References to Kant's moral philosophy are to the following editions:

Critique of Practical Reason, trans. Beck, L. W., 3rd edn. (New York: Macmillan, 1993).Google Scholar

Foundations of the Metaphysics of Morals, trans. Beck, L. W. (New York: Macmillan, 1990).Google Scholar

The Metaphysics of Morals, trans. Gregor, M. (Cambridge: Cambridge University Press, 1991).Google Scholar

¹ The phrase is used by Husserl, cited in Habermas, Jürgen, ‘Morality and ethical life’, in Moral Consciousness and Communicative Action, trans. Lenhardt, Christian and Weber Nicholson, Shierry (Cambridge, MA: MIT Press, 1990), p. 203.Google Scholar By reputation the Kantian moral agent is an autonomous rational individual who can be thoroughly independent of society and culture. See also Murdoch, Iris, The Sovereignty of the Good (London: Routledge and Kegan Paul, 1970), p. 80.Google Scholar

² Habermas, Jürgen, Theory and Practice, trans. Viertel, John (Boston, MA: Beacon Press, 1973), p. 150.Google Scholar

³ 'What is Orientation in Thinking?’, final footnote.

⁴ Anthropology from a Pragmatic Viewpoint, trans. Gregor, M. (Nether lands: Nijhoff), pp. 324–5.Google Scholar

⁵ ‘Is the human race constantly progressing?’ in Kant's Political Writings, second edition, ed. Reiss, H. (Cambridge: Cambridge University Press, 1992), p. 177.Google Scholar

⁶ Religion within the Bounds of Reason Alone, ed. Hudson, H., and Green, T. (New York: Bobbs-Merrill, 1970), p. 89Google Scholar

⁷ Harré, Rom, ‘The demography of the kingdom of ends’, Philosophy, 69/266 (January, 1994)Google Scholar proposes that the social nature of language as proposed in one interpretation of Wittgenstein is a successful remedy for some of Kant's problems. Similarly, Habermas and Karl Otto Apel propose a version of a language-based ideal speech situation as a ‘dialogic’ alternative to Kant's account. Others who have argued for the ‘intersubjectivity’ of value include Korsgaard, Christine, ‘The reasons we can share: an attack on the distinction between agent-relative and agent-neutral values’, Social Philosophy and Policy, 10/1 (January, 1993), 24–51.Google Scholar However, this last paper is concerned with issues arising from work by Nagel and Parfit, and does not consider the issues to be raised in this paper.

⁸ Before explaining aspects of the categorical imperative it may be useful to make clear that although this paper does not consider hypothetical im peratives at any length, a contrast and comparison with them is implicit in the discussion. When the next section describes the rational character of the categorical imperative, it does not imply that hypothetical imperatives are entirely irrational. Similarly, although all moral action is said to be rational, this does not imply that all rational action is moral. For Kant, action is purposive and rule governed (Metaphysics of Morals, 50/223.) This rule or maxim is made up of a goal and a plan, where the latter organizes available means to achieve the goal. Kant uses this structure of action to distinguish categorical from hypothetical imperatives by maintaining, among other things, first, that hypothetical imperatives govern actions where the goal is materially determined by inclination, desire, or our animal (rather than rational) being or interests, and therefore, second, that they cannot be universalized since they apply only to those who share the particular goal. Having a goal that is materially determined by inclination or some other interest does not make the action entirely irrational, for the agent still has to reason out the plan and order the available means to achieve a given materially determined goal; but we may say that such an action is not fully rational since its end is materially determined by the inclinations, desires or interests that the agent is subject to. And Kant intends to contrast hypothetical with categorical imperatives in which the goal is not materially determined, is universalisable and rational. See also Salim Kemal,’ The need for plans, projects and reasoning about ends: Kant, and Williams, ,' International Journal of Philosophical Studies, 2/5 (1997), 187–216Google Scholar, which also argues for the primacy of reason over inclination, desire or happiness and discusses the relation of hypothetical to categorical imperatives. Here it may be useful to clarify that the contrast cannot always be explained fruitfully in terms of a distinction between instrumental and ends reasoning. The satisfaction of inclinations ultimately depends on happiness, which is the best sustained order and relation of satisfactions of inclinations; and Kant sees happiness as itself a kind of ends reasoning even though it is not moral.

