Theory and Practice
Immanuel Kant's 1793 essay "Theory and Practice" is his attempt to defend his own moral and political theory against the charge that it is simply an idle academic exercise that cannot be brought to bear upon the real world in any useful way.1 He is concerned, in particular, to answer two charges -- the charges that his theory is (1) motivationally unrealistic, involving an account of moral motivation that is at odds both with scientific psychology and with all plausible philosophical accounts of rational deliberation and (2) not usable in either the design or critique of actual social institutions. Following a general discussion of what may be meant by the topic "theory and practice," Kant structures his essay as a response to the challenges to his own theories that are to be found in the writings of three other thinkers: Christian Garve, Thomas Hobbes, and Moses Mendelssohn.
Why is it worth caring about this essay? There are, I think, two reasons. First, it might be instructive to find out what the greatest philosophical mind of the eighteenth century had to say about the topic of the present volume. Second, a return to Kant is timely. There are few things more trendy these days than Kant-bashing, for he is often regarded as the patron saint of individualistic liberalism, Enlightenment rationalism, the idea of the "unsituated self" and a variety of other heresies that communitarians, virtue theorists, and feminists among others enjoy condemning. It might be instructive to see how Kant himself responded to the kind of Kant-bashing that was current in his own day -- some of it not all that different from our contemporary forms.
Introduction: What Is the Problem of Theory and Practice?
Introductions are generally Kant's strong point. Though nominally put forth to lay out clearly the topic for discussion and to set the reader up for what is to follow, they are often overly compressed and obscure. The introduction to "Theory and Practice" is, alas, somewhat in this mold. Although it makes (in a reasonably clear way) some important distinctions, it also contains much that is obscure and, as an introduction to what is actually to follow, somewhat misleading.
One thing is reasonably clear: Kant is at some level worried about the moral philistine -- the businessman, the politician, the military officer who prides himself on his role as a hard-headed, no-nonsense, realistic man of affairs (Geschäftsmann) and who, in pursuing his objective of greed or power or victory, either ridicules morality and moral theory as irrelevant to his practice or who conveniently adopts an account of morality exactly tailored to allow him to do whatever he pleases. (Such notions as "it's just business" or "it's just politics" or "military necessity" might function in this way.) This is the person who, when met with a challenge from the realm of moral principle, tends to respond dismissively with the smug cliché "yes, what you say may be true enough in theory, but it doesn't apply in practice."
What worries Kant the philosopher even more than these moral philistines, however, is the existence of philosophical doctrines that can be used to give a cover of intellectual respectability to the iniquities and deceptions practiced by such persons. There is very little that a philosopher can do directly to combat ordinary human venality and self-deception, but the philosopher can properly assume the task of unmasking the intellectual pretensions of those who would use or misuse philosophical doctrines in support of venality and self-deception. Such is Kant's objective in "Theory and Practice" -- an objective he pursues by attempting to demonstrate how bad theories (or misuses of good theories such as his own) can aid in the corruption of human life and human society.
But what is a "theory" and what is the "practice" to which it is related? Kant writes,
Kant is here making the very clever suggestion that, at least in the domain of morality, the very distinction between theory and practice -- and thus the idea that there could be an important gap between them -- is incoherent.2 An activity or institution is properly called a "practice" (Praxis), claims Kant, only if it is viewed as the instantiation of some general principles (i.e., some theory); and a moral theory is adequate only to the degree that it provides a rational reconstruction -- in terms of general principles -- of those practical judgments that constitute our ordinary moral consciousness. As Dieter Henrich puts the point, "[Kant] speaks of the theory as being inherent in moral consciousness and action itself. As such it eo ipso is effective in a practical way."3 Thus a bad theory, to use Rawlsian language for the method that Kant is here adopting, is a theory that fails to put us in reflective equilibrium with respect to our pre-theoretical moral convictions. Such a theory will indeed deviate from practice, but this is because it is a bad theory (bad because of this very deviation) and not because it is somehow inherent in the idea of moral theory itself that it should be practically useless. This account of the necessary connection between moral theory and practical belief recalls, of course, the first two sections of the 1785 Grundlegung where Kant speaks of his method as involving the "transition from the common rational knowledge of morals to the philosophical" and of the "transition from the popular moral philosophy to the metaphysics of morals."4
If we do conceptualize moral theory as the best rational reconstruction of our shared moral beliefs, then one worry about a theory/practice gap can indeed be met: the possible gap between moral theory and ordinary moral consciousness. (This gap is one that, I would argue, faces the utilitarian much more dramatically than it does the Kantian; and demonstrating the gap is surely the main point of all the well-known scapegoat and other counterexamples to utilitarianism.5).
As clever as Kant's suggestion is, however, it surely does not address all aspects of the theory/practice challenge to morality. For there are at least two different senses in which a moral theory can be thought to fail in practice. The first sense, one already discussed, involves the charge that a particular theory (e.g., utilitarianism) does not account for our considered moral judgments and thus cannot be the best rational reconstruction of our moral consciousness.
A second sense of the charge, however, involves a possible gap between our moral consciousness itself and the real world -- the world of empirical reality. Consider, as an illustration, the retributive theory of punishment. This theory is, it could be argued, the best rational reconstruction of our considered moral judgments about punishment -- e.g., the common moral belief that the guilty deserve to suffer. However, it is possible that this judgment itself presupposes a variety of false or self-deceptive views -- e.g., about the nature of crime and criminals, about the legal process, about what actually happens to people when they are punished, and about the nature of the societies in which people comfortably make the judgment that the guilty deserve to suffer.6
This is a different kind of theory/practice gap than the one already discussed, but one that should not be hastily dismissed. The idea is that a person who lives too much in the world of theory may negligently think that the world in which he actually lives admits of clear application of theory when in fact it does not. Such a person may even come to a distorted view of the world by seeing the world only through the spectacles of his theory -- thinking his theory is consistent with the facts because he does not realize that he is unable to accept as a fact anything that is inconsistent with his theory. (Paranoids, seeing all helpful gestures as threats, are masters of this; but the tendency is also present in those who are mentally normal. Think of those who see all welfare recipients as chiselers, all poor people as lazy, all criminals as free and responsible, and -- to shift ideologies -- all women as really desiring the independent and autonomous status that (supposedly) comes from having a career.7)
Kant is not indifferent to such problems in "Theory and Practice" and suggests that the existence of such people shows, not a weakness in theory, but a weakness in human nature -- the problem that some people simply lack the "natural gift" of judgment:
Kant has part of the story here, but surely there is more that needs to be said. It is true that some people simply lack judgment. It is also true, however, that many people make faulty judgments, not simply because they lack some "natural gift,j" but because they are caught up in complex webs of false consciousness and self-deception -- webs perhaps built and encouraged by certain philosophical theories. Kant clearly saw this in 1793 because the insight plays a profound role in his Religion, published in that year.8 But in "Theory and Practice," published in the same year, the insight is missing.9
There is still one additional passage in Kant's introduction that requires comment. Kant writes:
On one interpretation, Kant is clearly correct in what he says here. If I am really morally required to do X, then the fact that it would now be very difficult for me to do X or the fact that things happened to come out well in the past when I did not do X is irrelevant in determining my duty. Such factors could at most be relevant in excusing me for the nonperformance of my duty.
