formula of universal law. By saying that there is only one, Kant apparently means that there is only one underlying command of reason and that the categorical imperative given here is just one way of expressing that command as a formula. It is, however, not at all clear what exactly the argument is which is to establish this conclusion that there is only one categorical imperative. Kant does not, for instance, use the favorite technique of mathematicians and logicians, which is to assume that there is more than one and then derive a contradiction. Perhaps the argument instead uses the claims made in the previous paragraph. One such claim was that the categorical imperative, in contrast to the hypothetical, represents as necessary the conformity to universal law alone. Kant might then be thinking that this kind of representation of the conformity as necessary is what all categorical imperatives have in common. So it is perhaps merely to this representation to which Kant is alluding when he concludes that there is only one categorical imperative.

After having just concluded that there is only one categorical imperative, in this paragraph Kant suggests that it might be possible to derive all imperatives of duty from the one categorical imperative given in the previous paragraph. Kant comes very close to contradicting himself here. The imperatives of duty to be derived from the one categorical imperative must themselves be categorical imperatives, for imperatives of duty are moral imperatives and the categorical imperative is the only imperative of morality. But there is no contradiction because Kant is here not asserting that such derivations have been, or even can be, done, but only asserting that if they can be done then we can indicate what we think by the concept of a categorical imperative and what it wants to say. The suspicion of a contradiction also fades if we interpret, as in the commentary above, the previous paragraph's claim about there being only one categorical imperative as a veiled statement about the underlying command of which the given categorical imperative is but one formula.

In this paragraph, we get the second formulation of the categorical imperative. The first formulation, stated a couple paragraphs earlier, commands you to act only according to that maxim through which you at the same time can will that the maxim become a universal law. The second formulation commands you to act as if the maxim of your action should (or is to) become through your will a universal law of nature. There are many difference between these two formulations: the first, but not the second, has 'only' ('nur'); the first has 'according to' ('nach') while the second has 'as if' ('als ob'); the first has 'at the same time' ('zugleich') but the second does not; in the first your willing goes through your maxim but in the second your maxim goes through your will; in the first you will that your maxim become ('werde') a universal law while in the second your maxim should (or is to) become ('werden sollte') a universal law (of nature). It is hard to say which, if any, of these differences is of philosophical importance. It might be useful to ask, too, why Kant is giving us multiple formulations (and there are more to come).

In this paragraph, Kant provides a lead-in for the four examples to follow. The examples are evidently intended to illustrate the use of the two formulations of the categorical Kant gave in the preceding paragraphs on page 52 (4:421). The examples are manifestly not being used to prove or justify any specific claim about the foundations of morality, for that would be contrary to Kant's position on the use of examples. (He explains in the attached footnote that the division of duties is only for ordering those examples. He also tells us in the footnote that perfect duties are those that permit no exceptions to the advantage of inclination and that there are both outer and inner perfect duties.)

One of the best things about these examples that Kant gives starting in this paragraph is that we get along with the examples some nice samples of maxims. The sample maxim in this first example is: from self-love I make it my principle to shorten my life when its longer duration threatens more misery than it promises pleasantness. This particular maxim specifies:
1. a motive or incentive: self-love;
2. an agent or actor: I;
3. a result or effect: a shortened life
4. conditions or circumstances: a life of continuing misery rather than pleasantness.
It might be worth noting that, in this sample maxim, the maxim contains a reference to an adopted principle (consisting chiefly of items 3 and 4). But Kant has earlier said (pp. 15, 52 (4:400, 420-1)) in the footnotes) that maxims themselves are subjective principles and contain a practical rule. This is perhaps an example of Kant's carelessness or, more favorably, looseness, in using terminology.

In this paragraph, Kant gives his second example of a perfect duty, this time a perfect duty to others rather than to oneself. In both of the examples so far, Kant might want to say something like this: the maxim cannot be willed as a universal law of nature because the resulting maxim would contradict itself so that it is impossible for the maxim to become a universal law of nature and thus impossible to will that the maxim should become so. He does not, however, actually say this. For instance, in the first example, it is nature, not the maxim, which contradicts itself; Kant then infers from this self-contradiction in nature that nature would not endure as nature; and he then finally infers that it is impossible for the maxim to occur as a universal law of nature. In the first example, the maxim never contradicts itself, but at most conflicts with the highest principle of duty. In this second example, though, the universality of the maxim-become-law would make promising impossible and then from this impossibility Kant infers that the maxim must contradict itself. Because of these differences between the examples, perhaps all we can say for sure is that Kant thinks that in trying to will the maxim as a universal law of nature a contradiction and/or impossibility of some kind would arise such that the maxim cannot become a universal law of nature and thus cannot be willed to become such.

In this second example, there are also differences in the structure of the sample maxim. For instance, the maxim here does not specify a motive or incentive. The motive, however, is very probably the same as the motive explicitly specified in the first sample maxim, for Kant also characterizes this second sample maxim as a principle of self-love (54.18 (4:422.24)).

In this paragraph, Kant gives what is generally considered to be the weakest of the four examples. This reputation might be due in part to the overly general phrasing which Kant gives to the assertions in the reported argument. Kant speaks, for instance, of 'all powers' ('alle Vermögen' at 56.2 (4:423.14)) when he perhaps only needs to speak of some talent. Earlier in the example, Kant speaks of 'one talent' or 'a talent' ('ein Talent' at 55.9 (4:422.37)) and of 'natural gifts' ('Naturgaben' at 55.16 (4:423.5-6)). So, rather than phrasing the reported premises in terms of every power, faculty, or capability, Kant might have limited the scope to just one or a few capabilities or aptitudes that are, relative to a rational being's many other capabilities, special in some way. If Kant had used this phrasing instead, the premise (and intermediate conclusion) claiming that a rational being necessarily wills that all its powers be developed would be somewhat less implausible. Even with this change, though, Kant still has much work left to do to support the other premise about purposes and to shore up the inference from that premise about purposes to the intermediate conclusion about necessary willing. But perhaps insisting on this kind of support for these premises misses the point; Kant is perhaps not trying to establish anything with these examples, but only trying to illustrate the operation of the first two formulations of the categorical imperative.

In this paragraph, which contains the last of the four examples in this initial batch of examples, Kant again locates the problem not in the maxim itself but rather in trying to will the maxim as a universal law of nature. As in the third example, Kant thinks that the maxim can be rendered successfully as a universal law, without impossibility or contradiction; but a (fully?) rational being cannot will that the maxim become a universal law because the being can foresee that situations would arise in which the will would conflict with itself, willing indifference on the one hand but also willing assistance on the other.

Kant has just finished his enumeration of four examples of (putative) moral duties. In this paragraph, he first cautions us that in giving these four examples he is not to be understood to be asserting that the duties actually exist. Kant is still holding back full commitment to the actual existence of morality; he is still operating only on the assumption that morality is something real. So, in giving the examples, Kant is only really claiming that if these duties exist then they exist as they do because rationality imposes limits on what can be willed. Kant then makes a very important assertion: the canon of moral judgment of action in general is that one (i.e., a rational being) must be able to will that a maxim of action become a universal law. In the rest of the paragraph, he elaborates on this canon. He says that some maxims cannot be thought as universal laws of nature and also (presumably because they cannot even be thought) cannot be willed as universal laws of nature; these kinds of maxims, such as those in the first two examples, have an inner impossibility. Other maxims, such as those in the third and fourth examples, can be thought as universal laws but still cannot be willed as universal laws because the rational will would contradict itself. Kant traces the differences in impossibility between these maxims back to the constitution of the actions (57.9 (4:424.3-4)) that the maxims reference. Unfortunately, Kant does not similarly elaborate on this bare mention of the source of the differences. Instead, he ends the paragraph with a claim about the completenesss of his analysis: these two kinds of impossibility, impossibility in thinking and impossibility of willing, effectively, through the examples in their dependence on the one principle (i.e., the categorical imperative), capture all the types of obligation exhibited in duties.

This paragraph about making exceptions for ourselves is notable for several reasons.

First, it tells us a bit more about what Kant means when he says that moral laws are absolute and universal. If a person makes an exception for herself, then the law is no longer absolute, holding under all conditions; for it then becomes conditioned upon to whom the law is being applied: if applied to her, it does not hold; if applied to others, it does hold. And the law is no longer universal, since not everyone is covered by it and some are exempt from it. In the Groundlaying, Kant does not say in an explicit way much about the concept of fairness, but it comes close to the surface in this discussion of exceptions and gives us another way to understand the absolute and universal character of moral laws.

Second, this discussion about the making of exceptions provides Kant with his first opportunity to reference different and opposing points of view attached to aspects of the will. The will has, so to speak, one foot in reason and another foot in inclination and, since reason and inclination often fail to agree, the will finds an antagonism dwelling within itself. As you will see, these opposing points of view will again play a very important role later in the Third Section of the work.

As is appropriate for a paragraph given to summary, Kant repeats yet again that he is working under the assumption that morality is real, that there are actual moral duties. Only on this assumption does he conclude that the categorical imperative, with the sparse content he has identified, expresses these duties. Consequently, he also has not asserted the real existence of the categorical imperative; his assertions only amount to the claim that if duties exist then the categorical imperative, as he has characterized it, exists. It might be interesting at some point to count the number of times Kant either outright says or implies that he is working under this assumption that morality is real. But even without the actual figure, it should be evident from his having mentioned it more than once that this assumption is an important part of his overall strategy, and we should continue to keep it in mind as we go along.

Having stated in the previous paragraph that he has not yet shown a priori that the categorical imperative is real, in this paragraph Kant reminds us why the proof of its reality must be a priori and what such a proof requires. It must be a priori, independent of anything empirical, because the categorical imperative must be unconditional and must hold for all rational beings and because anything empirical has neither unconditionality nor universality. This exclusion of the empirical requires that the categorical imperative cannot be based on human nature, for every property of human nature is empirical. So, since the categorical imperative cannot be based on any property of human nature, it also cannot be based on any special property of human nature. Another way to think of this requirement, and this might really be what Kant is getting at by emphasizing 'special' in this paragraph, is to note that what is special to human nature is unique to human nature and therefore not something possessed by other rational beings; but then, if the categorical imperative is based on what is special and unique to human nature, the categorical imperative might not hold for other rational beings; but, as the formula of a moral law which is universal, it must hold for all rational beings; so the categorical imperative cannot be based on any special property of human nature. The crucial step in this reasoning is perhaps the move from the attribution of uniqueness to the claim that what is based on that uniqueness might not hold for all rational beings. In other words, is it true that if P is a property unique to rational beings of type B and an imperative is based on P then the imperative might not hold for rational beings of all types, that is, cannot be the categorical imperative?

In this paragraph, Kant stresses the difficulty of offering an a priori proof of the reality of the categorical imperative. Because it is to be a priori, the proof cannot be based on the usual suspects often claimed as the source of morality: human nature, sense, feeling, emotion, happiness, instinct, imagination, convention, religion, divinities. The source, like moral laws themselves, must be pure, found neither in Heaven nor on Earth, and must therefore be a priori. Without this pure source for its foundational principles, morality has no commanding authority. This, then, is the 'precarious position' philosophy is in: with so much, all of morality and everything that depends on it, at stake, philosophy has the task of finding the route to a secure foundation for morality, a route that avoids the impure dead ends so often suggested and yet a route that still preserves the a priori purity of the source.

