Hostname: page-component-cd9895bd7-gbm5v Total loading time: 0 Render date: 2024-12-26T20:48:35.519Z Has data issue: false hasContentIssue false
  • English
  • Français

Necessitation and Justification in Kant’s Ethics

Published online by Cambridge University Press:  01 January 2020

Mark Timmons
Mark Timmons*
Affiliation:
Memphis State University, Memphis, TN, 38152, USA
Get access Rights & Permissions [Opens in a new window]

Extract

In the Groundwork of the Metaphysic of Morals, Kant claims that hypothetical imperatives are analytic and that categorical imperatives are synthetic. This claim plays a crucial role in Kant’s attempt to establish moral ‘oughts’ as categorically binding on all rational agents, for by classifying moral statements according to this distinction, Kant hopes to uncover the sort of justification required to establish such statements. However, Kant’s application of the analytic/ synthetic distinction to imperatives is problematic. For one thing, this distinction was developed by Kant in connection with indicative, subject-predicate statements, which would seem to cast doubt on the idea that imperatives can be either analytic or synthetic. Moreover, Kant’s claim that hypothetical imperatives are analytic and categorical imperatives are synthetic seems to conflict with other claims in his moral works. For example, in the Groundwork, Kant claims that hypothetical imperatives are (compared to categorical imperatives) unproblematic since they can be established on the basis of experience (G, 419-20, 49-50). But this is incompatible with the idea that hypothetical imperatives are analytic, since presumably all analytic statements are a priori — established independently of experience.

Type
Original Articles
Information
Canadian Journal of Philosophy , Volume 22 , Issue 2 , June 1992 , pp. 223 - 261
Copyright
Copyright © The Authors 1992

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 See Groundwork 417-20, 84-8. I will incorporate references to Kant’s work into the text, supplying both Akademie edition numbers and page numbers of the translated edition in that order. The English language editions used and my abbreviations for them are: Kritik der reinen Vernunft, 1781, trans. by Smith, N.K. Critique of Pure Reason (New York: St. Martin’s Press 1929)Google Scholar, abbr. KrV; Grundlegung zur Metaphysik der Sitten, 1785, trans. by Paton, H.J. Groundwork of the Metaphysic of Morals (New York: Harper & Row 1948)Google Scholar, abbr. G; Kritik der praktischen Vernunft, 1788, trans. by Beck, L.W. Critique of Practical Reason (New York: Bobbs-Merrill 1965)Google Scholar, abbr. KpV; Die Religion innerhalb der Grenzen der blossen Vernunft, 1794, trans. by Greene, Theodore M. and Hudson, Hoyt W. Religion Within the Limits of Pure Reason Alone (New York: Harper & Row 1960)Google Scholar, abbr. R; Metaphysik der Sitten, 1797, trans. by Gregor, Mary The Metaphysics of Morals (Cambridge: Cambridge University Press 1991)Google Scholar, abbr.MdS.

2 Some commentators have denied that the analytic/synthetic distinction applies to imperatives at all. See, for example, Beck, L.W. A Commentary on Kant’s Critique of Practical Reason (Chicago: The University of Chicago Press 1960), 87Google Scholar, and Wolff, R.P. The Autonomy of Reason (New York: Harper & Row 1973), 141Google Scholar.

3 Although (on my view) Kant analyses ‘oughts’ as claims about what a completely rational agent would intend or do, such claims, because they represent idealizations are to some extent normative. Still, I think it is useful to characterize such claims as at least broadly descriptive, though I do not insist on some sort of sharp descriptive-normative distinction. Moreover, this does not jeopardize the strategy of analyzing ‘ought’ statements in the way I propose, because the analysis is not meant to reduce, in the sense of analyze away, moral and other normative statements. Rather the analysis is meant to illuminate the application of the analytic/synthetic distinction to ‘ought’ statements.