On another interpretation, however, Kant may well be mistaken. It is not unreasonable to suggest, for example, that the proper specification of the moral duty itself may sometimes quite properly take account of empirical variations in circumstances. Consider, for example, the duties that we have -- in both morality and law -- to do what is reasonable in certain circumstances, where reasonableness is partially understood in terms of what might happen to us (the risks we run) in acting in a certain way.10 Consider also our concept of a "white lie" -- a lie about a minor matter that will do no harm and perhaps great good. Philippa Foot writes instructively on such matters:
Except for the odd -- and to me quite unjustified -- cynicism of the final sentence, Foot seems to be onto something very important here -- something that Kant, with his well-earned reputation for rigidity, often, if not always, misses.12 If moral theory is indeed a rational reconstruction of our ordinary moral consciousness, and if it is a part of that consciousness that (for example) social lies are sometimes permissible,13 then Kant can maintain that moral rules never admit of exceptions only if he is willing to accommodate these exceptions into the specification of the rules themselves. And the first formulation of the categorical imperative may often allow this, since the maxim "to pay a false compliment when so doing will build a person's confidence" (for example) is probably universalizable. There are, of course, no exceptions allowed to the categorical imperative itself (morality is indeed rigid in that sense), but it does not follow from this that the categorical imperative (first formulation), when screening specific maxims for their moral permissibility, would never allow us to make exceptions to the rigid rule "never lie." Indeed, the universalization form of the categorical imperative probably would not identify the rule "never lie" as the correct moral rule about lying.14
Foot has insight on these matters, and Kant is -- in my view -- simply mistaken if he does not to some degree incorporate it into his theory. The issue is complicated, however, and the cynicism of Foot's final sentence may suggest that she has a (misplaced?) Kantian scruple or two of her own here.
So much (at last) for Kant's essay, where Kant will seek to establish that his own theory, properly understood, does pass the test that he himself imposes in the essay -- namely, that the theory be motivationally realistic and applicable in the design and critique of actual human institutions:
Christian Garve was one of Kant's contemporaries -- a writer of popular philosophy (a philosophe) whose moral seriousness was respected by Kant. Even though Garve published some criticisms of Kant that Kant regards as in part misinterpretations, Kant treats these criticisms with general courtesy and does not deploy the full polemical force of which he is capable. This courtesy and respect may also be based on Kant's realization that not all of Garve's criticisms are simple misunderstandings and that some of them, indeed, are potentially serious and require a careful response.16
Kant's own moral theory stresses that duty is to be determined after abstracting from inclination and must (in cases of true moral worth) motivate in independence from inclination; he is famous for his claim that an action adds to the moral credit of the agent only if it is motivated by respect (Achtung) for duty.17 Garve (greatly under the influence of British moral philosophy) is the first in a long line of anti-Kantians who argue that this theory is fatally defective because it cannot give proper weight to the role of the emotions in morality. Although such later critics as Schopenhauer (in the 1841 On the Basis of Morality) will chide Kant for ignoring or downplaying sympathy and other examples of what Barbara Herman calls "motives of connection,"18 Garve is particularly interested in the emotions of self-regard -- particularly the desire for one's own happiness. The selfish desire for one's own happiness is, according to Garve, (1) recognized in common experience as the actual motive for all human action, (2) the only motive that yields a coherent account of rational deliberation, and (3) the only motive that can be "reconciled with the customary principles of psychological explanation (all of which are based on the mechanism of natural necessity)" (285, 68).19 Garve makes points (1) and (2) in the following passages quoted by Kant:
Garve is arguing that practical deliberation (including moral deliberation) is instrumental in nature -- actions being approved as rational to the degree that they represent efficient steps toward an outcome identified by the agent as "the good." But what does it mean, asks Garve, for an agent to identify some outcome or goal as good (das Gute)? Simply this: it is an object of preference (Vorzug) (or want, or desire). And what is an object of preference? Simply this: an object expected to produce happiness when attained. Thus (oversimplifying a bit) Garve seems to be offering the following equation: good = preference satisfaction = happiness. According to Garve, to say that I regard something as a good is imply to say that I prefer it, and the only intelligible ground for preferring something is the belief that it will ultimately play a role in my overall happiness. Thus respect for duty -- Kant's sole candidate for a motive that is truly moral -- either does not exist or may be reductively analyzed in terms of the desire for personal happiness. In short, I do my duty either because it pleases me to do so or because I fear that I will feel bad if I do not.20
Kant's response to Garve is complex and overlaps, at certain points, with Bishop Butler's well-known response to Hobbes's doctrine of psychological egoism.21 Kant's response, as I interpret it, can be distilled into three basic arguments: (a) The phenomenological evidence provided by honest introspection reveals that in fact we all recognize a sharp distinction between the motive of respect for duty and all motives concerned with our own welfare or happiness. (2) We do take satisfaction when we do our duty and do indeed fear bad feelings when we do not. Such feelings are not the basis of respect for duty, however, but indeed can themselves be understood only through the realization that we value duty for its own sake. (3) The apparent plausibility of Garve's argument rests upon a variety of conceptual confusions and ambiguities -- e.g., with respect to the meanings of such terms as "preference" and "the good." Personal inclination does play a role in determining duty, but not in the way that Garve thinks; the role it does play is not only consistent with Kant's theory but is required to make sense of that theory.
Let me now explore each of these three arguments in more detail.