In this paragraph, Kant continues to stress the importance of excluding anything empirical from the foundations of morality so that the a priori purity of morals will remain intact. He says, for instance (61.9-12 (4:426.9-11)), that this purity accounts for the proper worth of an absolutely good will, for the worth such a will has above all price. In terms of a principle of action which a good will uses, what this purity implies is that the principle is free from the influences of all contingent grounds, such grounds being those that only experience can provide. With Kant's discussion of making exceptions for ourselves in the recent (57-9 (4:424)) background, Kant seems to want to make the point that such a pure principle does not have its origin in desires, feelings, emotions, or attitudes and consequently is impartial, unbiased, and does not favor any particular rational being or group of rational beings over others. Though this exclusion of the empirical is a recurrent theme in the work, whenever it appears, as it does again here, we would do well to remember that Kant intends to exclude the empirical only from the establishment of the foundations of morality, from, for instance, the establishment of the highest principle of morality. He does not intend that the empirical be banished from everything relevant to morality; he does not, for instance, think that moral principles can be applied, without knowledge of the empirical facts, to real world situations (recall, for example, ix.10-15 (4:389.29-33)). Still, we should wonder whether the exclusion of the empirical is sufficient to guarantee that our principles of action are impartial and unbiased. If a practical principle has its origin only in pure reason, does it necessarily follow that the principle shows no favoritism? Kant seems to want to make this inference from the non-empirical to the unbiased. How can such an inference be supported? What can be said for or against it? How does Kant support it? Or is it to be an immediate inference not requiring any explicit support?

At the beginning of this paragraph, Kant poses the conclusion-question of whether it is a necessary law for all rational beings that they always judge their actions according to such maxims which they can themselves will should serve as universal laws. Kant apparently frames the conclusion as a question because he has already admitted (see 59.12 (4:425.8)) that he has not yet given a proof of the reality of the categorical imperative. Kant of course wants to answer that there is such a necessary law; the formula of that law would be a categorical imperative. But Kant realizes that before he can justifiably give that answer he must first establish an a priori connection between that law and the concept of the will of a rational being in general. In the rest of the paragraph, Kant argues that establishing that a priori connection requires working in the metaphysics of morals. In the course of arguing for that requirement, it becomes somewhat clearer why he must establish a connection between the necessary law and the will and why that connection must be a priori. Kant's reasoning might be along these lines: because the necessary law is to be an objective-practical law, i.e., a moral law, it must be unconditionally and universally binding on the will of every rational being; a binding having these features of unconditionality and universality requires the exclusion of anything empirical; so the binding must be a priori, i.e., the law and the concept of a fully rational will are necessarily bound together, not contingently connected by mere happenstance; a metaphysics of morals is an a priori investigation into the idea and principles of a possible pure will, a will directed only by reason (see xii.6-8 (4:390.34-5)); so only a metaphysics of morals could reveal such an unconditional and universal binding, i.e., could reveal the moral obligation of every rational being to judge its actions according to maxims that it can will as universal laws.

If Kant's reasoning in this paragraph is anything like the reasoning just outlined above, then he seems to be arguing that only a metaphysics of morals can reveal how it is possible for reason alone to determine in an a priori way the conduct that the will initiates. And the suggestion, at this point of the overall argument, is that reason can do this because reason is the source of the law and there is an a priori connection between law and will. Being a priori, this connection is not forged by anything empirical, not by desires, inclinations, customs, habits, or by random events and personal experiences. The connection is direct and immediate, depending only on one's status as a rational being endowed with a will. It is the possiblity of such a connection that Kant still needs to establish by taking a step into a metaphysics of morals.

In this paragraph, Kant starts building the specific conceptual framework for the next formulation of the categorical imperative which he gives a few paragraphs further down the road at 66.21 (4:429.10). After giving us another way, other than as practical reason (see 36.21 (4:412.29-30)), in which to think of the will, Kant introduces the following concepts:
end: what serves the will as the ground of its self-determination; if this end is given only through reason, then it must equally hold for all rational beings and is an objective end resting on a motive that holds for all rational beings; if the end rests only or also on non-rational incentives, then the end is subjective
means: what merely contains the ground of the possibility of an action that has an end as its effect;
incentive: the subjective ground of eager desires; incentives can only be the basis for subjective ends;
motive: the objective ground of willing; a motive can be the basis for an objective end;
formal practical principle: a practical principle that abstracts from all subjective ends, leaving only objective ends resting on motives that hold for all rational beings;
material practical principle: a practical principle that has as its basis a subjective (material) end, a relative end that rests on incentives that appeal only to this or that rational being.
Kant does not use all these concepts consistently throughout the work. For example, he sometimes uses 'motive' in contexts where, according to the descriptions he has given here, it would be more appropriate to use 'incentive'. In any case, what is most important to remember from this paragraph is that subjective, material, and relative ends are grounds only for hypothetical imperatives, not the categorical imperative, and why Kant thinks this is so (see 63.18-20 and 64.9-12 (4:427.23-4 and 427.36 - 428.1)).

Although there is no explicit argument in this paragraph, Kant might have intended that, the implied argument being so obvious, the reader supply the argument herself. Since stating the obvious can nevertheless sometimes be useful, I venture that the argument might run something like this: the ground of the determination of the will of a rational being can operate either through a hypothetical or categorical imperative; hypothetical imperatives depend on subjective, relative ends having only relative worth; categorical imperatives depend on objective, absolute ends having absolute worth; so if there were something whose existence in itself has absolute worth such that as an end in itself it can be the ground of determinate laws, then only in that something would lie the ground of a possible categorical imperative. So the argument is a kind of disjunctive syllogism. One disjunct, signifying the hypothetical imperative, was eliminated in the previous paragraph, leaving now only the other disjunct signifying the categorical imperative. Even if this argument works, it is important to note that Kant has only established a conditional conclusion: if there were an end in itself that has absolute worth and that could be the ground of determinate laws, then only in it would lie the ground of a possible categorical imperative. He has yet to establish the antecedent of this condition, namely, that there is such an end in itself.

In the previous paragraph, Kant arrived at a conditional statement. The antecedent of the conditional consists of three items: something whose existence in itself has absolute worth; something that is an end in itself; something that, as an end in itself, can be the ground of determinate laws. The consequent of the conditional is that only in the something specified in the antecedent would the ground of a possible categorical imperative lie. To make use of this conditional, Kant needs to find the something mentioned in the antecedent; if he can find it, then he will be able to conclude the consequent that the something he found is the ground of a categorical imperative. In the present paragraph, it looks like this is indeed Kant's strategy. He begins the paragraph with the assertion that the human being and a rational being in general exist as an end in itself. Later in the paragraph, he infers that rational beings are (called) persons because their nature marks them out as ends in themselves. If this reasoning checks out, then Kant has got hold of one of the items in the antecedent: it looks like persons are shaping up to be the somethings mentioned in the antecedent. The rest of the paragraph does not manage to establish the remaining two items in the antecedent, but Kant does seem to be moving towards them when he discusses absolute worth and practical principles, seemingly saying that without persons as ends in themselves there would be no absolute worth and without absolute worth no highest practical principle.

In this very important paragraph, Kant arrives at the third formulation, the humanity or persons-as-ends formulation, of the categorical imperative. The paragraph can be understood as having three sections: a conclusion about a principle, a possible argument for the principle, an emphasized statement of the principle.

The paragraph begins with a conclusion drawn apparently from assertions contained in the previous paragraph. This conclusion tells us what a highest practical principle, if there is to be one, must be like. Such a principle, which would also be a categorical imperative for beings such as ourselves who have imperfect wills, must be the kind of principle that constitutes an objective principle of the will. This constituting of an objective principle comes about from the representation of that which necessarily for everyone is an end and is such an end because it is an end in itself. This objective principle, duly constituted in the specified way, can then serve as the universal practical law. This conclusion, then, with which the paragraph begins, characterizes a possible highest practical principle as one that has been made into an objective principle by thinking of what is an end in itself and therefore of what is necessarily an end for every rational being.

In the second section of the paragraph, Kant might be giving an argument for the actual existence of the possible principle characterized in the first section of the paragraph. In barest outline, the argument might go something like this: rational nature exists as an end in itself; I, as a (rational) being with a rational nature, must conceive of my own existence as a rational nature; (we postulate that) all other rational beings must conceive of themselves as rational natures for the same reason that I must conceive of myself in that way; so we have the representation or thought of something (namely, ourselves as rational natures) that is an end for everyone because it is an end in itself; so from this representation can be constituted an objective principle. There is a great deal going on in this very compact argument: there is movement from the individual to the universal, from me to everyone; there is movement from the human kind of rational being to any kind of rational being; there is movement from the subjective to the objective; and there is movement from a single ground to all laws of the will.

In the third section of the paragraph, Kant explicitly and emphatically states the third formulation of the categorical imperative. This third imperative is supposed to be a formula for the objective principle that Kant has apparently just argued for in the second section of the paragraph. Although they are not obscure, the following notes are still worth making: first, the imperative commands not just that we only use others always as ends in themselves but also that we only use ourselves as ends in ourselves; second, the imperative does not prescribe that we never use others as means but only that we never use others merely as means. These two notes can perhaps be combined by saying that the consideration of any rational being as an end in itself may never be left out of account.

In this "paragraph," Kant references the four examples given from 53-56 (4:421-23).

In this paragraph, Kant gives his first example illustrating the operation of the third formulation of the categorical imperative. The example also appears to contain an argument. This argument, when attached to the content of the example, is evidently supposed to elicit the reader's agreement that if there is a categorical imperative then the imperative points to a duty not to commit suicide in order to flee from difficult circumstances. To get the argument and the example to work together so as to elicit this agreement, the argument and example need to make contact at some point. This contact point occurs with using a person as a mere means. The example alleges that to kill oneself in order to flee a difficult circumstance is to use a person merely as a means to preserve a tolerable condition until the end of life. And one of the argument's intermediate conclusions, given above in the analysis, is that the human being is not something that can be used merely as a means.

Through the example and arguments in this paragraph, we learn more about what it means to treat or use a person as an end in herself and not merely as a means. It emerges that she must in some way contain or share the end of the person who is acting toward her, though it is also not super clear what this might mean. Kant's language is not sufficiently precise to pin it down. The verb he uses is 'contain' ('enthalten'), and the usage is encumbered with issues of time, agreement, and possibility. At one point he seems to say, for instance, that if the person does not contain the end of the action in herself at the same time (68.1 (4:429.32)), then she is treated merely as a means. Does this mean that she is still treated merely as a means if she, at a later time, would contain the end of the action? Kant at 68.4-5 (4:429.34-430.1) also seems to be thinking of the concept of containment at least in part in terms of possible agreement, but then says nothing more about possible agreement. So here we have again another case in which the argument is too short and incomplete but also in which there is a glimpse of a promising line of thought. These are typical features of the Groundlaying and are the marks of an enduring classic, of a text that has an abundance of insight but that leaves the details for others to work out. And this is indeed what has happened; the humanity formulation, especially, of the categorical imperative has been a very fertile ground for subsequent research in moral philosophy.