4 See, for example, Paton, H.J. The Categorical Imperative (Philadelphia: University of Pennsylvania Press 1971)Google Scholar, ch. 24; Aune, Bruce Kant’s Theory of Morals (Princeton, NJ: Princeton University Press 1979), 35-7Google Scholar; and Jr.Hill, Thomas E. ‘The Hypothetical Imperative,’ Philosophical Review 82 (1973) 429-50CrossRefGoogle Scholar. These authors seem to agree that for Kant, ‘oughts’ are equivalent to descriptive statements about completely rational agents, but fail to explain the equivalence. Their failure is due to not having appreciated the importance of practical necessitation. The only writer I know of who has appreciated the role of this concept in Kant’s moral epistemology is McCarthy, Michael H. See his ‘Analytic Method and Analytic Propositions in Kant’s Groundwork,’ Dialogue 15 (1976) 565-82CrossRefGoogle Scholar; ‘Kant’s Application of the Analytic/Synthetic Distinction to Imperatives,’ Dialogue 18 (1979) 373-91; ‘Paton’s Suggestion that Kant’s Principle of Autonomy Might be Analytic,’ Kant-Studien 70 (1979) 206-24; and ‘The Objection of Circularity in the Groundwork III,’ Kant-Studien 76 (1985) 28-42. However, McCarthy mistakenly analyses necessitation on the model of inter-personal coercion. For a critique of McCarthy’s view, see my ‘McCarthy on Practical Necessitation in Kant,’ Kant-Studien 80 (1989) 199-208.

5 According to Paton, ‘Once [Kant] has established the principle that a rational agent as such, if he wills the end, must necessarily will the means, Kant finds no difficulty —perhaps he should have found more — in turning it into a hypothetical imperative: he takes it for granted that if anything is what a rational agent as such would necessarily do, it is also what a rational agent ought to do, should he be tempted otherwise. Exactly the same assumption is made in the case of the categorical imperative ... ‘ (247). Because Paton interprets Kant as taking it for granted that ‘oughts’ are to be understood in terms of what a (fully) rational agent would do, he more or less passes over this important doctrine without much discussion. What is needed, then, is a study of Kant’s understanding of ‘ought’ statements, which I undertake in this paper.

6 See the McCarthy articles cited above in n. 4.

7 Cf. MdS 222-3, 48-9 where Kant claims that moral obligation (i.e., moral ‘ought’ statements) contain necessitation. In various places, Kant claims that duty, which he defines as ‘an action to which someone is bound’ (MdS 222, 49) implies necessitation. See, e.g., MdS 379, 185; 394, 197-8; and 401-2,203.

8 This claim is controversial within the ranks of Kant commentators. Atwell, John E. in Ends and Principles in Kant’s Moral Thought (Dordrecht: Martinus Nijhoff 1986), 50-1CrossRefGoogle Scholar, argues that maxims are not intentions; Onora O’Neill, ‘Consistency in Action,’ inN. Potter, T. and Timmons, M. eds., Morality and Universality (Dordrecht: D. Reidel 1985) 161-7CrossRefGoogle Scholar, argues that maxims are fundamental intentions — very general intentions that guide agents to accept more specific intentions on the basis of which they perform specific actions. Space does not permit me to defend my own view here but see Herman, Barbara Morality as Rationality (New York: Garland 1990)Google Scholar, ch. 2 for an interpretation of maxims that supports my claim.

9 Kant holds that in addition to actions that we have duties to perform or omit, there are also ends or goals that we have duties to adopt or avoid adopting. He elaborates a doctrine of obligatory ends in MdS. See especially the Introduction to the Doctrine of Virtue.

10 Kant says that we make something our maxim as a result of practical reflection. He points out at G 427, 94, that desires and inclinations (i.e., settled desires) arise from feelings of pleasure and displeasure, and from these ‘with the cooperation of reason, there arise maxims.’ At KpV 79, 82, he claims that an interest (which an agent takes up as a result of reflecting on her desires) is the basis, or provides the reason, an agent has for adopting a maxim. Thus, maxims—conditional intentions of the form, ‘I Will ___, if___(where the blanks are filled with a specification of the action and circumstances respectively) are adopted for reasons. This implies that agents could adopt the same maxim but do so for different reasons. (Kant’s example of the shopkeeper at G 397, 65 can be used to illustrate this point.) However, for purposes of moral evaluation of an action, Kant sometimes formulates maxims where the agent’s reason for adopting the maxim is made explicit; e.g., ‘From self-love, I make it my principle to shorten my life if its continuance threatens more evil than it promises pleasure’ (G 422, 89). But notice that in this example, the agent’s reason — self-love — is not part of the maxim; rather, Kant says that from this motive the agent makes it her principle to shorten her life, etc. For more on this, see Allison, Henry E. Kant’s Theory of Freedom (New York: Cambridge University Press 1990)CrossRefGoogle Scholar, ch. 5.