(1) The phenomenology of moral experience. According to Kant, the nature of moral motivation often reveals deep psychological conflicts between what we see as our duty and what we want or desire (in any ordinary sense of "want" or "desire'). Think, for example, of duties owed to strangers (people for whom we have no feelings) or to enemies (people toward whom we have negative feelings and who may even be in a position to harm us). We clearly can have such duties and sometimes at least find ourselves acting because of them in the full consciousness of the fact that, given our feelings, we would prefer to be acting otherwise. The specially conflicted nature of moral motivation (the aspect of it that makes us see duty as a categorical imperative22) simply does not emerge on Garve's account -- a flaw that prevents it from being an accurate picture of our shared moral phenomenology. Kant writes,
We all, of course, experience frequent conflicts between our various desires (should I continue the pleasures of smoking or aim instead for a long if boring life?), but the conflict between duty and any desire seems different in kind -- different because it at least appears as a conflict with my empirical self (as a bundle of given wants) and not within my empirical self (one want against another). The difference in kind is marked, at least in part, in the special pride I feel -- the special respect I feel for myself -- when I choose duty over inclination.23
(2) The pleasures of being moral. This last argument by Kant could simply be taken as evidence that the ultimate motivation for moral behavior is indeed selfish. It surely feels good to take pride in one's actions. Thus perhaps the real motive for moral behavior is the desire to get the feeling of pride and sublimity that comes from doing one's duty. (At the very least one may be motivated to avoid the painful guilt feelings that come from not doing one's duty.) Does not moral motivation then boil down to the search for personal happiness after all?
Kant's answer here is clever. He does not deny that we often obtain considerable happiness through acting morally -- particularly the pride or self-satisfaction attendant to such behavior. But what is it about such behavior, asks Kant, that gives us the special kind of pleasure involved? Surely it is pleasure -- prideful pleasure -- that we take in seeing that we are creatures capable of appreciating moral duty and being motivated by it! Otherwise what would we have to be proud of? Seeing ourselves as motivated by a desire to get moral pride would necessarily preclude our ever getting moral pride, and thus -- if moral pride truly exists -- we must actually be motivated by duty (or we must at least deceive ourselves into thinking that we are so motivated). The good feelings generate neither duty nor respect for duty; they are rather generated by duty itself:
Thus, at least in the realm of moral phenomenology, Kant seems to win the battle with Garve. He is far closer to being right in his description of the internal life of the normal moral agent.24
Praising Kant for the accuracy of his phenomenological account of morality should not, of course, be confused with an endorsement of his rich (bloated?) metaphysical account of such matters.25 Although Kant has an elaborate and controversial account of the contra-casual ("noumenal") freedom that he thinks must be present in the world in order for moral motivation to be distinguishable from typical desire-based motivation, we do not need to accept all this to grant Kant's phenomenological point. The phenomenological point is simply this: normal human beings, though often motivated by means/ends rationality, are sometimes motivated in a different way. They understand moral reasons, conduct moral conversations and listen to moral arguments, and are then moved to act on the basis of these reasons and arguments. This is at least part of what Kant means by "moral autonomy," and this part surely requires no controversial or unscientific assumptions about non-natural causation. (Assume, if you want, that responsiveness to moral reasons is encoded in human nature through some complex interplay between biology and conditioning.) Such a minimalist account of moral autonomy is, for example, totally compatible with a Freudian theory of the personality -- where the "superego" plays the role of respect for duty. Freud's theory of moral motivation overlaps Kant's to a substantial degree (it was perhaps influenced by it), and anyone who thinks that Freud has insight on moral phenomenology must agree that Kant has insight here as well.26
One additional point, obvious but sometimes missed, is worth stressing. It is easy to misstate the nature of the tension between duty and inclination because the phrase "acting out of respect for duty" is in some cases a misleading way for Kant to put his motivational point. Kant is not suggesting that, against all inclination, one is to go out looking for duties so that one can act with moral worth. This is not how duty arises in the life of the normal person. In the typical case, one wants to do something (e.g., spend a large amount of money on a new wardrobe); and one wants to do it in the perfectly ordinary sense in which "wants" means something like "desires" or "will take pleasure in." (What one wants to do in this subjective sense is, when properly specified, what Kant calls a "maxim.") Before acting on such a subjective maxim, however, the moral person who respects duty will use the categorical imperative to test the maxim for its moral permissibility (i.e., to make sure that the maxim is not contrary to duty). If the maxim passes the test, the person is then free to act on it. If the person now acts, is it true to say that he acted from duty? In one sense, the answer is surely no; for the best explanation for his new wardrobe is "he wanted it" and a crazy and unbelievable explanation would be "having it was required by duty" or (even stranger) "he bought the wardrobe because buying it was morally permissible." Yet surely the action is properly motivated and surely satisfied the Kantian sense of acting out of respect for duty. The moral: respecting duty should not be interpreted as the shunning of all personal pleasures and desires for happiness in order to respond to the direct call of duty. It should rather be interpreted as the commitment to pass all such desires through a certain kind of screening or filtering device -- the device provided by the categorical imperative and its demand for universalization. When one uses this screening device and accepts its verdict as final, one has -- in the important sense -- acted out of respect for duty. For the Kantian, the inner voice of morality functions more as a censor than as a drill sergeant.27
(3) The seductive charms of the selfishness theory. If the kind of egoism-hedonism taught by Garve fails to capture our ordinary moral consciousness, then why are Garve and others still inclined to reject the Kantian account? The answer, surely, is their belief that the Kantian theory has such deep flaws that it must be rejected even if this requires the rejection of our ordinary moral consciousness as erroneous. What kind of support might they have for such a negative assessment? Such support could be either conceptual (Kant's theory is incoherent) or scientific (Kant's theory is incompatible with the naturalism of scientific psychology). I have already indicated why I think that the latter argument fails, so let me briefly consider the former.