In this paragraph, Kant illustrates a contingent or meritorious duty to oneself. The duty is to accede to predispositions to greater perfection that exist in humanity. Kant actually never speaks here of talents; but because he introduced (at 67 (4:429)) this set of examples as paralleling the earlier (53-56 (4:421-23)) quartet of examples, and because he spoke there of talents in connection with this kind of duty to oneself, this duty to accede to such predispositions is typically interpreted as a duty to develop one's talents. Through this example, we learn that, to comply with the humanity formulation of the categorical imperative, an action must harmonize with humanity in our person. We also learn that the action must further humanity as an end in itself. Unfortunately, and as before, Kant does not elaborate; we are left to figure out for ourselves what harmonizing and furthering humanity might mean.

In this paragraph, Kant gives the fourth example of a duty, this time of a meritorious duty to advance the ends (happiness directly or intermediate ends that contribute to the final end of happiness) of others. Kant emphasizes that to comply with the requirements of the humanity formulation of the categorical imperative, there must be a positive agreement with humanity as an end in itself. It is not enough that one's action does not subtract from the happiness of others. Kant might even be interpreted here as insisting that one's action must have an impact on the happiness of others and that the impact must increase the happiness rather than leaving it at the same level. Although requiring this kind of impact is very demanding, it is a possible, and perhaps even plausible, way of making sense of what Kant means by a 'positive agreement'.

The main event in this paragraph is Kant's arrival at the fourth formulation, now typically called the formula of autonomy, of the categorical imperative: the idea of the will of each rational being as a will giving universal law. Kant identifies this as the 'third practical principle of the will' (70.18-9 (4:431.14-5)) and on the next page (71.24 (4:432.2)) as the 'third formula of the principle'. As we know from page 37 (4:413), principles and formulas are not the same thing; so it would seem that Kant is being rather loose with his terminology here. It is also possible that when he wrote this particular passage, he had lost count of the number of formulas he had already given; or perhaps he was thinking of principles rather than of formulas. In any case, as a result, scholars are somewhat divided over the actual number of categorical imperatives. I have identified this occurrence as the fourth formula of the categorical imperative, counting the universal law (52 (4:421)), the universal law of nature (52), and the humanity (66 (4:429)) formulations as the first three. Other scholars (and perhaps Kant, too) view the universal law of nature formula as so close to the universal law formula that they categorize the former as a variant of the latter formula, thus arriving at only three formulas so far.

In this paragraph, Kant introduces, though not by name, the concept of autonomy. At this early stage, the concept appears as self-lawgiving ('selbstgesetzgebend' at 71.1-2 (4:431.22)). Kant wants to emphasize in this paragraph that the will is not just subject to or bound by the moral law. The will is also the author of the moral law; moreover, precisely because it is the author, the will is subject to the moral law: by its authorship of the moral law, it imposes that moral law on itself.

In this paragraph, which is chiefly a summary of assumptions already made and of their consequences, Kant makes it clear that the categorical imperative, especially this latest formulation of it, excludes all interest as an incentive. The moral law and the categorical imperatives that express it are not binding on us because of any interest that we might have. Kant reminds us why this is so: he is investigating the concept of duty (8.11 and 36.14 (4:397.6 and 412.24)) and this investigation requires that we assume that the categorical imperative is indeed categorical, that is, is the moral imperative (43.14 (4:416.13)) commanding us to act from duty; this assumption then has the consequence that all admixture of interest as an incentive be excluded from the categorical imperative's commanding authority to specify our duty (13.14 and 38.29 (4:399.35 and 414.35)). But now Kant points out that only this last formulation of the categorical imperative makes this exclusion of interest clear. It makes it clear by tying the authority of the imperative, its obligatoriness, to the authorship of the moral law rather than tying the authority to an interest of some kind. Before moving on to argue for this claim that the latest formulation of the categorical imperative really does indicate this exclusion of interest and that this exclusion marks the difference between hypothetical and categorical imperatives, Kant reminds us again that the assumption that there really are practical propositions which are categorical imperatives has not yet been proven.

In this paragraph, Kant gives us more information about the relation between will, interest, and the moral law. He makes a distinction between a will that stands under law and a will that is highest in law-giving. The difference is that a will standing under or subject to law can be bound to the law by an interest; the interest is what binds the will to the law, what makes the law obligatory for the will. This is the case, for instance, with hypothetical imperatives which prescribe that I do X because I want Y. My interest in the thing (Y) that I want binds my will to the imperative so that I will to do X in order to get Y. But this is not the case with categorical imperatives, in which the will is acting as a legislator for all rational beings. Kant argues that in this kind of case, a will that is giving law at the highest level, for all rational beings, cannot be dependent on an interest such as a want or a desire; it cannot be dependent on an interest because any interest might not be shared by all rational beings. The only way to make sure the interest is shared by all would be to limit the interest to the condition of a validity for universal law. But then, if the interest is constrained by universal law, it is really the law upon which the will is dependent and not the interest. And, finally, because the will is also the author of the law, the will is really only dependent on itself as a rational nature.

In this paragraph, Kant helps us see more clearly why this autonomy formulation of the categorical imperative has an advantage over the other formulations as far as the exclusion of interest is concerned. Kant gives us this help by converting the proposition and by saying that the conversion is in some way even better.

The proposition, first of all, might be something like: if the commanding principle of each human will is that of a will giving universal law through all its maxims, then it (i.e., the principle) is the basis for a categorical imperative. The converted, or turned-around, proposition is then something like: if there is a categorical imperative, then it can only command everything be done from the maxim of one's will as a will that could view itself as giving universal law. For both propositions, the support is the same: the principle is not based on any interest and therefore is unconditioned (and hence suitably categorical).

The question now is why the converted proposition is even better. Kant does not tell us why it is better, but there are a couple of ways in which the converted proposition sits better with Kant's overall strategy and sits better with Kant's specific point here about the exclusion of interest. It sits better with Kant's overall strategy in that, by putting the reference to the categorical imperative in the antecedent of the conditional, the conversion makes it clear that Kant has not been asserting that there is a categorical imperative. The converted proposition sits better with Kant's specific point about exclusion because in the converted proposition the will is represented as being aware (looking on itself, having itself as object) that it is legislating universally. Mentioning this awareness tends to bring to the mind of the reader the self-legislation and authorship elements in the autonomy formulation and thus tends to bring to mind the rational self, not interest, as the source of the necessity, the obligatoriness, of the moral law.

In the paragraph, Kant gives an explanation of why all previous efforts to discover the principle of morality had to fail. They had to fail, Kant explains, because it never occurred to anyone that the moral laws to which a human being is bound by duty are the human's own, authored and issued by the human's own will, and yet are also universal in scope, applying to all rational beings. And this never occurred to anyone because they were always thinking of humans only as subject to the law. Thinking in this way necessarily leads to the mistaken view of duty that some interest, an enticement or constraint, has to be what makes the law binding on humans. Thinking in this way, as only subject to the law, necessarily leads to that mistaken view because the obligatoriness of the law is by this way of thinking seen as coming from something else, from something other than the individual human will. Finally, the resulting view is mistaken because what we arrive at by this thinking is not duty; instead, we arrive at a necessity to act from a certain conditional interest which, no matter to whom it belongs, can never be the basis for a moral command.
It should be noted that this explanation, though ostensibly about why all previous attempts to find the principle of morality had to fail, also manages to offer a fuller account of why or how authorship of the law excludes interest. Kant initially raised this issue of exclusion back on page 71 (4:431), but he did not elaborate. With this explanation, he has somewhat rectified that omission. Authorship of the law makes it binding on the author (and everyone else) without the need of any associated interest because the law is then not seen as something foreign imposed on the will from the outside; instead, the law flows out of the will which, as a rational will, guided by principles, constitutes and structures itself through its own legislating and regulating activity.

Though there is no argument in this paragraph, one should not pass over it too quickly. At the most general level, the paragraph is just the statement that one concept leads to another. But there are a number of interesting features built into these concepts. The first concept, that of a rational being, includes:
a. the necessity that the rational being consider itself in a certain way so that it is not enough that the rational being does so consider itself;
b. the giving of universal law through all, not just some, maxims of the rational being's will;
c. the rational being's consideration of itself, an awareness of its own activity.;
The status of the connection between this first concept and the concept of an empire of ends (sometimes alternatively called a kingdom of ends or a realm of ends) is unclear; Kant only says that the first leads to the second and that the second depends on the first. Beginning in the next paragraph, Kant starts to clarify the connection. This clarification is especially important because the first concept (of a rational being) is closely aligned with the autonomy formulation of the categorical imperative, the second concept (of the empire of ends) with the empire formulation, so that if the former concept leads to the latter concept there might also be a connection of some kind (perhaps a logical connection) between the two corresponding formulations of the categorical imperative.

In the previous paragraph, Kant stated that a certain concept of a rational being leads to the concept of an empire of ends. In this paragraph, he begins to expand on the latter concept, explaining what he means by it and arguing how it is possible to think such a concept. By an empire Kant means the systematic union of different rational beings through common laws. This seems to say that an empire exists when various rational beings have banded together and the resulting group is organized and regulated by a set of shared laws. Kant then begins to argue that, if one first thinks away the differences (e.g., physical appearances and personal tastes) between rational beings, it is possible to think such an empire because law determine ends according to their (i.e., the law's) universal validity. Kant's thought process here might be that by thinking away, abstracting from, personal differences between rational beings, only ends compatible with universal law will remain; these compatible ends will then be amenable to coordination and regulation by universal law.

The conclusion apparently arrived at in this paragraph is very similar to the conclusion arrived at in the previous paragraph. In this paragraph, the conclusion appears to be that by treating rational beings as ends in themselves an empire of ends arises only as an ideal. The previous paragraph's conclusion was that if we abstract from personal differences and private ends of rational beings, then it is possible to think an empire of ends. These conclusions are very similar because existing only as an ideal would seem to mean existing only in thought. The similarity of the conclusions suggests that there might really only be one argument here spread across two paragraphs. This one argument would then have three premises, one from the previous paragraph and two from the current paragraph, supporting the one conclusion.
It should perhaps be noted that Kant's rendition of the humanity formulation of the categorical imperative here (as the first premise which begins the paragraph) is slightly different from earlier renditions. Kant here uses the word 'behandeln' ('treat') rather than 'brauchest' ('use') as at 67.1 ('brauchst' at 4:429.12), while in the first two examples (at 67.9 and 68.1 (4:429.19 and 429.32)) he used conjugations of 'bedienen' ('make use of'). So one might ask whether there is any philosophical difference between 'using a person as an end in herself', 'treating a person as an end in herself', and 'making use of a person as an end in herself', and between 'using a person only as a means', 'treating a person only as a means', and 'making use of a person only as a means'.