11 In other places (KpV 19, 17; MdS 225, 51) Kant means something different in calling maxims subjective and does not use the objective/ subjective terminology to distinguish maxims from imperatives. In those places, a maxim whose adoption by an agent is based on her desires and inclinations generally is called subjective; a maxim is objective when it ‘qualifies for such a giving of universal law’ (MdS 225, 24).

12 Here I am focusing exclusively on what Gilbert Harman calls ‘principles of revision,’ which he contrasts with maxims of reflection (the term ‘maxim,’ of course, not being used in Kant’s sense of the term). See his Change in View (Cambridge, MA: MIT Press 1986), ch. 1. In the context of deliberation, we can distinguish two possible phases. In the phase that often precedes actually revising one’s intentions, there is a process of reflection in which one, e.g., envisions alternative courses of action, possible outcomes of each action, perhaps tries to vividly imagine what some of those outcomes would be like, and so forth. Maxims of reflection prescribe the various sorts of things one should consider in practical thinking, principles of revision put constraints on how one is to revise her intentions. But in addition to errors that violate these two sorts of prescriptions, a further locus of error concerns the effectiveness of (possibly correct) practical reasoning. For instance, the sway of desire may interfere with following through and acting on the basis of one’s practical reasoning.

13 The phrase ‘reason completely determined the will’ (die Vernunft den Willen gänzlich bestimmte) is apt to mislead here and should not be confused with Kant’s talk of ‘reason determining the will in a practical law directly’ (Die Vernunft bestimmt in einen praktischen Gesetze unmittelbar den Willen) (KpV 25, 24; cf. KpV 31, 31; 46, 48; 48, 49; 62, 64; 71, 74; 78, 81), or again with his claim that ‘The moral law is the sole determining ground of the pure will’ (Das mora/ische Gesetz ist der alleinige Bestimmungsgrund des reinen Willens) (KpV 109, 113; cf. KpV 28, 28). Talk of reason or the moral law being the direct or sole determining ground of the will has to do with Kant’s view of moral motivation — his view that reason itself can be practical. In referring to the sort of necessitation involved in both hypothetical and categorical imperatives, Kant talks of reason completely (or infallibly or decisively) determining the will which should not be understood as claiming that reason alone, independently of all inclination, moves the will. Rather, the idea is that imperatives or ‘oughts’ tell an agent what it is rational to do given her reasons for action — whether they be desire-based or non-desire-based. So, for example, Kant describes the formal principle of hypothetical imperatives as: ‘Who wills the end, wills (so far as reason has decisive influence on his actions) also the means …’ (G 417, 84-5) and again: ‘Who wills the end, wills also (necessarily if he accords with reason) the sole means which are indispensably necessary and in his power’ (G 417, 84-5).

14 At MdS 226, 51, Kant writes: ‘The supreme principle of the doctrine of morals is, therefore: Act on that maxim that can also hold as a universal law. Any maxim which does not so qualify is contrary to morals’ (my emphasis). This does not mean that Kant's categorical imperative, requiring us to act on a certain sort of maxim, is solely a test of the morality of maxims. Rather, Kant’s principle is properly taken as a test of the moral permissibility of actions; the test requires that we focus on maxims from which actions flow.