9 See Religion within the Bounds of Reason, where Kant says that the choice and its opposite were within their power.

¹⁰ Determination by an agent's own nature precludes freedom as surely as does the determination by an object in nature since the agent's behaviour results from ‘antecedent time[ ] which, with what happened in it, is no longer within our power.’ (Religion within the Bounds of Reason Alone, Essay I) Their actions are expressions of their nature only in the way a turnspit performs movements at the order and speed determined by its machinery. Just as the turnspit does not determine anything by itself, similarly an agent determined entirely by its own nature is only pre-determined rather than free, and such predetermination is inconsistent with autonomy where the latter involves being able to choose an act and its opposite.

¹¹ In the Foundations of the Metaphysics of Morals, Kant explains autonomy in these terms. In Religion within the Bounds of Reason Alone Kant writes that freedom is present in human beings at birth though birth be not the cause of it. This also furthers his conception of moral action because the latter depends on an ability to act freely.

¹² Critique of Judgement, §43.

¹³ Foundations of the Metaphysics of Morals, p. 412.

¹⁴ Ibid.

¹⁵ In the Critique of Practical Reason Kant explains freedom as a capacity to accept or reject the moral law or natural inclinations. He urges that ‘though freedom is certainly the ratio essendi of the moral law, the latter is the ratio cognoscendi of freedom’ (p. 118). Freedom is the basis of moral law since agents cannot pursue moral imperatives unless they are free. But agents know their freedom only through the choices they make. It is not simply a matter of following the moral law because it is rational but includes agents' capacity to choose whether or not to let their actions be determined by inclinations. If people choose to follow their inclinations, they can be accused of immorality for failing to follow the moral law. Further, we cannot expect to explain the source of this freedom. A genetic explanation that refers to causes would be self-defeating where freedom is supposed to contrast with causal determina tion. An explanation in terms of our possessing a capacity for freedom will be useless. The existence of freedom and reason is simply a fact we must accept, whose nature we understand by examining its exercise in choosing ends when we behave intentionally.

¹⁶ See Metaphysics of Morals, p. 223 and Foundations of the Metaphysics of Morals, p. 427.

¹⁷ Critique of Practical Reason, p. 19, Beck, 20–1.

¹⁸ See ‘Need for plans: Kant and Williams’.

¹⁹ In Foundations of the Metaphysics of Morals Kant develops this con ception of universality in terms of agents' disinterestedness. See p. 436f.

²⁰ The categorical imperative is a description of the operation of reason, and so of what rational action would consist in, rather than itself a normative proposal. In his book on Hegel's Ethical Thought (Cambridge: Cambridge University Press, 1991), Allen Wood contends that Kant's deduction moves from a description to the claim that everyone should act in certain ways. This misconstrues the descriptive nature of the categorical imperative itself. Kant's argument says only that, given this description of rational behaviour, any rational action, performed by any agent, will satisfy the categorical imperative. He is not, then, importing a normative content into the general description, as if he were arguing (or assuming where he should be arguing) that because agents can behave rationally, therefore they should always do so. Such an argument would have to look for reasons for being rational and would therefore be self-defeating, and no argument can work here. Kant's derivation is rather a matter of formulating the description of rational action as a standard that any rational action must satisfy (see Metaphysics of Morals, p. 218, where Kant describes the imperative as ‘a merely theoretical cognition of a possible determination of choice, that is of practical reason’). He does also argue, elsewhere, that no other notion of action based on reason, such as the ones enjoined by the hypothetical imperative, can be fully rational, and that the possibility of rational action is available to us as the fact of reason. These and other claims defend his description against other construals of reason, but none of them makes the ‘derivation’ of the categorical imperative into an extrapolation of a normative conclusion from a mere description. Kant is also clear that some agents will feel the standard of rational action as an imperative because their ‘reason is not the sole determinant of the will’ (Critique of Practical Reason, p. 18). Why such creatures, capable of acting according to reason or nature, should prefer to act according to reason is, of course, the real issue, an answer to which allows Kant to show the validity of the moral rational law for imperfectly rational creatures like ourselves.

²¹ Imperfectly rational (human) beings feel the rules of rational behaviour as commands because their imperfection requires them to constrain themselves to act rationally rather than follow their natural inclinations. For Kant this constraint does not imply that human agents possess an imperfect reason; rather they possess reason imperfectly: they do not always behave according to rational principles because they sometimes follow their inclinations instead; but they do not follow an imperfect reasoning. The categorical imperative remains a description of how they would behave if they were acting rationally.

²² Again remembering that we have set aside consideration of hypothetical imperatives which, because their goal is materially determined, are rational in means but not in ends, and whose ends cannot be universalized — so that their imperatives apply only to those who share these ends.