Rather abruptly, Kant here categorizes rational beings that belong to an empire of ends as either a member or a head. In this paragraph at least, the only difference between a member and head is that a member is subject to the laws whereas the head is as lawgiving subject to the no will of another. This is not as clear as we would like. Several questions come to mind. Are these categories of member and head exclusive, or can a rational being be both at the same time or at different times? Can a rational being be promoted or demoted, so to speak, from one category to the other? Can there be more than one head? Is there a limit to the number of heads? Could all members of an empire of ends be heads so that there would be no members? Is it correct to say that being subject to no (lawgiving) will of another is equivalent (in some sense) to being subject to no laws? Could there be a law given by the will of only one rational being, and would other rational beings be subject to that law? Would it make a difference if that law could only be given by that one rational being? What does it even mean to say that one is not subject to the will of another? Is any special importance to be attached to the 'as lawgiving' ('als gesetzgebend' at 75.12-3 (4:433.37)) qualifier in the statement about the head?

In this paragraph, we learn a bit more about a rational being who is either a member or head in an empire of ends possible through freedom of the will. In particular, we learn that both member and head must consider themselves always as lawgiving. This apparently means that, when they consider themselves as rational beings with freedom of the will and in an empire of ends, the consideration must be of themselves as lawgiving. We also learn that a rational being can maintain the place or seat of a head only if the rational being is a fully independent being that has no needs and whose power, adequate to its will, is not limited. This seems to mean that a head cannot remain a head merely through the maxims of its will; a head must also be completely independent, without needs, and without any restrictions on its power, a power which is sufficient to accomplish what it wills. Finally, there is no indication in the paragraph that a rational being who has all these qualities of a head in an empire of ends is or would remain a head; these qualities are necessary to be a head but not necessarily sufficient.

It is easy to see that there is quite a lot going on in this paragraph, but it is not so easy to see what kind of structure to give to it.

The paragraph opens with a conclusion, but it just hangs there, no clear support either before or after it. One supposes that the conclusion is to follow from what Kant said about the empire of ends two and three paragraphs prior (skipping over the two immediately preceding paragraphs about member and head) to the current paragraph. In those paragraphs, Kant characterized an empire of ends as arising from shared laws among individuals who have banded together into a group. Kant might be thinking, though he does not say so, that morality is to be understood as a set of laws regulating the actions of such individuals in a group. Adding in this unstated conception of morality gives the conclusion at least some support. It is also possible that Kant instead, or perhaps in addition, wants to support the conclusion by thinking of morality as a set of duties, and this is why in the latter half of the paragraph he takes the time to say that duty applies to members of an empire of ends.

More interesting are the inferences in the paragraph, though here, too, how exactly they are to work is mysterious. They are interesting because they suggest logical connections between the various formulations of the categorical imperative. The first inference (signaled by the paragraph's second 'also' at 76.3 (4:434.10))goes roughly from lawgiving, a key component of the autonomy formulation, to universal law as in the first formulation. The second inference (signaled by the paragraph's third and final 'also' at 76.5 (4:434.12)) then goes from universal law back again to universal lawgiving. The reciprocity of these inferences suggests that the two formulations are equivalent in some way, perhaps logically equivalent.

It is also something of a mystery how Kant arrives at the claims, at the end of the paragraph, that duty does not apply to the head and that duty applies equally to each member in an empire of ends. As to the former claim, Kant has already informed us that the head is a fully independent being; he might then be putting this information together with earlier claims (see, for example, 8.13, 39.12 (4:397.8, 414.9)) to the effect that inclinations afflict dependent beings having imperfect wills and that inclinations are thus hindrances to achieving agreement between subjective maxims and objective principles. The result is that the head, as fully independent, is not afflicted with such hindrances so that its maxims necessarily agree with objective principle, and therefore the head does not experience the necessitation of duty. How the latter claim of equality, or of duty applying in equal measure to members, comes about is even more mysterious. Perhaps Kant was thinking along these lines: duty is necessity to act out of respect for the law; the law originates in rational beings insofar as they are rational; rationality is the same wherever it is found; so all rational beings are equal insofar as they are rational; duty applies to rational beings only insofar as they are rational; so duty applies to all rational beings in the same measure, that is, equally.

This paragraph is especially notable for two reasons. First, in the premises, there is an inference (signaled by 'weil' at 76.20 (4:434.24)) from a rational being otherwise not being able to think of other rational beings as ends in themselves to a rational being having to consider the relation between rational beings as a relation between lawgivers. This inference suggests that there might be a logical connection of some kind between the autonomy formula of the categorical imperative and the humanity formula. Expressed as a conditional, the inference goes like this: if the rational beings in the relation do not have to be considered always at the same time as lawgiving, then they could not be thought as ends in themselves. Contraposing this, we get that if they can be thought as ends in themselves, then they must be considered always at the same time as lawgiving. If this is correct, then being able to be thought as ends in themselves is sufficient for having to be considered always at the same time as lawgiving, and having to be considered always at the same time as lawgiving is necessary for being able to be thought as ends in themselves. So it looks like the humanity formula might be logically dependent on the autonomy formula insofar as the latter is necessary for the former.

Second, the paragraph is notable for introducing the idea of dignity. Kant does not say precisely what dignity is, but, given the context in which it is introduced, there is a hint that dignity attaches to the capacity to give (universal) laws to oneself and to obey only those laws that are, or can be, so given.

In this paragraph, Kant states that everything in the empire of ends has either a price or a dignity. These two categories seem to be presented to us as exclusive of each other: if something has a price, then it does not have dignity; and if something has dignity, then it has no price. Kant also seems to give us a functional way to determine if something has a price: if X can be put in place of Y as an equivalent, then X has a price (and so does Y since X and Y are equivalent). But to make this really functional and practical, we would also have to know exactly what it means to put something in the place of something else and also have to know what kind of equivalence is required.

In this paragraph, Kant identifies two kinds of price: market price and affective price. There is no indication here that this division is exhaustive; there might be other kinds of price, too. Kant, however, does seem to give us another way, in addition to the equivalency method given in the previous paragraph, to think of price in general; he seems to want to say that everything that has a price has only a relative worth, perhaps even wanting to say that price is relative worth. This relative worth he sets against inner worth. Kant does not say exactly what inner worth is, but at the end of the paragraph he seems to identify it with dignity. (It is possible, though, that Kant only means that whatever has inner worth also has dignity.) So now, adding in what we learned about dignity in the previous paragraph, we have that dignity is basically the (inner) worth that attaches to the capacity to make universal law and to obey only such law. Almost in passing, Kant also appears to let us know something very important: that whatever has this inner worth constitutes the sole condition under which something can be an end in itself. So what is it that has this inner worth that is dignity and that is that sole condition? It is the capacity. The capacity to make universal law and to obey only such law has that inner worth and is the sole condition for something to be an end in itself. So if something, for example a rational being, has this capacity, and if worth is transferable from capacity to possessor of that capacity, then the rational being has inner worth, has dignity, and is an end in itself.

It should be noted here that Kant seems to have backed off the apparent exclusivity of price and dignity suggested in the previous paragraph that first introduced the distinction. That paragraph reads as if what has a price cannot have dignity (and vice versa) because what has a price has an equivalent while what has dignity has no equivalent. But in this paragraph, Kant says that what constitutes the condition under which alone something can be an end in itself has not merely a relative worth, a price; it has an inner worth, dignity. This 'not merely' ('nicht bloß' at 77.16 (4:435.3)) suggests that something that has price might also have dignity.

In this paragraph, we learn still more about dignity, which is turning out to be a very complex concept. In particular, we learn that dignity has to be more broadly construed than merely as the worth of the capacity to make universal law and to obey that law; dignity is also the worth that attaches to the condition, namely morality (or the moral way of thinking), which enables the capacity, which makes it possible to put the capacity into actual practice in concert with other rational beings considered as ends in themselves. It remains to be seen whether moral actions, too, have dignity. Kant says they have inner worth, but this is not enough to conclude that they have dignity; for there might (though 77.17 and 79.14 (4:435.4 and 436.3) suggest not) be other kinds of inner worth besides dignity.

We also learn something new, or perhaps only confirm something that we have suspected for some time, about maxims. We learn that maxims are, or can be, dispositions (78.8 (4:435.15)). This is important to note, because it means that maxims do not have to be consciously formulated. Rational beings can be said to act from a maxim even if the rational being has not explicitly given linguistic expression to that maxim.

Kant begins this paragraph with a question, prompted apparently by the discussion in the previous paragraph: what entitles virtue to make such lofty claims? Then, essentially summarizing the points made in the previous paragraph, he answers the question by saying that the entitlement is due to the part in universal lawgiving that virtue allows to the rational being. Kant then supports this answer with a new argument. The key premise in the argument is perhaps the claim that nothing has a worth except that which the law determines for it. What this claim means is not at all clear, and Kant again does not elaborate. At 36.16 (4:412.26), however, Kant did say that everything in nature works according to laws, and he seems comfortable with a general teleological view (see e.g., 4.14-19 (4:395.4-7)) of nature. Perhaps, then, what he means to convey by this key premise is that the value or worth of anything derives from its proper operation or functioning well. And this functioning is directed or governed by laws. So the worth of anything is ultimately due to the laws regulating the operation of that thing.

Also of note in this paragraph are the hints of connections between the various categorical imperatives, or at least between concepts central to those various formulations. At 79.5-7 (4:435.33-5), Kant states that the rational being was already determined by its nature to be an end in itself and for just that reason to be lawgiving in an empire of ends. This statement could be interpreted as saying that being an end in itself is the basis for being a lawgiver in an empire of ends, for it gives the former as a reason for the latter. This interpretation suggests that the humanity formulation might be logically more fundamental than either the autonomy or empire formulations. (Note that this is not what was suggested in my commentary for paragraph 66 above.) On the other hand, perhaps Kant intended that we can conclude a rational being's status as lawgiver in an empire of ends from the rational being's status as an end in itself because the former (i.e., lawgiver) is necessary for the latter (i.e., end in itself), so that, having the latter, the former can safely be concluded. Finally, there is the hint at the end of the paragraph, that autonomy, as the ground of dignity, is what is most fundamental. Perhaps Kant was thinking of autonomy as the capacity of a rational being with a will to be the author and legislator of laws. This capacity is then most fundamental because it provides the basis for authorship and thus the original source for the unconditional value that characterizes an end in itself; and it is most fundamental because the capacity provides the basis without which law-giving activity could not take place.