15 This form of ethical intemalism is called reasons internalism: there is a necessary connection between ought and reasons for action. Sometimes ethical intemalism is explained as the view that there is a conceptual connection between obligation (or recognizing an obligation) and being appropriately motivated to act. For our purposes, any differences between motivational internalism and the sort of reasons intemalism I attribute here to Kant is of no importance. Following Darwall, Steven Impartial Reason (Ithaca: Cornell University Press 1983)Google Scholar, I think it is plausible to suppose that reasons for action are to be understood counterfactually in terms of how one would be motivated under certain conditions. So the difference between these two types of internalism may not amount to much anyway.

16 See n. 5 above.

17 McCarthy, ‘Paton’s Suggestion,’ presses this objection against Paton’s view. Again, the same point would apply to the views contained in Hill and Aune.

18 The universal law formulation of the categorical imperative expresses those formal considerations that provide rational agents with reasons for action independently of what they might happen to desire, viz., that one’s maxim have the essential properties of law: necessity and strict universality. A perhaps more illuminating characterization of the such non-desire-based reasons for action is provided by the Principle of Humanity formulation of the categorical imperative which provides a substantive characterization of such reasons. Kant claims that humanity is an end in itself — its value is independent of any contingent ends or goals we may adopt and thus serves as reason for action independently of desire and inclination.

19 Notice that according to my analysis of ‘oughts,’ the subject concept of the correlated descriptive statements is the concept of an incompletely rational agent reasoning is a fully rational manner or, equivalently, a fully or completely rational agent (will). There is textual support for this. In the Groundwork, Kant writes: ‘If there is such a [moral] law, it must already be connected (entirely a priori) with the concept of the will of a rational being as such’ (G 426, 94). What Kant goes on to claim is that mere analysis of the concept of the will of a rational being (or, equivalently, of a fully rational will) does not include adopting the sort of universalizable maxims required by the moral law. More on this in the next section.

20 This claim requires some comment. There is unclarity and disagreement over whether Kant should be understood as claiming that particular hypothetical imperatives, e.g., ‘If you intend to gain weight, you ought to eat more,’ are analytic or whether only the principle of heteronomy is analytic. I claim both are analytic simply because particular hypothetical imperatives are instantiations of the principle of heteronomy. However, L.W. Beck, 87, argues that only the principle of heteronomy is analytic, particular hypothetical imperatives (he thinks) are synthetic. Against Beck, it should first of all be noted that Kant claims that particular hypothetical imperatives are analytic (see, e.g., G 117-19, 84-6). But more importantly, Beck’s reason for claiming that particular hypothetical imperatives are synthetic is that means-ends reasoning is required to establish claims about what actions are required to achieve whatever ends an agent sets for herself, and such claims are synthetic. And, of course, it is doubtless true that these means-ends claims are synthetic, but this fact only shows that one component in the antecedent of a conditional is synthetic, and this is irrelevant to whether the conditional as a whole is analytic. After all, the following proposition is analytic: ‘If John gains weight and gains weight only if he eats more, then John eats more’ despite the fact that both conjuncts of the antecedent are themselves synthetic propositions.

Within the category of hypothetical imperatives, Kant distinguishes imperatives of skill from imperatives of prudence. The former concern some specific chosen end and the means necessary to achieve that end, while the latter concern the end of happiness which Kant presumes that all incompletely rational agents have. Regarding the latter, he says that since we lack any determinate notion of happiness (not being omniscient, we are not able to tell what will lead to our own happiness), prudential imperatives are not, strictly speaking, analytic. ‘If it were only as easy to find a determinate concept of happiness, the imperatives of prudence would agree entirely with those of skill and would be equally analytic’ (G 417, 85). But the fact that we do not know how to achieve happiness does not show that imperatives of prudence fail to be analytic. Given that for any individual there is a fact of the matter concerning what would make her happy, we can say that there are hypothetical imperatives that specify the relevant ends, satisfaction of which would constitute the agent’s happiness. Such imperatives are analytic. Because our knowledge about what would make for our own happiness is limited, there are analytic truths that we do not or perhaps cannot know.