²³ See the Foundations of the Metaphysics of Morals, p. 427 for Kant's distinction between purpose and means as parts of the maxim. Recently other writers have developed interesting and powerful accounts of maxims. See O'Neill, Onora, Constructions of Reason (Cambridge: Cambridge University Press, 1989)Google Scholar and Towards Justice and Virtue (Cambridge: Cambridge University Press, 1996), Herman, Barbara, The Practice of Moral Judgment (Cambridge, MA: Harvard University Press, 1995)Google Scholar and Bittner, R., What Reason Demands (Cambridge: Cambridge University Press, 1974).Google Scholar In the Critique of Practical Reason, §1 Kant distinguishes maxims from practical laws, but arguably practical laws also involve maxims. Kant later distinguishes both hypothetical and categorical imperatives from maxims, which leaves maxims in an odd position. In his Commentary on Kant's Critique of Practical Reason L. W. Beck suggests that Kant has three options: maxims, practical laws, and maxims that are practical laws; and this seems a suitable description.

²⁴ It may be ambiguous, in the sense that it may not be clear to the agent or observer whether a subject is taking aspirin to cure a headache or prevent a heart attack, but without some such description the event remains something less than an action.

²⁵ Assertions such as ‘Eat beef’ or ‘Do not walk on the cracks of a pavement’ are not maxims. ‘Eat beef to gain healthy amounts of protein’ or ‘Avoid the cracks in the pavement to stave off bad luck’ may be maxims because the antecedent sets out a means to the goal in the consequent. The ends may be irrational — animal protein may be bad for people — or the means may be questionable — healthy amounts of protein are available more efficiently from pulses or other kinds of meat — but without both elements, the rule of action will fail to be a maxim at all.

²⁶ A maxim may well be rational subjectively, as a hypothetical imperative, in that it pursues the most rational organization of available means to a materially determined goal, and therefore will be ‘universalizable’ over all subjects who share that determined goal, but it will fail to be (objectively) moral because its end or goal is not universalizable or fully rational.

²⁷ By saying that the maxim must be one that the agent adopts because its goal is one that every rational being can always adopt, I mean to capture the sense that moral agents follow moral maxims because they are moral.

²⁸ Foundations of the Metaphysics of Morals, p. 426.

²⁹ See the parallel from the Critique of Pure Reason about the impossibility of more than one person having the same experience. Kant raises the issue in criticizing the second paralogism, and rejects the metaphysical claim that it sometimes supports; but he is careful not to disavow the unity of the self or agency (cf. A356ff. and B407). An excellent discussion of the issues involved in identity claims in relation to the first Critique is to be found in Doepke, F. C., The Kinds of Things. A Theory of Personal Identity Based on Transcendental Argument (Chicago: Open Court, 1996).Google Scholar

³⁰ Foundations of the Metaphysics of Morals, p. 427.

³¹ See Critique of Pure Reason, Transcendental Deduction and the Third Paralogism.

³² Foundations of the Metaphysics of Morals, pp. 445–6.

³³ Kant recognizes this as an issue since he has already raised similar issues in relation to cognition where, in addition to distinguishing the noumenal realm from its phenomenal counterpoint, he also identifies the formal ‘I’ of the ‘I think’ accompanying all representations in the transcendental unity of apperception as only the representation of a subject and not the premiss from which to draw conclusions about actually existent souls or being (see the Paralogism of the first Critique). That formal unity of apperception, moreover, stands as a counter to a fragmented self that could not sustain any single complex thought (see Critique of Pure Reason §§18–26). Issues of a unified self, its representation, and its identification with a will, including a transition from noumenal to phenomenal considerations, are all items in the account Kant wants to give of knowledge and action. And given the distinctions he makes, it is not surprising that he should suggest that beyond actions we need to consider the agents who will carry out those actions — and the agents may be of different kinds: purely rational agents, imperfectly rational agents, those possessed of holy wills, and those having only ordinary wills, those subject to natural inclinations and those given wholly to pursuing natural ends or happiness being some varieties.

³⁴ See the extended discussion of ‘disposition’ in Religion within the Bounds of Reason Alone.

³⁵ Metaphysics of Morals, p. 223.

³⁶ Foundations of the Metaphysics of Morals, p. 434. Italics added.

³⁷ See, for example, Kant's reminder that ‘self-consciousness would make [an] automaton a thinking automaton, but it would be a mere illusion if the consciousness of its spontaneity were taken for freedom’. Critique of Practical Reason, p. 101.