Kant begins this paragraph with a reference to three ways of presenting the principle of morality. These three ways apparently refer to three of the five formulas of the categorical imperative that he has given so far. The five are:
1. universal law
2. universal law of nature
3. humanity
4. autonomy
5. empire of ends.
The three Kant has in mind are probably 1, 3, and 4. Because 2 is so close to 1, Kant probably views 2 as a minor variation on 1. And because 5 is very close to 4, both explicitly making prominent use of the concept of a lawgiving, Kant perhaps takes 5 to be a minor variation on 4. Kant says of the three that they are at bottom only so many formulas of the same law; he would presumably say the same of the two variations, too. Unfortunately, Kant does not go on to say more about the relation of the formulas to the one law and does not say what the one law is. Instead, he gives us a few words about the relation of the formulas to each other: each one (of the three) by itself unites in itself the other two. Kant again does not elaborate, moving on instead to a difference between the formulas. This difference seems to have to do with how closely the various formulations bring an idea of reason to intuition and then to feeling.

One of the puzzles in this sequence of paragraphs is why Kant abruptly switches gears here and starts talking about maxims rather than continuing his discussion of formulas. The connection between the two is evidently that formulas are to help us select and then act from maxims that are compatible with the formulas. So perhaps Kant means to convey here that these features of maxims, form being the first listed, provide a kind of hook or point of application which formulas can latch onto in order to render judgment on the maxim.

In this paragraph, Kant says that all maxims have a matter. This matter is an end; so all maxims aim at some end. But whatever relative and voluntary end a maxim aims at, the rational being must, according to the humanity formulation of the categorical imperative, serve as the limiting condition of that maxim's relative end. This would seem to mean that the maxim's relative end (i.e., the contingent goal it aims to achieve) must be compatible with treating those rational beings affected by the actions undertaken as means to that end, and affected by the end itself, as persons or ends in themselves.

Kant starts this paragraph by listing the third feature that all maxims have in common: a complete determination of all maxims through the autonomy and empire of ends formulas. Having listed the three features that all maxims have in common, Kant speaks of the progression as through the categories of unity, multiplicity, and allness. These three categories, which are probably references to some of the categories from the Critique of Pure Reason, seem to parallel the three listed features of maxims. Kant then switches back to the earlier point (made before the listing) about bringing an idea of reason closer to intuition and by that to feeling. In particular, he says that for moral judging it is better to proceed by the strict method, using the universal law formula; but if access to the moral law is to be obtained at the same time, then it is very useful to lead the same action through the three named concepts, in this way getting closer to intuition. By this progression, then, Kant might be referring to a kind of handy procedure for evaluating maxims and the actions specified in them. Maxims, because they have the three listed features, can be considered from the perspective of each of the three parallel categories. Each category is tied to a listed feature, unity to form, multiplicity to matter, and allness to a complete determination, and each of those features is tied to a formula. Because of these linkages between category, feature, and formula, by leading a maxim and its specified action through all three concepts of unity, multiplicity, and allness, we in effect fully and intuitively evaluate the maxim in terms of its compatibility with each of the three above-mentioned formulas of the categorical imperative.

In this paragraph, Kant reviews the arguments for the two universal law formulations of the categorical imperative. These categorical imperatives were originally introduced beginning at 52.3 (4:421.6). Several items are noteworthy.

First, the arguments expressed here (and in the review generally) are not exactly the same as arguments expressed earlier in the text. So there are several possibilities: the arguments in the review are new; the arguments in the review are the same as some of the arguments earlier in the text, but they have been re-expressed so that they appear new; the arguments in the review are not really arguments but reports of arguments.

Second, the categorical imperatives in the review frequently are not phrased in exactly the same way as they were earlier in the text. This is very obviously the case, for instance, with the universal law of nature formula here in which it is required that the maxim have itself as a universal law as an object; this reference to an object is not present in the version given at 52.19 (4:421.18). Kant also identifies this latest version as the formula of an absolutely good will; so this latest version might actually be a sixth formula.

Third, with regard to the first argument here, the conflict mentioned in the first premise (statement 1 above) is between the maxim and itself while in the second premise (statement 2 above) the conflict is between the will and itself.

In this paragraph, Kant reviews the argument (originally given differently on p. 66 (4:428-9)) for the humanity formulation of the categorical imperative. As a glance at the analysis of the argument reveals, the argument is very complex, and it might not have the precise structure indicated in the analysis. It might be, for instance, that statements 8 and 9 in the analysis jointly derive statement 10 rather than 10 following directly from 9 alone. And it seems like statement 7 has a role to play in getting to statement 10, too, but it is not clear how it fits in the argument. In any case, to make progress in understanding the argument, it is important to keep the following identities in mind: the will is the will of a rational being; a rational being is itself an end in itself and can have other ends besides itself; an end in itself is a rational being; a subject of a will is a rational being; a subject of ends is an end in itself. A shorter way to think of these identities is perhaps to think of the following quadruple identity: a subject of a will is a rational being is an end in itself is a subject of ends.

In this long paragraph recapping previous results, Kant first establishes the possibility of an empire of ends which is analogous to the actual empire of nature. This analogous relationship might be the same as that in the references to an analogy at 79.26 and 81.23 (4:436.12 and 437.16). Kant realizes that the analogy is not perfect; there are some dissimilarities, the most notable of which is that the empire of ends is merely possible or ideal (see 75.8 (4:433.32)) whereas the empire of nature is real. Despite this mere possible existence, Kant manages to arrive next at the position that the law expressed in the empire of ends formula remains in full force, apparently because the empire of ends would exist if the maxims through which it can become actual were universally imposed as prescribed by the categorical imperative. The paradox associated with the full force of this law recalls Kant's earlier discussion of the strange idea of an absolutely good will (see p. 4 (4:394-5)), a will now recognized as determined only by respect for laws anchored in the idea of autonomy, not a will determined by desires, wants, or needs affected by the effects of actions or by any other purely contingent factors (such as some incentives, impulses, and motives) that can influence the will. Though paradoxical because it is hard to understand how the will can be directed only by an idea, the laws in an empire of ends and the idea of autonomy underlying those laws bring morality into focus as the relation of actions to possible universal lawgiving through maxims of the will. This conception of morality, along with its concepts of obligation and duty, remains possible despite the paradox because the external relations of things, such as the effects that our actions can have, play no role in assessing the absolute worth of the human being.

In this paragraph, having finished the review of his progress in the search for and establishment of the highest principle of morality, Kant is now in a position, drawing on what has been shown so far, to explain some things. First, there needs to be an explanation of how under the concept of duty there is thought a subjection to law but also represented at the same time a sublimity and dignity of the person who fulfills all her duties. Kant's explanation of this seeming incongruity under the concept of duty is that a person, an autonomous rational being, is not just subject to the law; such a being is subject to the law only because it is a lawgiver, and it is this status as lawgiver that accounts for the sublimity and dignity represented under the concept of duty. Explaining further, Kant adds that respect for the law is the incentive which can give action a moral worth. And we can see how sublimity and dignity can arise out of this respect for the law when we realize that the idea of a will giving universal law is the proper object of respect insofar as the dignity of humanity consists just in this capability to be a lawgiver to oneself. In other words, when we realize that in acting from respect for the law we are actually, at another level, also expressing respect for ourselves as lawgivers, we also come to the realization that in fulfilling our duties we at the same time manifest our sublimity and dignity.

In this paragraph, Kant explicitly identifies the principle of the autonomy of the will as the highest principle of morality. So he has now found what he set out looking for. This principle apparently underlies all of the formulations of the categorical imperative, not just the autonomy formula. All the various formulas merely express this principle, through commands, as imperatives. As if to emphasize this, Kant starts the paragraph by pointing out that autonomy is the property of the will by which the will is a law to itself and then by explicitly expressing the principle in wording reminiscent of the universal law formulation of the categorical imperative. It is not altogether clear what exactly the principle is, for it is not stated as a complete sentence (i.e., what follows the colon at 87.13 (4:440.18) is not a complete sentence because a grammatical subject is missing). But the principle is presumably something like this: a rational being with autonomy of the will must always choose so that the maxims of its choice are comprehended in the same willing at the same time as universal law. This principle is, Kant says, a synthetic and a priori proposition. Kant also suggests an alternate version (87.16-18 (4:440.20-2)) which is also a synthetic a priori proposition. This alternate version refers to the practical rule (of choosing in a certain way) and runs something like this: the will of each rational being is necessarily bound to the practical rule as a condition.

There are three keys or steps to understanding what Kant means by calling the principle a synthetic a priori proposition. First, the principle needs to be broken up into three components, a subject concept, a predicate concept, and the linkage between the two concepts. The subject concept of the principle is 'a rational being with autonomy of the will', and the predicate concept is 'always choosing so that the maxims of its choice are comprehended in the same willing at the same time as universal law'. Now, second, by saying that the principle, which is a proposition, is synthetic, Kant means that analysis of the subject concept, no matter how far that analysis looks into the concept, does not reveal the predicate concept. In short, the subject concept does not contain the predicate concept. Third, by saying that the principle is also a priori, Kant means that the linkage between the two concepts is necessary and not contingent. In the principle, this necessity is conveyed by the 'must' (or in the alternate version by 'necessarily bound'). So, in saying that the principle is synthetic a priori, Kant is saying that the connection between the subject concept and predicate concept cannot be severed and yet is not a connection based on one concept containing the other. (The difficulty of ascertaining the exact nature of this connection is what prompted Kant at 50.11-18 (4:420.12-7) to list this difficulty as the second explanation of why the question of the possibility of a categorical imperative, unlike the other imperatives, is in need of a solution.) Because the principle is a synthetic a priori proposition, a full understanding of it requires a critique of the subject (i.e., a rational being with a will); this would be carried out in a critique of pure practical reason.

Although analysis of the concept of an autonomous rational being cannot reveal the concept of always choosing so that one's maxims can be willed as universal laws, Kant says that the principle of autonomy is revealed as the sole principle of morals simply by analysis of the concepts (such as a good will and duty) of morality. The following proposition is therefore analytic a priori: the principle of morals is the principle of autonomy. Kant has arrived at this by analyzing (see 8.4-16, 36.12-15, 51.1-8 (4:397.1-10, 412.22-5, 420.18-23)) the concepts of morality, finding in it the concept of the categorical imperative and then finding behind the categorical imperative the principle of autonomy.

In this paragraph, now that he has already found the highest principle of morality in the principle of autonomy, Kant begins his discussion of heteronomy. Heteronomy is a state of the will in which an object of the will, rather than the will itself (as in autonomy), gives law to the will. An object can do this giving of law through its relation to the will, but the result is never more than a hypothetical imperative that says that one ought to do something because one wills something else. The first of Kant's two examples has to do with lying and in it Kant contrasts what a hypothetical imperative says and what a categorical imperative says. The hypothetical imperative says that one ought not lie if one wants to retain one's honorable reputation; the categorical imperative says that one ought not lie even if one could get away with it without any damage to one's reputation. The important part of the example, however, is what comes next in Kant's discussion. He says that the categorical imperative must abstract from any object so that the object has no influence on the will. The object of the will in this particular example is presumably the honorable reputation (or the desire or inclination for it). In the hypothetical imperative, this object, the honorable reputation, exerts influence on the will through its relation to the will: it is because one wants the honorable reputation that one ought not lie. This influence amounts to heteronomy: the object, the honorable reputation, is giving the law not to lie to the will. The categorical imperative, in contrast, abstracts from any object, effectively ignoring the matter or incentive of an honorable reputation, so that the will is not influenced by the object and gives to itself and by itself (without a relation to anything else) the law not to lie. This freedom from influence characterizes autonomy, not heteronomy, and Kant's second example about others' happiness illustrates how the will uses this freedom: the will checks to see if the maxim not to promote others' happiness is compatible with willing the maxim as universal law. Presumably for the reason given or reported in Kant's fourth example of duties (56-7 (4:423)), such a maxim not to promote others' happiness cannot be willed as a universal law and so the will excludes such a maxim from universal lawgiving.