21 What I’m calling the principle of heteronomy is usually called ‘the formal principle of hypothetical imperatives’ or, more simply, ‘The hypothetical imperative.’ But Kant does claim that all so-called heteronomous practical principles, i.e., principles that prescribe actions for desire-based reasons, ‘can only give rise to hypothetical imperatives: “;I ought to do something because I will something else”’ (G 441, 108). So ‘principle of heteronomy’ is appropriate to use for the most general principle associated with particular hypothetical imperatives.

22 See especially G 417-20, 84-8.

23 There are at least three interpretations of the principle of heteronomy. A subjectivist version would replace ‘recognizes’ with ‘believes’ or possibly ‘justifiably believes,’ while an objectivist version would simply delete ‘recognizes.’ The version I favor is a hybrid version of the principle. The objectivist version won’t do, since whether or not adopting a maxim is rational for a person at a time depends on the person’s perspective. It may not be rational for a person who is completely unaware of the fact that performing a certain action is necessary to achieve some end to adopt an intention to perform that action. Indeed, doing so may be positively irrational! A subjectivist version, where all that is required is that the agent believes that doing A inC is necessary for her bringing about E, seems wrong. After all, if S irrationally believes that some action is necessary for achieving some end, she isn’t being rational in adopting the maxim to do the action. A more sophisticated subjectivist version that refers to what S justifiably believes may be adequate. I think my own hybrid version is a better interpretation of this principle, but shall not argue for that claim here. Nothing in what follows turns on my preference for a hybrid over a sophisticated subjectivist version of the principle of heteronomy. The principle of autonomy also admits of at least three interpretations and I shall present a hybrid version.

24 Actually, as they stand, H and H* are falsifiable and so certainly not analytic. Consider this example. I will that I now drink an ice cold beer. The only one available is the one Jones possesses and he’ll part with it only if I pay him $100. But it surely doesn’t follow that I ought to pay him the $100 even if I have the money. Following Hill, we should recognize that the rational force of a hypothetical imperative requires that one intend to take the necessary and available means or give up the end.

25 There is an historical reason for working with a minimal conception of rationality. Kant’s ethical views are (I think) best understood as a response to the ethical empiricism of Hume. At the heart of Hume’s ethical views is a conception of practical rationality from which he concluded that reason is and ought to be slave of the passions. Kant, of course, takes issue with Hume on the practical powers of reason. But instead of simply packing the concept of rationality in order to easily analyze out of it the principle of autonomy, Kant takes Hume’s challenge seriously and attempts to argue that there is a necessary, although synthetic, connection between the concepts of rationality and acting on universalizable maxims.

26 Actually, these principles do double duty for Kant: they function both as principles of practical reasoning and as principles governing correct judgment. Construed as principles of the first sort, they are concerned primarily with the rational revision of one’s intentions (maxims). Textual support for this reading of these principles can be found in Kant’s remarks about the moral law determining the will (G 400, 68; 402, 69-70); in his claims that the principle of autonomy can be expressed as ‘Maxims must be chosen as if they had to hold as universal laws of nature’ (G 436, 104); and in claims about willing in accordance with hypothetical and categorical principles (G 416, 84). Such passages support the idea that, for Kant, imperatives are principles governing correct modifications of intentions. However, other passages make clear that Kant also thinks of the principle of autonomy as a principle governing correct judgment or belief about right and wrong action. See, e.g., G 402-04, 70-1, where he says that the principle of autonomy functions at least implicitly as ‘norm of judgment’ in ordinary moral judgment.

I see nothing problematic about construing H and A as serving both roles. However (below) I reformulate H in a way that makes its role as practical principle more explicit.

27 By talking about a stable cognitive system I don’t mean to suggest that the system is unchanging or stagnant; rather, I have in mind a dynamic system that enjoys diachronic stability as a result of revisions within the system being made in ways that respect internal consistency and closure.