On this issue, we have the familiar Hobbes-Butler dog and pony show with Garve and Kant as stand-ins. Garve argues that it makes no sense to think that a person could every act unless he prefers to act in that way and that this truth makes any theory other than psychological egoism/hedonism incoherent. Kant responds that this argument plays on two different meanings of "prefer." "I prefer X" may simply mean "I have some reason for choosing X." Kant can happily concede that all actions are in this sense based on preference, for this is just another way of saying that all actions are motivated; the claim that all actions are motivated may be tautological. ("Jones had a preference for his duty" is in this sense harmless and unobjectionable.) A quite different meaning of "I prefer X," however, is "I think it will give me pleasure to do X." Kant cannot, of course, regard "It will give Jones pleasure to do his duty" as a harmless and unobjectionable analysis of moral motivation. He does not really have to confront this possibility, however, since the claim is not self-evident (not even intrinsically plausible) and Garve gives no good argument for it. Although it is true that duty is a matter of preference where preference is simply equivalent to motivation, this does not entail (is not even an argument for) the claim that duty is a matter of preference where preference is equivalent to pleasure.
After arguing that Garve also trades on ambiguity in the phrase "the good" (confusing instrumental good with final good, conditional good with unconditional good, and physical good with moral good), Kant leaves this final assessment of Garve's attack on his theory of moral motivation:
Kant claims that the section on Garve was addressed to the topic of theory and practice as it impacts on the moral individual -- "in relation to the well-being of each man" (277, 63). This long second section of the essay is directed to the person as a political being -- as a being, not simply worrying about the possibility of moral motivation, but seeking to enjoy the benefits of a basic social structure that guarantees fundamental rights at a constitutional level. If Kant's theory cannot help in this political task, it will have failed in one important way to apply in practice. Thus Kant needs to show that it can help, and this is the primary task of the section.
Kant sees this task as having two main aspects -- aspects that do not sit comfortably together. He wishes (1) to demonstrate that his theory can help in the design of just (i.e., rights-respecting) civic institutions and (2) to demonstrate that his theory shows why citizens have no right to revolt against legal authority even when that authority violates their rights and ignores or perverts the justice of their institutions.
Although I think that most of Kant's arguments against revolution are confused (e.g., involving simplistic positivistic doctrines of sovereignty), I plan to ignore this issue in the present essay and concentrate solely on his account of justice -- drawing on the material on revolution only when it is useful to the other project. I do this for the following reason: since even before the publication of "Theory and Practice," the material on revolution has received essentially all of the attention. (It is indeed the only aspect of this essay that has not been generally neglected.) Kant's contemporaries eagerly awaited the work because they wanted to see if Kant, the world's leading philosophical liberal, would support the French revolution. Later commentators, generally uncomfortable with Kant's apparent refusal in the essay to support revolution under any circumstances, have generally probed the consistency of this refusal with his general moral and political doctrines. Anyone now writing on this topic would need to survey this enormous body of literature -- a task not possible in the present context.28
Although the second section of the essay is said by Kant to be "against Hobbes," Hobbes's thought receives no detailed commentary. The negative reference to Hobbes essentially serves one main purpose: Kant wants to put his reader on notice that, in spite of his agreement with Hobbes that citizens have no coercive rights against political authorities, he does not agree with Hobbes that citizens have no important rights of any kind against the state. (Recall the United States Constitution here. It grants citizens many fundamental rights against the state, but among these is not the right of revolution.) Thus, the primary point of this section of Kant's essay is to identify the basic rights of citizens and to place those rights within a general scheme of justification -- i.e., within a theory of justice.
Kant's primary purpose in this section is to defend the principles of a liberal social and political order -- i.e., a basic social structure that will be just in the sense that it will protect the fundamental natural rights of all citizens -- particularly the inalienable right of freedom.29 If such a liberal theory is to survive the challenge that it works in theory but not in practice, it must provide a test or criterion for such a social order -- i.e., some device that will allow us to recognize when a social order does respect rights in the proper way. For Kant, a model of universal agreement (adapted from social contract theory) is such a device. He writes,
Kant is, I think, outlining some very important ideas here -- setting the groundwork for the constructivist methodology of Rawls and the related conversational methodologies of Habermas and others.
If a just society is essentially (as Kant says) a society that treats the natural right of freedom as sacred, this freedom cannot be interpreted simply to mean the freedom to do what you want or the freedom to be bound only by rules to which you actually consent. This account of freedom would allow no punishment of criminals (who wants to be punished?), no taxation (who wants to be taxed?), and no duties imposed on those citizens of a democracy who did not vote with the majority (where is their actual consent?) Such an account would, in short, allow no coercion; and even just states, as states, are necessarily coercive. Coercion (forcing people to do what they do not want to do) is thus a political necessity. However, certain forms of coercion are clearly inconsistent with respect for human freedom. For Kant, a clear example is paternalism:
For the state to compel as a parent (treating subjects like "immature children," 290-91, 73) is necessarily to fail to respect the freedom of citizens. For the state to thwart the actual desires of citizens and to compel them even when they do not actually consent, however, is not necessarily to fail in such duties of respect. The difference? In the former case, it is impossible to imagine all citizens capable of moral reasoning consenting, after extended argument and conversation, to a strong principle of paternalism.30 (It is also impossible to imagine all such citizens consenting to permanent humiliation and powerlessness based on hereditary class or racial status.) It is, however, not difficult to imagine them all consenting to some legal principles that require punishment, some principles of taxation, and the general principle of majority rule. This, then, is Kant's test for justice: a principle of coercion is just if it is possible that every citizen could consent to it in an environment of mutually respectful moral conversation.
Is this a good test? When it seems to work well, it is tacitly operating on one very important assumption -- an assumption that perhaps keeps the test from being as foundational as Kant thinks. The assumption is that the imagined grounds for consent or dissent be reasonable. The powerful idea, I take it, is this: To respect all persons as moral equals, we must consider (anticipate if necessary) the kind of objections that any thoughtful and morally sensitive person might make against our proposed coercion and regard the coercion as unjust if we see such objections as possible grounds for a reasonable refusal of consent. Mere imagined refusal or assent -- regardless of its reasonableness -- would surely not do, however. I can, for example, imagine persons failing to see the importance of free political speech for the preservation of just institutions and thus I can imagine them agreeing to coercive curtailment of political speech. I would not want to conclude from this, however, that such coercion is just. Like Kant, who defends the "freedom of the pen" as the most basic political right (304, 82), I would want to maintain that free political speech is a requirement of justice even in the face of my thought experiment (wherein I can imagine all citizens agreeing to its unreasonable curtailment). Thus surely the operative test must really be this: a coercive rule is just if it is possible that all citizens could agree to it on grounds that are informed and reasonable.