Kant uses this opening paragraph to set up his discussion of the various ways in which heteronomous principles go wrong. Kant's claim that human reason has previously (that is, as Kant would have it, before Kant) tried all possible incorrect ways is perhaps exaggeration. But it does get the reader wondering. Have all the wrong ways been tried already? Must all the wrong ways be tried first before the correct way is found? Why has human reason failed to find the correct philosophical account of the foundations of morality? Is human reason bound to fail in its search if it has not been corrected by a critique, corrected by the already-published (1781) Critique of Pure Reason, or by the yet-to-be-written (1788) Critique of Practical Reason, for which latter work Kant is now paving the way with this Groundlaying? If so, why is it bound to fail?

In this paragraph, Kant categorizes all principles from a certain point of view. The point of view is apparently that of a human reason that operates with principles. (Recall that one way of characterizing a rational being is to say that such a being can act from, or on the basis of, principles.) From this point of view, all principles are either empirical or rational. Empirical principles, from the principle of happiness, are built on physical or moral feeling. Rational principles, from the principle of perfection, are built either on the rational concept of perfection as a possible effect or on the concept of a self-standing perfection (such as the will of God). Any of these principles is supposed to be a determining cause of our will, a decisive factor in locking in the maxims that our will chooses to follow.

In this paragraph, Kant first argues that empirical principles are not suitable as the basis for moral laws because moral laws have universality and necessity. This is not a new argument; as far back as the Preface (see p. viii (4:389)), Kant argued that nothing empirical can establish any kind of universality or necessity. If empirical principles are not suitable, then they are presumably also objectionable. So the rest of the paragraph is probably not an argument designed to prove that empirical principles are objectionable. Rather, Kant is probably just explaining or describing the ways in which such principles are objectionable. Another reason for thinking that Kant is not arguing the point is that he does not make the case that any one of these specific ways is more objectionable than any other specific way; at most, he claims that one of them is most objectionable, but he does not argue for this claim.

In this paragraph, Kant singles out among the rational grounds for morality two concepts: the ontological concept of perfection and the theological concept of perfection. Kant does not explain either of these concepts, and does not identify by name any advocates of them. By the ontological concept he might mean a view according to which we are duty-bound to realize or actualize in ourselves as many perfections as possible. By the theological concept, he might mean a view according to which our duties are specified by and originate in the commands of an all-perfect being such as God. In any case, as a basis for morality, Kant finds both concepts objectionable. The ontological concept, he says, is empty, indeterminate, useless, and promotes circular reasoning. The theological concept fares no better; in fact, Kant thinks it worser. It tries to derive morality from the intuition, which we do not have, of a divine and perfect will. This leaves the concept in a jam: if we try to derive perfection from our own concepts, then we engage in circular reasoning, having to use a pre-existing moral standard to judge in advance that the divine will is morally perfect (recall pp. 29-30 (4:408-9)); if, on the other hand, perfection is not derived from our own concepts, then the result is a faux morality based on desires for glory and dominion and on representations of power and vengefulness. So the theological concept depends either on circular reasoning, as does the ontological concept, or depends, which in Kant's eyes is even worse, on desires.

In this paragraph, Kant explains why he would choose, if he had to make a choice, the concept of perfection in general over the concept of moral sense. (He does not even consider the principle of happiness, apparently because this principle, unlike approaches based on moral sense or perfection, does actual damage to morality. (See the first stretch of parenthetical material in the current paragraph and also x-xi (4:390).) Kant prefers the concept of perfection because it removes the decision of the question from sensibility and moves it to the court of pure reason. This removal protects the indeterminate idea of a will good in itself from contamination by empirical elements associated with sensibility and preserves it until the idea can be more closely examined by pure reason.

Kant explains in this paragraph why he is going to offer no extensive refutation of still other theories of the foundations of morality. He seems to think it sufficient to have already gone into some detail regarding the mistakes of the principle of happiness, of moral sense theory, and of perfection theories. He states that the refutation of any other theory is so easy and already so well-understood that repeating the refutation here would be superfluous. But Kant does end the paragraph by pointing out what does need to be known, namely, that all these other theories fail because their principles set up heteronomy of the will as the ground of morality.

In this paragraph, Kant appears to give an argument that is supposed to support the inference from the claim that an imperative expressing a heteronomous rule (i.e., a rule prescribed to the will on the basis of a natural affinity for an object) is conditioned to the claim that such an imperative can never command morally or categorically. That is, the argument is to fill the inferential gap between these two claims. There are two bits of textual evidence for this interpretation. First, the paragraph in which the argument occurs begins with talk of a rule (93.19-20 (4:444.1)) and ends with similar talk of a rule (94.24 (4:444.23)). Second, the argument itself, though very obscure in its precise structure, seems to try to make the case that, when an object determines the will, the object does so through some kind of natural affinity or receptivity that the subject has for the object; the upshot is that it is really nature that gives the law to the will. This reliance on nature then makes the rule that prescribes pursuit of the object into a contingent rule unsuitable to be an apodictic (i.e., absolutely necessary) moral rule. So, basically, an imperative that expresses such a heteronomous rule is conditioned because it depends on our natural makeup, which could be different, and so such an imperative cannot be a moral imperative that commands categorically for all rational beings.

Based on his discussion of the last few previous paragraphs, Kant concludes in this paragraph that the absolutely good will, which must have a categorical imperative as its principle and which is indeterminate with regard to all objects, will contain merely the form of willing in general; and this form of willing is autonomy. The part about such a will being indeterminate with regard to all objects (e.g., not decisively influenced by any object of desire such as a desire to achieve or maintain a certain social standing) is evidently due especially to his just-ended discussion of heteronomy which characterizes a will that is determined by objects. Kant elaborates a bit on the meaning of this conclusion. He says that the suitability of the maxim of each good will to make itself into universal law is itself the sole law that the will of each rational being imposes on itself, without putting underneath as a ground any incentive and interest of the maxim. Breaking this down, we seem to have the following:
1. Every good will is guided by some maxims.
2. These maxims have a suitability or fitness.
3. The suitability is a suitability of the maxims to make themselves into universal law.
4. This suitability is itself the sole law that the will of each rational being imposes on itself.
5. Such a will imposes the law on itself without making any incentive or interest of the maxim a ground of the law.
These five items, then, together capture what Kant means by the conclusion that opens the paragraph, the exclusion of incentive and interest in 5 leaving behind only the form of which Kant speaks.

Kant wraps up the Second Section by pointing out what has and has not been shown so far. The following have not been shown, established, maintained, or provided:
1. how a synthetic practical proposition a priori (such as the principle of autonomy) is possible;
2. why such a proposition is necessary;
3. the truth of such a proposition;
4. a proof of such a proposition.
The answers or solutions to 1 and 2 do not lie within the boundaries of a metaphysics of morals. Kant does not say so, but, presumably, the same would go for establishing 3 and providing 4. What has been shown by developing the generally accepted concept of morality includes the following:
A. the autonomy of the will attaches in an unavoidable way to the concept of morality;
B. the autonomy of the will lies at the basis of the concept of morality.
Kant then concludes, and so this now also counts as having been shown, that if we accept morality as something real, then we must accept the principle of autonomy as the highest principle of morality. So Kant is not claiming to have shown that morality is real but only that if it is real then the principle of autonomy must be its foundational principle. Kant has arrived at this claim, he says, only by an analytic method, by developing, examining, probing, analyzing, the concept of morality. To get further than this claim, that is, actually to establish the reality of morality, could be done by showing that:
I. the categorical imperative is true;
II. the autonomy of the will is true;
III. the principle of the autonomy of the will is absolutely necessary as an a priori principle.
Showing I, II, and III requires a possible synthetic use of pure practical reason. But before we can reliably use pure practical reason, we must first subject it to a critique. Kant intends the impending Third Section to be a movement toward such a critique.

Whatever it is that Kant does in this paragraph, whether merely inference or a complete argument, or explanation or definition, Kant also gives a simile. The simile appears to have these parallel structures:
1a. natural necessity is a property of the causality of non-rational beings;
1b. this property is the property to determine the being to activity through the influence of foreign causes;
2a. freedom is a property of the causality of rational beings with a will;
2b. this property is a property to determine the being to activity independently of foreign causes;
So, in the simile, natural necessity and freedom play the same determining role. This seems to be the main point that Kant wants to get across to the reader by means of the simile; that is why he emphasizes 'freedom', 'natural necessity', and 'determining' in the text. Kant also emphasizes 'will'. Does the will have any matching entity, an analogue, in the simile? To answer this, keep in mind that the will is a kind of causality, is practical reason, and is a capacity possessed by at least some rational beings.

In this paragraph, Kant gives an elaborate argument to establish the conclusion that free will and a will under moral laws are one and the same. There is much that can be discussed in this argument, but the gist of the argument is that freedom of the will depends on moral law being present in the will (as shown by the explanation of freedom in terms of causality and by the simile between natural necessity and freedom) so that there can be no free will that is free of the moral law. In other words, free will cannot be separated from moral law because free will cannot be separated from causality and causality cannot be separated from law of some kind, either natural or moral; the explanation of freedom in terms of causality (from the previous paragraph) tells us that free will can be separated from natural law; so it must be the moral law from which free will cannot be separated. Whether the argument is good or not probably depends most on whether the argument's transition from law to moral law via causality is fallacious; the explanation and simile from the previous paragraph are presumably intended to smooth over and make plausible this transition.

In this paragraph, Kant also speaks of a negative explanation of freedom and of a positive concept of freedom, and the latter is said to flow out of the former. He does not elaborate on what he means by characterizing the explanation as negative, the concept as positive, and the latter as flowing out of the former. He seems to say that the explanation is unfruitful because it is negative, and he seems to say that because the concept is positive it is richer in content and more fruitful. Perhaps Kant means that the explanation is negative in that it specifies that causality (of the kind in question) is not determined by foreign causes so that the freedom that is a property of this kind of causality is a freedom from foreign determining causes. So this explanation of freedom does not tell us how this freedom operates internally, only that it is not operated on by foreign determining causes; consequently, the explanation is unfruitful for discerning the essence of this freedom, for discerning how it does operate internally.