28 See, for example, G, 424, 91.

29 For more on this see Harman, 3-4, in which he argues that reasoning should be distinguished from argument.

30 Notice that this analytic statement does not wear its analyticity on its sleeve, so to speak; the predicate concept here is not obviously contained in the subject concept. In Kant’s Logic, 1800, trans. R. Hartman and W. Schwartz (Indianapolis and New York: Bobbs-Merrill 1974), 118, Kant distinguishes explicit from implicit analytic statements. ‘The identity of concepts in analytic judgments can be either explicit (explicita) or non-explicit (implicita) ... Implicitly identical propositions ... are not void of consequences or fruitless, for they clarify the predicate which lay undeveloped (implicite) in the concept of the subject through development (explicatio).’ The sort of analytic statement with which we are concerned, then, is of the implicit variety. (I thank my colleague Torn Nenon for this citation.)

31 That the very notion of doxastic rationality includes consistency and closure principles like D and D’ see Harman, ch. 2, and Cherniak, Christopher Minimal Rationality (Cambridge, MA: MIT Press 1986)Google Scholar, ch. 1. Cherniak distinguishes minimal from ideal rationality in belief and action and formulates weakened versions of the consistency and closure principles in connection with the idea of minimal rationality. Here, of course, we are interested in the notion of ideal rationality, but ideal for human beings, one type of what Kant would call a finite rational being.

32 I refer the interested reader to Aune, Bruce Reason and Action (Dordrecht, Holland: D. Reidel Press 1977)CrossRefGoogle Scholar, ch. 4, and to Marshall, John ‘The Hypothetical Imperative,’ American Philosophical Quarterly 19 (1982) 105-14Google Scholar.

33 The second premise is usually interpreted as expressing a belief on the agent’s part. Because I favor what I have called a ‘hybrid’ interpretation of the principle of heteronomy (see n. 23), I shall, for present purposes, construe this premise as expressing what the agent recognizes to be the case. One result of doing so will be that if, in deliberating about what to do, the agent were to reject the conclusion of a bit of reasoning exemplifying I, she will then be forced to abandon premise 1. By contrast, were the second premise to be understood as representing an agent’s (perhaps false) belief, then upon rejecting the conclusion she could hold onto the intention expressed in 1 and give up the belief expressed in 2.

34 Earlier I claimed that hypothetical and categorical imperatives are distinguished for Kant by the fact the former are associated with desire-based reasons for action while the latter are associated with non-desire-based reasons for action. (See above, 1.) But some cases seem to resist this neat division. Consider, for example, a case in which I morally ought to bring about some end E and I recognize that my doing A is the only way I can bring about E. In this case, wouldn’t H generate a reason for my performing A, but one that is not desire-based? (Terry Horgan raised this question in conversation.) I reserve full treatment of this issue for another occasion, but I’m inclined to suppose that in such a case, the performing of A (in the circumstances described) would be required by application of the principle of autonomy to the relevant maxim. If so, then the use of H to generate a reason for doing A is superfluous. Of course, if I also intend E owing to some desire-based reason, then I have both a desire-based and a non-desire-based reason for action.

35 The tests associated with the principle of autonomy require us to make the fictitious assumption that we have the capacity to legislate universally. As O’Neill, Onora (formerly Nell) points out in her Acting on Principle (New York: Columbia University Press 1975), 69Google Scholar, this assumption is indicated in the universal law formulation of the principle of autonomy, ‘Act as if the maxim of your action were to become through your will a universal law of nature’ (G 421, 89).

36 Notice that in this case, in order to generate the inconsistency in willing, there is no need to consider one’s maxim becoming universal law. This is because, so far as I can tell, Kant’s argument here is circular since his claim about what we would necessarily will as rational agents is really a moral claim in disguise. In the Metaphysics of Morals, Kant argues that we have a moral obligation to adopt a maxim of perfecting ourselves because The capacity to set oneself an end — any end whatsoever — is what characterizes humanity (as distinguished from animality)’ (392, 195). I construe this passage as an attempt by Kant to claim that certain essential features of human beings, viz., our end-setting capacities, provide any rational agent with a reason for action, regardless of that agent’s desires. Space does not permit investigation into this claim, but see Hill, Thomas E. ‘Kant’s Theory of Practical Reasoning,’ Monist 72 (1989) 363-83, esp. 377CrossRefGoogle Scholar.