If reasonableness is accepted as a constraint on possible consent, however, then consent itself perhaps cannot play quite the foundational role that Kant envisions. It will at least share the foundational role with an account of reasonableness. Although Kant speaks generally of the natural right of freedom, he clearly does not regard all freedoms as equally fundamental. (This is revealed in his own insistence on the priority of freedom of the pen.) It is impossible, however, to sort freedoms into categories of differential importance in terms of freedom itself. Some other value is required. My own view is that the best candidate is a concept of the human good -- a concept allowing us to rank as most important those freedoms that are intrinsically bound up with a reasonable account of the human good. Political freedom is more important than (for example) the freedom to collect string, because the former is closely tied to the human good in a way that the latter is not. And what about sexual freedom -- e.g., for homosexuals? We surely cannot even begin to make a case for such freedom as fundamental without developing an account of love, sexuality, intimacy, and the role they play in a full and good human life. If we see sexual freedom purely as freedom of recreational pleasure, we will trivialize it and make it a very poor candidate indeed for a fundamental liberty. Could we imagine all citizens reasonably consenting to repressive rules with respect to homosexuality? If sex is just recreation, I think the answer is yes. If sex is more deeply tied to the good life, however, I think the answer is no. In short, if liberal societies must rank liberties (some more fundamental than others), then strong forms of the neutrality principle advocated by Ronald Dworkin and others must be rejected.31
Kant seems officially committed to a neutrality principle (290, 72), in part because he finds it difficult to articulate a complex account of the human good. Either he sees it simply in terms of our moral powers (much too narrow -- to moralistic -- an account) or he thinks that it cannot be distinguished merely from the desire to be happy -- a value on which he pours (as usual) a large dose of contempt:
Even if one agrees with Kant that the concept of external right should not depend upon happiness (as mere pleasure), one does not have to conclude from this that the concept of external right should not depend on any account of the human good. (Indeed, in my view, no plausible account of the human good will attempt to interpret it merely as pleasure.) At one point in "Theory and Practice" Kant seems to recognize this. The issue is religious freedom and Kant's desire to defend it as a fundamental right:
Here Kant seems to impose a constraint on his consent test -- a constraint involving the value of "humanity's vocation and end." It is unclear just what this means, but it strikes me that it must in part be some notion of the human good that is not a mere matter of personal preference. What seems crucial to Kant here is not a general failure of possible consent to such a law, but rather a failure of consent on the part of persons who understand the nature and value of religion in human life. There are, then perhaps some beginnings here of the kind of account that Kant needs and that he develops in more detail in other writings.33 In "Theory and Practice," however, it is nothing more than a mere hint.
This final section of Kant's essay is not, in my view, really about what Kant says it is about: right and justice in international law. It is directed against Moses Mendelssohn, but not against anything Mendelssohn said about international law or "the cosmopolitan point of view." Most commentators take Kant's own gloss to heart, however, and note this final section of "Theory and Practice: merely as a brief anticipation of the doctrines that will be developed in detail in the 1795 Perpetual Peace.
What, then, is the real point of this section? It is, in my view, a counsel against moral despair. Note the question with which Kant begins the section -- a question having little to do with international justice but a great deal to do with the temptations of moral pessimism:
Mendelssohn had defended the pessimistic view,34 and this is the sense in which Kant's final section -- an attempt to meet the pessimism -- is truly contra Mendelssohn. There is noting more basic to Kant's moral and political outlook than his view of human beings as sacred or precious -- beings having the unique value of dignity (Würde) that is the foundation of their basic rights. If human beings do not deserve such respect, however, then Kant's theory -- which assumes that they do -- will fail in practice. The view that human beings are sacred is, of course, basic to the Christian worldview in which Kant was raised, but Kant the philosopher will not allow himself, through mere faith, to avail himself of its theological defense, that people are sacred because they are created in God's image. If people are genuinely sacred, it must be because of something about them that can be understood in a secular, empirical way.
But is this possible? Some philosophers, such as Robert Nozick, have argued that the demonstrated human capacity for unspeakable evil (e.g., the Holocaust) shows that human beings are dramatically unworthy of the kind of deference paid to them by Kant.35 Though Mendelssohn's view was not as bleak as Nozick's, neither is it terribly upbeat: "The human race as a whole swings slowly back and forth, and it never takes a few steps forward without soon afterwards relapsing twice as fast into its previous state" (307, 85).36
It is in responding to Mendelssohn's pessimism that Kant makes reference to emerging international law -- only as an example of a general point he wants to make about humanity as a progressive species: "Human nature never seems less lovable than in the relations among entire peoples" (312, 89). Yet Kant sees signs that moral tendencies in (at least some) people are gradually moving the world toward an international legal order, an order that will end the wars that have brought out the worst in humanity. If human history does indeed continue to move in this way, it will be easier (according to Kant) to see humanity as a moral and progressive species that is perhaps deserving of some special respect. But who really knows? Kant, rejecting (in philosophy) all appeals to religious faith, can do no better than close "Theory and Practice" with a statement of secular faith:
Kant had not, of course, seen evil on the scale that we have known in the twentieth century. Suppose he had known of the Holocaust. Would he have joined Nozick in seeing it as evidence of total human worthlessness? I think not; I think his moral faith would have survived even this. After all, unless human beings are precious, what is so bad about murdering six million of them? Properly to condemn the Holocaust -- as a crime against humanity -- requires a view of humanity as having a value at least very like Kant's value of dignity.38 Thus Nozick's moral pessimism is probably premature -- probably even inconsistent with his own activities as a concerned philosopher. Kant says this of Mendelssohn:
Thus here we find Kant, the great secular rationalist, adopting as an article of faith the view of humanity and its possibilities necessary to avoid moral despair and to leave the door open for whatever good may be realizable. (We must think and act as if what is true in theory is also possible in practice, since life is bearable and meaningful on no other assumption.) Here also we find Kant, one of the supposed founders of the philosophy of "possessive individualism," adopting a communitarianism of the most ambitious sort: hopeful and sympathetic fellowship with the entire human community in its present and future generations and admiration for those who (like Mendelssohn with respect to Judaism) seek to preserve the special values of their own subcommunities within such a ongoing scheme of human progress. Perhaps rationalistic libertarians are not Kant's only natural bedfellows after all.