Though the explanation is negative in that it does not tell us how freedom operates internally, a positive concept of freedom still flows out of the negative explanation. How might this positive flow happen? I suggest that it flows out of the idea of law. The explanation, to be sure, does not tell us in any detail how this freedom operates, except to say that it operates in a lawful way. But by combining this idea of lawful operation with the explanation's requirement that the causality not be determined by foreign, external causes, we come across the thought that the law comes from within, that the law is self-imposed. With this, we have arrived at autonomy of the will. And this is richer and more fruitful because autonomy of the will leads us to the principle of autonomy, to the categorical imperative, and then to the specific duties of morality.

In the previous paragraph, Kant ended by concluding that a free will and a will under moral law are one and the same. In this paragraph, Kant starts by concluding that if freedom of the will is presupposed, then morality together with its principle follows from this presupposition by mere analysis of the concept of freedom. This latter conclusion is apparently supposed to follow from the former conclusion. Perhaps Kant was thinking along the following lines. The former conclusion basically asserts this equivalence: a will is free if and only if the will is under moral law. This equivalence was established (in the previous two paragraphs) merely by analyzing the concept of freedom. This analysis turned up causality, then law, then autonomy, and then the categorical imperative and morality. In short, by starting with the concept of freedom, we arrived necessarily at the concept of morality. So far this is just a linkage between concepts; it does not show that freedom is real or that morality is real. But if we were to presuppose freedom, if we were to assume that freedom is real, then, by the equivalence, we could get to the reality of morality. And this last in the train of thought is basically the conclusion with which Kant begins the current paragraph. Much of the rest of the Third Section will be devoted to supporting the claim that this presupposition of freedom of the will is necessary.

In the rest of the paragraph, Kant points out more precisely the importance of this presupposition of freedom. In the first part of the paragraph, he argued again (see p. 87 (4:440)) that the principle of morality, that is , the principle of autonomy, is a synthetic proposition. It is synthetic because analysis of the concept of an absolutely good will does not turn up a property of the will's maxim requiring that the maxim always be universalizable (i.e., always can be willed as a universal law). So there needs to be something to bridge the gap between the concept of an absolutely good will, or of a rational being, and the concept of choosing maxims that can always be turned into universal laws. There needs to be something that necessarily connects these concepts; otherwise, morality is lost since morality has to be universal and necessary, not contingent. Kant thinks the richer and more fruitful concept of freedom (from the previous paragraph), the positive concept, can supply this necessary connection between the concepts of a free will and a universalizable maxim. So if there is a justification for presupposing freedom of the will, then there is a way to establish the highest principle of morality that was found in the last Section to be the principle of autonomy. This establishment would complete the stated purpose of the Groundlaying: to find the highest principle of morality and then establish it.

In this paragraph, Kant argues ultimately that freedom must be presupposed as a property of the will of all rational beings. The general strategy in the paragraph and also the details of the arguments (see the analysis above) are not clear. There appear to be two separate arguments in the paragraph, and the general strategy seems to be to argue first that it is not enough to attribute freedom to ourselves if we do not have sufficient reason to attribute it as well to others. In other words, we need to find a rational way to attribute freedom not just to oneself but also to all other rational beings that have a will. The argument for this claim occurs in the first part of the paragraph, before the proclamation 'I say now' ('Ich sage nun' at 100.12 (4:448.4). This proclamation and the subsequent assertion beginning with 'Now I maintain' ('Nun behaupte ich' at 100.19 (4:448.9)) announce the way that Kant has found to attribute freedom to all rational beings. The way he has found is to recognize that a rational being's being able to act only under the idea of freedom is, for practical purposes, the same as the being being actually free. Kant's second argument, which occupies the remainder of the paragraph, is then given in support of this way he has found and just announced. So the overall strategy in the paragraph is this: Kant argues that we must (if we want to establish morality) find a rational way to attribute freedom of the will to all rational beings endowed with a will; he then announces what this way is; he then argues for this way.

In this first paragraph of this subsection, Kant reminds the reader of what has just transpired: we at last traced the determinate concept of morality back to the idea of freedom; we were not able to show freedom (not the idea of it) to be something real in humans and in human nature; we, if we want to conceive a being as rational and endowed with consciousness of its causality with regard to actions (i.e., endowed with a will), must presuppose freedom; in this way we find that we must attribute, for just the same reason, to every being with reason and a will this property to determine itself under the idea of its freedom to act. To complete the summary, Kant might have added at the end of the paragraph that acting under this idea of freedom is, for practical purposes, the same as actually being free (which he said at 100.15 (4:448.6)).

In this paragraph, Kant begins to wonder and reflect on the argument he has just given that established the principle of autonomy on the presupposition of freedom. In particular, he asks why he, a rational being, should subject himself, and in so doing all other rational beings, to this principle. Kant does not know straightaway what the answer is, but he does know that it is not that an interest impels him to so subject himself; he knows this cannot be the answer because such an impelling interest would not give a categorical imperative. But he does also know that interest is involved in some way: he necessarily takes an interest, though is not impelled by it. He explains that he must necessarily take an interest because his reason is encumbered by sensibility. For him, as such a sensuously affected being, the practical necessity of action is only an ought, not a willing, and the subjective necessity of action and the objective necessity of action can come apart. In other words, humans do not always do what they are supposed to do; so some kind of interest must be involved in order to point them, but not to drive or impel them (for they are free), in the right direction.

In this paragraph, Kant continues his reflection on the argument he has given, via the presupposition of freedom, for the principle of autonomy. He begins his reflection in this paragraph by supposing what appears to be the case, and what appears to be the case is that the argument has only presupposed the moral law in the idea of freedom and that we could not prove the reality and objective necessity of the law for itself. Even if this supposition matches reality, though, Kant does not downplay his achievement: even if what appears to be the case really is the case, the moral law has at least been determined more precisely than ever before. But he does allow that, on the supposition that the appearances are real, the validity and practical necessity of subjecting oneself to the moral law has not been supported by the argument. Kant will soon argue, however, that the appearances are not real; that is, the reality is that the argument does not presuppose (at least not in an objectionable way) the moral law and does prove the reality and objective necessity of the moral law. As he goes on to argue that the appearances do not match reality, we should note whether Kant manages to provide satisfactory answers to the three questions for which we cannot at present provide satisfactory answers and because of which we have gotten no further regarding the validity and practical necessity of subjecting ourselves to the principle of autonomy. Those three questions, again, are:
1. why must universalizable maxims be limiting conditions of our actions;
2. what is the basis of the worth we attribute to acting according to universalizable maxims;
3. how does this worth alone constitute or contribute to the personal worth that we believe that we feel that we have.

In this paragraph, Kant continues to develop his reflective supposition that the moral law has only been presupposed in the idea of freedom that was used to establish the principle of autonomy. He points out that we can take an interest in a personal characteristic, even if we will not benefit from it, as long as the benefit will be distributed in a rational manner. We can, for instance, take an interest in a person's virtuous character and subsequent virtuous conduct, even though we will not benefit directly from that conduct, and can judge the person to be worthy of the happiness distributed to her on account of her virtuous conduct. Kant notices, though, that this judgment of worthiness, of being deserving of happiness because of virtuous conduct, is due only to a prior acceptance on our part of the presupposed importance of the moral law: if we did not already have a commitment to morality, then we would not judge her worthy in this way. So the fact the we can take such interest in a personal characteristic (such as virtuousness) only reveals our prior commitment to the moral law; it does not show that our commitment is warranted or what the source of moral obligation is. Our taking an interest does show, though, what the source of moral obligation is not; it is not interest, for Kant's point is that interest in the personal characteristic results from, and is not the cause or origin of, our prior acceptance of the moral law.

From his reflections over the past few paragraphs, Kant arrives in this paragraph at the admission that there is a kind of circle in the argument establishing the principle of autonomy via the concept of freedom, and there seems to be no way out of the circle. Because he admits that there is a circle in the argument, he is in this paragraph not trying to argue for or prove the existence of the circle; rather, he explains or describes what the circle is and how it comes about. The circle is this: in the order of ends, we conceive ourselves as under moral law and so we assume that we are free in the order of efficient causes; and because we attribute freedom to ourselves we conceive ourselves as subject to moral laws. So the circular sequence is from moral law to freedom and then from freedom to moral law. Kant seems to think that the circle comes about because the concepts of the freedom of the will and of the lawgiving of the will are reciprocal concepts in that freedom and the will's own lawgiving are both autonomy. Kant says that reciprocal concepts are such that one cannot be used to explain or provide a ground for the other; at most, for logical purposes, one concept can bring different appearing representations of the same object to a single concept. In a parenthetical span, Kant mentions the reduction of fractions as an analogy. The analogy is evidently supposed to illustrate reciprocal concepts, something along the following lines. Consider the fraction 1/2, which is already fully reduced and represents one-half of a unit. Then consider 2/4 and 3/6; these are different appearing representations of the same value that is represented by 1/2. By the analogy, then, the concepts of freedom and lawgiving presumably correspond to 2/4 and 3/6 because the concepts represent autonomy just as the unreduced fractions represent the value one-half. Then, apparently, neither 2/4 nor 3/6 can explain or provide a ground for the other because neither is in reduced form. But since both can be reduced to 1/2, they, like the concepts of freedom and lawgiving, can bring their differing representations of the same object or value (i.e., one-half or autonomy) to a single concept, namely, the concept of a half of something or the concept of autonomy of the will.

In this paragraph, Kant suggests a possible way to escape from the circle, an escape which at first seemed impossible. He suggests, in particular, that there are two standpoints, one in which we conceive ourselves through freedom as a priori efficient causes, the other in which we represent ourselves according to our actions as effects. The suggestion, then, is that we take up one or the other of these standpoints when we conceive or represent ourselves in one or the other of these ways. Kant does not yet elaborate as to how, precisely, this suggestion gets us out of the circle. Perhaps he wants to say that the argument for the principle of autonomy based on the presupposition of freedom only occurs from within one of these standpoints, perhaps from within the first standpoint in which we conceive ourselves as free and efficient causes.

In this paragraph, Kant gives two abbreviated accounts of arguments that he gives elsewhere. The first such account starts with the two claims: first, that all representations that come to us without our choice give objects to cognize only as they affect us; second, and because of the fact described in the first claim, what the objects might be in themselves remains unknown to us. From these two claims Kant infers that ('mithin daß' at 105.24 (4:451.4)), as far as representations of that kind are concerned, we can only arrive at cognition of appearances, never of things in themselves. From this distinction between cognition of appearances and things in themselves, Kant then infers that('so folgt von selbst, daß' at 106.9 (4:451.12)) we must admit and assume that the things in themselves are not appearances and that they are behind the appearances. He also reasons that ('daß, da' at 106.13 (4:451.15)), because we can never become acquainted with the things in themselves except as they affect us, we can neither step closer to them nor ever know what they are in themselves. The conclusions (106.16-21 (4:451.17-21)) of the first account are then that there is a world of sense and a world of understanding, that the world of sense can be very diverse according to the difference of sensibility in various observers, and that the world of understanding always stays the same and is the ground of the world of sense.