37 I assume that intending that some state of affairs, E, be the case, involves intending at least some of the ‘normal and predictable’ results of E’s obtaining. And certainly intending that E be the case involves intending anything that logically must be the case if E obtains. For more on this, see O’Neill, Acting on Principle, 70 ff.

38 In this case, as I’ve just remarked, Kant is presupposing that we all, from time to time are in need, and relative to this presupposition, the maxim: ‘I will that others help me if I am in need,’ and the maxim (that one is committed to in willing the non-helping maxim as a law of nature): ‘I will that others not help me if I am in need’ are inconsistent maxims. Although many of Kant’s maxims amount to conditional intentions of the form, ‘I will ___ if___,’ the ‘inevitable’ maxim Kant is here concerned with seems better expressed by the form: ‘I will____ When____,’ which makes the presupposition of being in need clear. Perhaps for Kant, then, practical reasoning takes place against the background of certain assumptions about, for example, the human condition, including our being creatures who have needs and our not being self-sufficient. If so, then practical reasoning takes place against the sort of presupposition mentioned above.

39 Kant’s talk of maxims that we ‘inevitably’ adopt is problematic. Here, in formulating U, I am simply formulating one of the key ingredients that operates in one of Kant’s Groundwork examples, and for our present purposes this is all we need. What to make of maxims one inevitably adopts (on the basis of wants one inevitably has) and whether U should be taken to represent part of Kant’s considered understanding of his universalization test requires some study.

40 For a discussion of how Kant attempts to establish that one’s own perfection and promoting the well-being of others are fundamental unconditionally rational maxims of ends, see Potter, Nelson T. ‘Kant on Ends That are at the Same Time Duties,’ Pacific Philosophical Quarterly 66 (1985) 78-92CrossRefGoogle Scholar.

41 Obviously much more needs to be said in making clear the sort of inconsistency involved here. There is a rather large body of literature devoted to making sense of this sort of impossibility. My own contribution to that literature is contained in ‘Contradictions and the Categorical Imperative,’ Archiv für Geschichte der Philosophie 66 (1984) 294-312.

42 There is much to say about the connection between the constraints on intention revision represented by the principle of autonomy on the one hand, and rationality on the other. Certainly my comments don’t tell the whole story. For an illuminating discussion of Kant’s theory of practical rationality that attempts to explain in some detail why the requirement of adopting only universalizable maxims represents a constraint on rationality in the realm of deliberation and action, see Reath, Andrews ‘The Categorical Imperative and Kant’s Conception of Practical Rationality,’ Monist 72 (1989) 384-410CrossRefGoogle Scholar.

43 More precisely, Kant tells us that the concept of positive freedom ‘directs us’ to the third term. On one reading, the third term in question is ‘rational being qua member of an intelligible world.’ (See G 450-453, 118-121.) For discussion, see Allison, ch. 12.

44 Cf. R 26, 21, quoted below in III.

45 As we have seen, Kant does claim that this question asks how we can ‘conceive the necessitation of the will’ (G 417, 84), but precisely what all is involved in so conceiving the will is none too clear.

46 Of course, someone who denied that moral statements are truth valued can still raise this issue. That is, unless one holds that moral statements are completely arbitrary, one can deny the truth-valuedness of moral statements and raise the issue of whether moral statements can be shown to be ‘valid’ or ‘correct,’ where these notions are not identified with truth.

47 Thus, the issue of scope concerns the truth of particular ‘ought’ statements. However, as we shall see, there are interesting differences between this and the truth issue, and distinguishing them is useful for untangling some of what Kant writes.

48 There is a strong and a weak sense in which a principle might be authoritative for an agent. A principle is weakly authoritative just in case the agent has a reason to comply with it. A principle is strongly authoritative for an agent just in case the agent has a sufficient or overriding reason to comply with it. Obviously, the project of showing that a principle is strongly authoritative is more demanding that showing that it is weakly authoritative. Space doesn’t permit pursuing this matter here, so in what follows, I will interpret the authoritative problem weakly, though certainly Kant’s intention is to establish the stronger claim in connection with moral ‘oughts.’