1. The full title of Kant's essay is "Über den Gemeinspruch: Das mag in der Theorie richtig sein, taugt aber nicht für die Praxis" ("On the Proverb: That May be True in Theory but Is of No Practical Use"). The standard citation source for this essay is Bank VIII (volume 8) of Kant's gesammelte Schriften, hrsg. Konigliche Preussische Akademie der Wissenschaften (Berlin and Leipzig: Walter de Gruyter, 1904--). In the present essay I am mainly relying on the translation by my colleague Ted Humphrey in his Perpetual Peace and Other Essays (Indianapolis: Hackett, 1983), 61-92. All page references to Kant's essay will be given in the following form in the body of the text: the page number from the Academy edition will be given first, followed by the page number from the Humphrey translation -- e.g., (275. 61). [BACK]
2. Kant says that morality is quite different from speculative metaphysics in this regard. The claims of speculative metaphysics (e.g., claims to detailed knowledge about the nature and will of God) tend to involve "mere empty ideas that have either no practical use whatsoever or even one that would be disadvantageous. In such cases, therefore, the proverbial saying could be perfectly correct" (276,62). [BACK]
3. Dieter Henrich, ed. Kant, Gentz, Rehbert: Über Theorie und Praxis, Einleitung von Dieter Henrich (Frankfurt am Main: Suhrkamp Verlag, 1967), 10. ("Er spricht von der Theorie, welche dem sittlichen Bewusstsein und Handeln selbst innewohnt. Als solche ist sie eo ipso in praktischer Wirkung.") This very useful volume contains Kant's essay, the passages from Christian Garve to which Kant responds, and essays written in response to Kant by Friedrich Gentz and August Wilhelm Rehberg. Henrich's introduction is very useful. [BACK]
4. Grundlegung zur Metaphysik der Sitten (Foundations of the Metaphysics of Morals, trans. Lewis White Beck, Indianapolis: Bobbs-Merrill Library of Liberal Arts, 1959, First Section and Second Section). [BACK]
5. Kant is perhaps more confident than he should be that it will be easy to identify shared moral beliefs -- a confidence perhaps generated by his generally addressing too narrow a sample even of his own culture. But unless we propose to treat other persons with contempt and not attempt to engage them in moral conversation at all, what alternative do we have except to search for some points of shared agreement? Idealized models of conversation and agreement (in Kant, Rawls, and Habermas) can perhaps be of assistance here. [BACK]
6. I expand on this idea in my essay "Marxism and Retribution," Philosophy and Public Affairs 2:3 (Spring 1973), reprinted in my Retribution, Justice and Therapy: Essays in the Philosophy of Law (Dordrecht: Reidel, 1979). See also Life Sentences, Rage and Survival Behind Bars, ed. Wilbert Rideau and Ron Wikberg (New York: Times Books, 1992). I wonder how many people, after reading this book and seeing what life in an American prison is really like, will continue to feel comfortable saying "they deserve it"? There is, of course, this very important point still to be said in defense of the retributive theory: only by regarding deserved suffering as the norm for legitimate punishment can we see the terrible injustice of what we are actually doing. For more on this point see my "Three Mistakes About Retributivism," Analysis (April 1971, reprinted in Retribution, Justice and Therapy. [BACK]
7. Stephen Carter has noted that the way in which minority academics are perceived is to a large degree a function of the current debates over the justice and wisdom of affirmative action in university hiring. See his Reflections of an Affirmative Action Baby (New York: Basic Books, 1991). [BACK]
8. "[People] may ... picture themselves a meritorious, feeling themselves guilty of no such offenses as they see others burdened with; nor do they ever inquire whether good luck should not have the credit, or whether by reason of the case of mind which they could discover, if they only would, in their own inmost nature, they would not have practiced similar vices, had not inability, temperament, training, and circumstances of time and place which serve to tempt one (matters which are not imputable), kept them out of the way of those vices. This dishonesty, by which we humbug ourselves and which thwarts the establishing of a true moral disposition in us, extends itself outwardly also to falsehood and deception of others. If this is not to be termed wickedness, it at least deserves the name of worthlessness, and is an element in the radical evil of human nature, which (inasmuch as it puts out of tune the moral capacity to judge what a man is to be taken for, and renders wholly uncertain both internal and external attribution of responsibility) constitutes the foul taint in our race," Religion innerhalb Grenzen der blossen Vernunft (Religion Within the Limits of Reason Alone,) trans. T. Greene and H. Hudson (New York: Harper and Row, 196 0), 33-34. [BACK]
9. Except, perhaps, in passing. In discussing the defects in Christian Garve's account of moral consciousness (284-85, 68), Kant speculates that Garve knows in his heart that Kant's account is correct but is mislead by his head -- i.e., by some faulty theoretical assumptions in speculative metaphysics. [BACK]
10. So-called "Good Samaritan Statutes" often impose duties of this nature: "A person who knows that another is exposed to grave physical harm shall, to the extent that the same can be rendered without danger or peril to himself or without interference with important duties owed to others, give reasonable assistance to the exposed person unless that assistance or care is being provided by others" Vt. Stat. Ann. tit 12, Section 519). [BACK]
11. Philippa Foot, "Are Moral Considerations Overriding?" in her Virtues and Vices (Berkeley: University of California Press, 1978), 186-87. A very rich discussion of the issues raised by Foot is to be found in Samuel Scheffler, Human Morality (Oxford: Oxford University Press, 1992), esp. chapter 4. [BACK]
12. Kant does not always miss this point, of course. Indeed, he characterizes an imperfect duty as one that "permits exceptions in the interest of inclination" (Foundations, 39, note 10). What he means by this is that an imperfect duty such as charity allows each person to exercise some choice over the persons or causes which will be the beneficiaries of his charity (e.g., cancer research or food for the hungry) and the nature and quantity of the sacrifices that will be made (e.g., volunteer work or financial contributions). "Exception" (Ausnahme) is a poor word choice here, since Kant does not really mean to say that one may make exceptions to the duty out of inclination but rather that some inclinations with respect to persons, time, place, and manner may enter into the specification of the duty itself -- the duty being understood informally as a duty to do something for somebody sometime. So specified, the duty allows us some choice -- based on inclination -- of how to fill in the variables. The issue of the limits that might be placed on such choices raises interesting questions, but they must be left for another paper. For a rich discussion of Kant on the duty never to lie, see Christine M. Korsgaard, "The Right to Lie: Kant on Dealing with Evil," Philosophy and Public Affairs 15:4 (Fall 1986), 325-49. For more on the perfect/imperfect duty distinction see my Kant: The Philosophy of Right (London: Macmillan, 1970), 51-53, and especially Lewis White Beck, A Commentary on Kant's Critique of Practical Reason (Chicago: University of Chicago Press, 1960), 147ff. [BACK]