The second abbreviated account ultimately concludes that the human being may not claim to know, according to the cognition that the human has of itself through inner sensation, what it is in itself. In support ('Denn da' at 106.24 (4:451.24)) of this conclusion, Kant says that the human does not create itself and receives its concept of itself empirically; this being the case, the human being informs itself about itself through the inner sense and consequently through the appearance of its nature and the way in which its consciousness is affected. From this information, the human being puts together a subject constituted from appearances and must assume something else, an I, that is the ground of the subject. Kant ends this second abbreviated account by concluding that the human being must count itself as belonging to both the world of sense and to the world of understanding (or to an intellectual world). It belongs to the world of sense in regard to mere perception and receptivity to sensations. It belongs to the world of understanding in regard to what is pure activity in it, that is, to what reaches consciousness not through sense but in an immediate way; but the human being has no further acquaintance with this intellectual world.

It is important to note that Kant gives these two abbreviated accounts here. The second, about the self, the I, seems to be a specific application of the first account so that the first account might initially seem sufficient for introducing the two standpoints. But this is not so; the second account is needed. The first account is indeed sufficient to introduce the two worlds, the world of sense and the world of understanding. These two worlds, however, are not the same as the two standpoints. They are related; the two standpoints are tethered to the two worlds, but they are not the same. The standpoints are points of view which we adopt when we think of ourselves in certain ways, when we think of ourselves as belonging to one or the other of the two worlds. So the standpoints make an essential reference to us, to how we conceive of ourselves. It is to explain this reference to ourselves that the second abbreviated account about the self is needed and why Kant included it.

In this paragraph, Kant brings it to our attention that the two-world theory that he has just outlined is a theory that even people of ordinary understanding have or would subscribe to if it were explained to them. Their only mistake is in trying to make the world of understanding into something that can be perceived and known. Kant believed that a critique of reason is needed in order to prevent such mistakes. This critique, for instance, will keep us from making the mistake of thinking that one world causes the other world, even though people of ordinary understanding may have a tendency to think that there is such a causal relation between the two worlds. What the exact nature of the relation is between the two worlds in not clear. Kant several times uses the metaphor of something being behind ('hinter' at 106.10, 107.17, 121.20 (4:451.12, 452.2, 459.28)) the appearances that we experience in the sensible world. The metaphor is apparently supposed to suggest that the intelligible world is the ground of, is responsible for, is the source and origin of, the world of sense. But the intelligible world is not the cause of the sensible world, at least not the same kind of cause as the kind of cause found in the natural necessity that determines events in the sensible world. As Kant later (e.g., 120.9-10 (4:458.36-7)) says, reason would overstep its bounds if it tried to explain the relation between the two worlds in terms of the causality involved in the natural necessity that governs the world of sensible objects that we experience. A critique of pure reason is therefore needed in order to keep reason from making claims it is not entitled to make and to keep us consequently from falling into error.

In this paragraph, Kant distinguishes our reason from our understanding, and both of those from our sense. Reason is the most elevated of the three, being pure self-activity and having a pure spontaneity. Reason also shows its highest status by being that which distinguishes the world of sense from the world of understanding and in so doing prescribing limits to the understanding itself. One such limit would be to limit the understanding in such a way that it does not make the mistake mentioned in the previous paragraph, the mistake of trying to make something invisible (e.g., a thing in itself) into something visible, into an appearance, into an object of intuition.

In this paragraph, Kant starts to make use of the distinction he has made between the two worlds and of the claims he made in the previous paragraph about reason as a faculty possessed by rational beings. From that distinction and those claims, Kant now maintains that a rational being must, as an intelligence, look at itself as belonging to the world of understanding. And then from this claim, he concludes that there are two standpoints. These standpoints are points of view from which the rational being considers itself. It is important to recognize that Kant has argued (or reported the argument) for the two-world theory before he concludes that there are two standpoints; so he seems to be building the two standpoints on top of the two-world theory, not the other way around. The adoption of one or the other standpoint has a bearing on the kind of laws under which the rational being acts. If the rational being is considered as belonging to the world of sense, then the laws are natural laws and so the will is heteronomous. But if the rational being is considered as belonging to the intelligible world, then the laws are independent of nature, not empirical, and are grounded in reason alone and so the will is autonomous.

In this paragraph, which is also part summary, Kant concludes that a human being, as a rational being, has to think of the causality of its own will as under the idea of freedom. That is, we have to think of ourselves as having free will, as freely acting agents. This conclusion is an important result because Kant uses it to shore up his argument (near the beginning of the Section) for the principle of autonomy established from the concept of freedom. As if to underline this importance, Kant provides a brief summary that shows how this latest piece of the puzzle fits into the larger picture of the overall argument that began at the beginning of the Third Section. The pieces now seem to fit together in the following way: rational beings, as rational, belong to the intelligible world rather than to the world of sense; human beings are also rational beings; so human beings, as rational beings, belong to the intelligible world; freedom is independence from the determinate causes of the world of sense (to which human beings, as sensuous beings, belong); so human beings, as rational and thus belonging to the intelligible world, must think of the causality of their will under the idea of freedom; but the concept of autonomy is inseparable from this idea of freedom; and the universal principle of morality is inseparable from that concept of autonomy. So, in sum, our rationality gets us into the world of understanding and, as members of this intelligible world, we are free from the kind of causality that governs the world of sense, and with this freedom (which is itself a different kind of causality, a causality of will rather than of things) comes autonomy and then morality.

In this paragraph, Kant lays out how the circle is escapable. The circle threatened to undermine the principle of autonomy and so morality. It threatened this because the circle alleged that autonomy is built on top of freedom but also that freedom is built on top of autonomy or moral law. So neither freedom nor autonomy has any independent support. What Kant has done by bringing in the two standpoints, one of ourselves as free and belonging to the world of understanding and the other of ourselves as obligated and belonging to both worlds, is to show how freedom can receive some independent support; by separating our view of ourselves into these two standpoints, we can see that, from the standpoint associated with the intelligible world, we must attribute freedom to ourselves as rational beings. Then, with freedom securely in place, autonomy can be built on top of it, followed at last by morality. The two standpoints themselves are supported by the distinction between the world of sense and the world of understanding and by the fact that humans are also rational beings belonging to an intelligible world governed by laws having their origin in reason rather than in nature. (The distinction between the two worlds is, in turn, supported by arguments in another work, the Critique of Pure Reason.)

In this paragraph, Kant argues that the laws of the world of understanding must be seen as (categorical) imperatives. In giving this argument, then, Kant essentially answers the question (of how a categorical imperative is possible) which is the title of this short subsection. There are many things that could be commented on in this argument, but two items stand out even on an initial reading.

First, Kant emphasizes that the world of understanding contains the ground of the world of sense and therefore of the laws of the world of sense. In the argument, these claims are presented as basic premises, but they are presumably supposed to be supported in some way by the two-world theory. This theory, if that is what it is, has always been a point of contention in Kant's ethics, indeed, in Kant's entire system of philosophy. Perhaps the major difficulty with it is that we are told that we can have no acquaintance with, and no knowledge of, the world of understanding, and yet we find Kant using premises such as these which seem to attribute to us specific knowledge of the world of understanding, in particular, knowledge of its relation to the world of sense.

The second item that stands out is that the argument exhibits an active self-referential perspective. Kant speaks of 'my actions' and 'my will'. The penultimate conclusions are that I will do something, will cognize myself as subject to the laws and to the autonomy of the will. And one of the final conclusions is that the laws of the world of understanding must be seen as imperatives for me. The use of this active self-referential perspective suggests that the argument's conclusion is to be taken not as an abstract theoretical result in speculative philosophy but rather as a practical result; it is to make an actual difference in how we live our lives, how we conduct ourselves, and how we make moral judgments. The moral imperatives and duties, to which we see ourselves as subject, obliged, or bound, are to have a real impact on our behavior.

In this paragraph, Kant explains how a categorical imperative, as a synthetic a priori proposition, is possible. His explanation makes use of the two-world theory, the two standpoints, and an analogy with synthetic a priori propositions in theoretical or speculative philosophy. The explanation goes something like this: reason provides us with the idea of freedom, and this idea makes it possible for us to think of ourselves as members of the world of understanding; when we think of ourselves in this way, we adopt the standpoint of freedom; if we were members of only this world of understanding, then all of our actions would always be in conformity with the autonomy of the will; but we at the same time intuit ourselves as members of the world of sense; when we think of ourselves in this way, we adopt the standpoint of obligation, and our actions ought to be in conformity with the autonomy of the will; this ought is categorical and represents a synthetic a priori proposition; the proposition is synthetic because in such a proposition the concept we have of ourselves as members of the world of sense, as affected by sensuous eager desires and inclinations, is supplemented by the concept we have of ourselves as members of the world of understanding (that is, it is synthetic because the concept of our intuited sensible self does not include or contain the concept of our intelligible self so that the latter concept has to be added to the former in the proposition); and the proposition is a priori because the concept of self that is added is of a pure practical will that is the highest condition of the sensuously affected will so that this added concept (of the pure will) can only come from reason, not from experience. This explanation parallels the explanation (in the Critique of Pure Reason) of how synthetic a priori propositions are possible in theoretical philosophy: there, too, concepts that do not originate in experience are added to intuitions of the sensible world and the result is a synthetic a priori proposition.

In this paragraph, Kant argues by an extended example that the practical use of common human reason confirms that the deduction is correct. Although Kant is not very clear about what exactly the deduction is or about what exactly the villain's wish is, the overall strategy seems to be to take an extreme example and show that the deduction holds up even in this extreme case. If it holds up in such an extreme case, then it will hold up in all other cases, which are typically much less unusual than the case of the most vile villain. So the practical use of common human reason confirms the correctness of the deduction.

This paragraph is explanatory as much as it is argumentative. This mixture is probably due to Kant's reliance here on the arguments given some years before in his Critique of Pure Reason. In any case, Kant's point seems to be that the subjective reality of freedom is secure: we do in fact think of ourselves as having free will. It is this fact that explains how we are able to make judgments about what actions we should take and about actions we ought to have taken but which we did not take. And Kant has argued that we must presuppose this freedom of the will if morality is to be taken as something real. But the objective reality of freedom is in question. What calls it into question is that we also (in addition to our subjective experience of freedom) experience the world as operating in a law-like manner according to natural necessity: pure water always freezes at 0 degrees Centigrade; unsupported objects always fall to the ground, etc. Kant even wants to say that we must experience the world in this law-like way; otherwise, experience itself would not even be possible. (And this is why the concept of natural necessity is not a concept of experience; natural necessity is not from experience because it is a precondition for experience in the first place.) So, because our physical actions take place in this natural world, it seems that they, too, would be governed by natural necessity, which would suggest that objective freedom is not real. So a kind of conflict or opposition is shaping up between freedom and natural necessity. We seem, subjectively, to be free, but the objective reality of natural necessity is strong evidence that we are not really, objectively, free. Moreover, the conflict cannot be resolved by appealing to experience; for, as Kant has argued in this paragraph, the associated concepts are not concepts of experience. The concepts have their origin in us, in our reason and our understanding, not in nature (and this reversal is the so-called Copernican Revolution from the Critique of Pure Reason).