49 There may be some uncontroversial sense in which the fact that someone ought to do something is a reason for her to act accordingly. But the agent who asks for a reason to be moral (or prudent) is asking what reason she has for so acting. Whether or not there is a necessary connection between its being the case that one ought to do something and one’s having a reason is disputed. Ethical intemalists affirm the connection; externalists deny it. Above (1), I classified Kant as an internalist, a claim argued for in my ‘Kant and the Possibility of Moral Motivation,’ The Southern Journal of Philosophy 23 (1985) 377-98.

50 Of course, for any proposed hypothetical imperative there is the interesting non- conceptual question about the truth of the means-end connection mentioned in the antecedent of the imperative. But, as explained above (seen. 20), the question of the truth of the means-ends connection is not the same as the question of the truth of the conditionally formulated imperative itself.

51 Wolff, 141-51, makes this complaint about Kant’s treatment of hypothetical imperatives.

52 Here, I’ve simply used N* as the basis of this formulation of my analysis of ‘oughts.’

53 We assume that the agent being addressed is an incompletely rational agent who, on the basis of her desires, intends that she control her blood pressure.

54 Talk of the scope of particular hypothetical imperatives can be understood in at least two ways depending on how they are formulated. We can distinguish between undetached and detached hypothetical imperatives. Imperatives of the former sort have the form: ‘If S___, then S ought _____; where the first blank is filled by a specification of the reasons that make the action (or end) specified in the second blank rational for an agent to do (omit). Now undetached particular hypothetical imperatives are addressed to all rational agents of a certain sort — agents whose will is affected by desire. Detached hypothetical imperatives have the form: ‘S ought _____; where the blank is filled as before. In this paper, I am concerned with the scope of detached hypothetical imperatives. And, of course, whether some such imperative correctly applies to an agent at a time depends on that agent’s desire-based reasons for action — reasons not universally shared by human agents.

55 However, such evidence will not be conclusive. In connection with moral motives or reasons for action (i.e., the thought that something is one’s duty), Kant claims that ‘it is absolutely impossible for experience to establish with complete certainty a single case in which the maxim of an action in other respects right has rested solely on moral grounds and on the thought of one’s duty ... [since] we can never, even by the most strenuous self-examination, get to bottom of our secret impulsions’ (G 407, 74-5). For similar reasons, we cannot establish with certainty that an agent has acted from some particular desire, or even which desires he has.

56 Kant begins G III with the claim the ‘Will is a kind of causality belonging to living beings so far as they are rational. Freedom would then be the property this causality has of being able to work independently of determination by alien causes.’ See Hill, ‘Kant’s Theory of Practical Reasoning,’ 363-83, for a discussion of the various negative features associated by Kant with negative freedom.

57 For a detailed reconstruction of Kant’s transition from the concept of negative freedom to the concept of positive freedom, see Hill, Thomas E. ‘Kant’s Argument for the Rationality of Moral Conduct,’ Pacific Philosophical Quarterly 66 (1985) 3-23CrossRefGoogle Scholar. According to Hill’s analysis, the transition is purely conceptual so that the claim that if a being is negative free, that being is positively free is analytic.

58 This claim, of course, is part of Kant’s so-called ‘reciprocity thesis,’ viz., that the notions of positive freedom and the moral law imply one another; that there is an analytic connection between the notions of autonomy and the moral law. Actually, the transition from the notion of autonomy to the notion of the moral law is problematic as many critics of Kant have argued. For an extensive discussion of this doctrine in Kant, and a reply to the critics, see Allison, Henry E. ‘Morality and Freedom: Kant’s Reciprocity Thesis,’ Philosophical Review 95 (1986) 393-425CrossRefGoogle Scholar; and especially Kant’s Theory of Freedom, ch. 11.

59 For their generous help, I wish to thank David Anderson, Robert Audi, Denny Bradshaw, Michael Gorr, Terry Horgan, Len Lawlor, Tom Nenon, Nelson Potter, John Protevi, John Tienson, Bill Tolhurst, an editor and two anonymous referees for this journal.