14. As I long ago argued in my Kant: The Philosophy of Right (supra note 12), I think that Kantian universalizability is best understood not as what Seyla Benhabib has called "a silent thought experiment" but rather in terms of models of conversation and agreement of the kind that one finds in the writings of John Rawls. (Kant's own best statement of such a model is to be found in the second section of "Theory and Practice.") Perhaps the question "Is lying under these circumstances universalizable?" is best interpreted as "Would lying in these circumstances be permitted by the best account of the rules of truth telling?" -- where "best account" is interpreted as "the account that would emerge from a certain model of conversation and agreement." Even if social lies are universalizable, however, it is less clear that they are consistent with the duty to treat all persons with the respect owed to them as ends in themselves. For the claim that lying runs into particular problems from the respect for persons (second) formulation of the categorical imperative, see Korsgaard, "The Right to Lie." In Kant's 1797 essay "On a Supposed Right to Lie from Altruistic Motives" (Über ein vermeintes Recht aus Menschenliebe zu lügen) he argues, unpersuasively in my judgment, that all lies fail the universalizability test. See also his argument in the Doctrine of Virtue (Tugendlehre) (chapter 2, section 1) that lying, involving a misuse of our faculty of communication, violates a duty we have to ourselves. (The best translation of the essay may be found in Immanuel Kant, Critique of Practical Reason and Other Writings in Moral Philosophy, trans. Lewis W. Beck [Chicago: University of Chicago Press, 1949], 346-50. The best translation of the passage from the Tugendlehre is to be found in Mary Gregor's translation of the entire Metaphysics of Morals (Metaphysik der Sitten) [Cambridge: Cambridge University Press, 1991], 225-27. A revised Beck translation of the Critique of Practical Reason alone was published by Macmillan in 1992.) For more on the conversation-agreement model of universalizability, see Seyla Benhabib, "Afterword: Communicative Ethics and Contemporary Controversies in Practical Philosophy" in The Communicative Ethics Controversy, ed. Seyla Benhabib and Fred Dallmayr (Cambridge: MIT Press, 1990). [BACK]
15. Kant is here, of course, doing nothing more than stating his famous principle that "ought" implies "can" -- a principle he never actually stated in the exact form in which it is often quoted: "One of Kant's most famous 'statements' -- "Thou canst because thou shouldst" -- does not exist in his writings in this neat form (see David Baumgardt, 'Legendary Quotations and the Lack of Reference,' Journal of the History of Ideas 7 : 116). But statements that express this inference less succinctly abound, e.g., Critique of Practical Reason 30:118-19; Critique of Pure Reason, A807-B835; Über den Gemeinspruch 8:287; Metaphysik der Sitten 6:380; Streit der Fakultäten 7:43-44; Vorlesungen über Metaphysik, ed. Kowalewski, 600; Opus postumum 21:16." (Beck, Commentary, 200n.) [BACK]
16. Kant's response is directed primarily to some explicit criticisms of his views made by Garve in the 1792 Versuche über verschiedene Gegenstände aus der Moral, der Litteratur und dem gesellschaftlichen Leben (Essays on Various Topics from Morals, Literature, and Social Life). The most relevant portions will be found in Henrich, Kant, Gentz, Rehberg. [BACK]
17. See, for example, Foundations, 16ff. Kant's actual claim is that acts have moral worth only if performed from a motive of respect for duty, but agent assessment is (at least in my judgement) his real concern in such passages. As he makes clear in his distinction between pflichtmässig actions (right actions or actions in accord with duty) and actions performed aus Pflicht (actions motivated by duty, he does not think that motives must enter into all relevant moral assessments of actions. See my "Kant's Concept of a Right Action," Monist 51:4 (1967), 574-98. [BACK]
18. The depth of the conflict between such motives and Kantian duty is often overstated. The Kantian can surely grant, for example, the value of these motives and actions based on them so long as they are pursued with the constraints of a basic structure that is just. By far the richest discussion of such matters is to be found in the essays of Barbara Herman -- who argues not merely that Kant can allow a place for such motivational considerations but that he must acknowledge them as "principles of moral salience" in order to apply the categorical imperative. See her "The Practice of Moral Judgment," Journal of Philosophy 82 (1985), 414-36, and "Agency, Attachment, and Difference," Ethics 101:4 (1991), 775-97. [BACK]
19. Some of Garve's other claims really are just simple misunderstandings of Kant, and I will not bother to discuss them in detail. Contrary to Garve, Kant is not a sour killjoy who counsels that human beings should renounce their desires for happiness; he simply argues that they must forget about such motives when determining their moral duties. Neither does Kant's belief that a moral world (e.g., a world governed by a just God) would distribute human happiness in proportion to human virtue commit him to the belief that either human beings or God is motivated by a desire for happiness in seeking this outcome. They are motivated, surely, by the perceived justice of such a pattern of distribution. [BACK]
20. The challenge posed by Garve is essentially the same challenge addressed by Kant in the long second footnote on pages 17-18 of Foundations. The footnote begins this way: "It might be objected that I seek to take refuge in an obscure feeling behind the world 'respect.'" Several of Kant's arguments in the note appear again in the response to Garve. [BACK]
21. Those not familiar with Butler's attempted refutation of the doctrine of psychological egoism that he imputes to Hobbes will find a good general discussion of the matter in the chapter on Butler in C.D. Broad, Five Types of Ethical Theory (Patterson, N.J.: Littlefield, Adams, 1959). [BACK]
22. Any rational demand -- even a prudential one -- can conflict with some powerful desires, and this is why (according to Kant) all demands of reason often appear as imperatives. Moral imperatives appear as categorical, however -- by which Kant means they appear to bind regardless of any of our desires and continue to bind even when they are in conflict with all of our desires. See Lewis White Beck, "Apodictic Imperatives," in his Studies in the Philosophy of Kant (Indianapolis: Bobbs-Merrill, 1965). [BACK]
23. Someone might well argue that Kant is making more out of this than he should. Might one not also feel considerable pride in being able to give up smoking? It is difficult to be moral, but it is also difficult to be prudent. Thus overcoming difficulties in either realm could be an occasion of enhanced self-esteem. Kant may still be on to something, however. Perhaps his adjectives "divine," "sublime," and "greatness" have a special phenomenological fit with duty that is not present for mere prudential rationality. The claims of prudence allow one aspect of the empirical self to dominate another. Perhaps the motive of duty allows one to transcend the given empirical self entirely. This at least seems to be Kant's thought. [BACK]
24. This is not to say that there are no important problems remaining for Kant's account. Kant has, I think, successfully established that there is a distinct moral motive of respect for duty -- a motive that is not reducible to a mere selfish desire for personal happiness. Another argument would be required, however, to show that this is the sole moral motive -- e.g., that sympathy and other "motives of connection" lack moral standing. Kant tends to confuse these motives with selfishness, and thus perhaps he wrongly believes that in giving reasons to discard selfishness as a moral motive he has also given reasons to discard these other motives as well. He is wrong about this, however. A separate argument would be required, and he gives no such argument in "Theory and Practice." [BACK]
25. In "Theory and Practice," Kant makes only one explicit reference to his metaphysical views about freedom of will and to his belief that these metaphysical views are necessarily presupposed to render morality as he understands it possible. This reference is in the brief footnote on freedom at 285, 68-69. [BACK]
27. There are, of course, numerous positive duties on the Kantian theory; for, with respect to any maxim that is not universalizable, the agent has a duty not to perform it -- feeling at that point the direct conflict between duty and inclination that Kant often describes. (For example, since, according to Kant, the maxim of neglecting others in distress is not universalizable, agents have positive duties of beneficence.) The point is that, for Kant, the starting point in developing duties is always some contemplated action that the agent seeks to perform for subjective reasons (e.g., personal happiness). thus the subjective dimension is always present in moral calculation (in the form of maxims) from the outset. In this way, Kant's deontology differs from, for example, that found in divine command theory -- a theory that starts, not with subjectively generated maxims, but with a list of duties imposed at the outset from "outside." [BACK]
28. For a recent discussion of Kant on revolution (and a generous listing of other studies on the topic) see 341ff. of Leslie A. Mulholland's splendid Kant's System of Rights (New York: Columbia University Press, 1990). [BACK]
29. According to Kant, the basic rights are freedom (freedom to be limited only to guarantee like freedom to others), equality (formal equality before the law -- the right of each citizen to have others coerced to protect his freedom and to have his own prospects in life limited solely by his own talents, industry and fortune), and independence (the right to function in a legislative -- i.e., voting-capacity with respect to the laws that bind him). Although, when abstractly stated, these are good candidates for basic rights, Kant's own interpretation of them is occasionally shallow and shows that to a considerable degree he simply adopted the prejudices of his age. For example: He sees differentials in wealth resulting from inheritance as merely a matter of "fortune," and he sees no problem in limiting the right to vote to adult males (children and women being called "naturally" unsuited at 295, 76). Even among adult males the right is to be limited to those who are their "own masters" -- i.e., property owners (not laborers). This latter constraint is the occasion for one of the most bizarre (and unintentionally humorous passages in Kant -- a passage (295-96,76) in which Kant takes pains to attempt to show why barbers and woodchoppers may not vote even though tailors and wig-makers may. It has something to do with the relation in which they stand to the commodities with which they work, but the whole discussion defies intelligibility. Even Kant concludes his discussion by saying: "It is, I admit, somewhat difficult to determine what requirements a person must meet so that he can be his own master." [BACK]
30. Think of strong paternalism as coercing someone on the basis of your own (or the community's) conception of the good. A weaker form of paternalism would involve coercing someone to help him realize his own conception of the good. The stronger the form of paternalism, of course, the more difficult it will be to justify within a Kantian framework. See Gerald Dworkin's "Paternalism" in Morality and the Law ed. Richard A. Wasserstrom (Belmont: Wadsworth, 1971). [BACK]
32. Kant's desire to keep all considerations of happiness or welfare out of moral and political theory may result from a confusion on his part over the concept of "determining ground." His theory of motivation requires that a desire for happiness may not motivate (and in that sense be a determining ground for) the moral will. It does not follow from this, however, that happiness cannot be a right-making characteristic and in that sense be a determining ground for moral duty. (Even a utilitarian could act out of a motive of respect for duty -- something not ruled out merely because the utilitarian defines duty partially in terms of happiness.) A full theory of the human good would, of course, involve much more than happiness. It would not, however, involve only elements that "could never be taught by experience." [BACK]
34. Kant is responding to views expressed by Moses Mendelssohn in his 1783 Jerusalem, oder über religiöse Macht und Judentum (Jerusalem, or on Religious Power and Judaism). Kant's own views on radical evil in his Religion show that he was by no means blind to humanity's dark side. [BACK]
36. As I read this passage I cannot help thinking of Eastern Europe -- particularly the "ethnic cleansing" currently taking place in the former Yugoslavia. We did not get to celebrate the lifting of totalitarianism for more than a few days before human bestiality raised its ugly head in a different form. It is hard not to become discouraged. [BACK]
37. As Philip Quinn pointed out to me in correspondence, having faith in the moral progress of the species does not alone provide reasons for respecting each individual member of that species. This individualized respect may require something very like the kind of religious faith that Kant wants to reject as foundational in his moral theory. Perhaps human dignity itself can only be defended as a "postulate of practical reason." I have started a very superficial exploration of the possible religious basis for Kantian dignity in my "Constitutionalism, Moral Skepticism, and Religious Belief" in Constitutionalism: The Philosophical Dimension, ed. Alan S. Rosenbaum (New York: Greenwood Press, 1988), 239-49, and in my forthcoming "Human Decency and the Limitations of Kantianism" (as yet untitled proceedings of the Sixteenth World Congress of the International Association for Philosophy of Law and Social Philosophy). [BACK]
38. But Kantian dignity does not quite, at least to my mind, capture it all. I have never felt more deeply the moral horror of the Holocaust than when I read of the murder (at Babi Yar) of the central character in D.M. Thomas's novel The White Hotel. So much that was wonderful and precious was lost in the death of this woman (a fictional representative of all the victims of that atrocity), and not all of it is (in my view) to be captured in some ideal moral autonomy (the core idea in Kant's concept of dignity). Kant is perhaps inclined to overmoralize the value incarnate in the human person. [BACK]