³⁸ See Theory and Practice, where Kant criticizes the empiricists for looking for moral motives through introspection: his criticism is that they are looking in the wrong place for the wrong thing when they turn to introspection.

³⁹ We might explain this inner dialogue as one between moral and psychological personalities, thereby denoting the kinds of conflict and issues that motivate moral agents, and do so without seriously fragmenting agency.

⁴⁰ By this means they can arrive at universalizable judgements — maxims which others could also follow — without using others as part of their calculation. Indeed, unless they had some such ability to recognize rational principles, they could not be moral. For even if the universality of a maxim were shown by others being able to follow a given rule, they would have to be able to distinguish a moral maxim from an accidentally accepted generalization which a number of people happen to accept.

⁴¹ Foundations of the Metaphysics of Morals, p. 427.

⁴² That is, they are negatively free from inclinations and determination by material conditions and positively free in being done for themselves, because they are universalizable, rather than for any interest.

⁴³ Foundations of the Metaphysics of Morals, p. 429

⁴⁴ See O'Neill, Onora, ‘Universal laws and ends in themselves,’ in Construc tions of Reason, Explorations of Kant's Practical Philosophy (Cambridge: Cambridge University Press, 1989)Google Scholar and Hill, Thomas E. Jr., ‘Humanity as an end in itself’, in Dignity and Practical Reason in Kant's Moral Theory (Ithaca: Cornell University Press, 1992).Google Scholar

⁴⁵ Foundations of the Metaphysics of Morals, p. 428. Kant also suggests here that unless there were some objective end, ‘nothing at all of absolute value would be found anywhere. But if all value were conditional — that is, contingent — then no supreme principle could be found for reason at all’ (Foundations of the Metaphysics of Morals, pp. 428–9). A supreme principle would be useless if it lacked all purchase because nothing satisfied its conditions. In this case, if nothing could be the object of universal law, then the imperative would be otiose. Conversely, if there is an object of the universal law, something that could fall under universal laws, such as rational nature as agency, then that rational nature would also show that the categorical imperative had validity and could legitimately guide action.

⁴⁶ Critique of Practical Reason, §90, p. 87

⁴⁷ See Metaphysics of Morals, pp. 387, 435, and Foundations of the Meta physics of Morals, pp. 436–9.

⁴⁸ Foundations of the Metaphysics of Morals, p. 431. It may need to be made clear here that treating respect for humanity in terms of respect for the ends which individuals choose does not imply that moral agents owe unconditional respect to others as moral persons merely on the grounds of others' capacity to pursue conditioned goods. In the Metaphysics of Morals Kant does ascribe to us a duty to the happiness of others, thereby seeming to propose an unconditional respect to others' capacity to choose unconditioned goods, but even there he also restricts happiness by requiring that it be suitable to the agent's capacity for acting morally. Thus, respect for others' choice of ends is always constrained by their moral being and behaviour.

⁴⁹ Anthropology from a Pragmatic Point of View, p. 277. In this instance, though, this seems to result from the fact of our nature.

⁵⁰ Foundations of the Metaphysics of Morals, p. 431.

⁵¹ Ibid., 428.

⁵² Ibid., 433.

⁵³ A832/B860.

⁵⁴ A833/B861.

⁵⁵ Critique of Judgement, §65.

⁵⁶ Foundations of the Metaphysics of Morals, p. 435.

⁵⁷ Kant associates this argument for a duty sui generis with the second Critique's postulates of God and the immortality of the soul, though by the time of Religion within the Bounds of Reason Alone for Kant the immortality of the soul seems to have become subsumed within the continued existence of the moral community while, in the Opus Postumum, he says that when we need to think of the categorical imperative as a command, we must think of it as a divine command — though for reasoning beings the categorical imperative is not a command at all but a description of, and standard for, their rational behaviour.

⁵⁸ In the first footnote of Religion within the Bounds of Reason Alone, Kant says that a reference to consequences is important even if it does not determine what is moral.

⁵⁹ ‘On the old saw: That may be true in theory but will not work in practice’, p. 60.

⁶⁰ Ibid., p. 61.

⁶¹ In part, these maxims are appropriate to the ‘subjectivity’ of the activity of judging. Lacking any rules, judgement cannot rely on objective criteria derived from the nature of objects. Judgement is the ability to discern order and rules in material, and while the order it discerns may be capable of objective validity, judging claims only validity over subjects. Subjective validity as such needs other subjective validation to bolster it — it is by nature intersubjective.

⁶² A134/B173–4.

⁶³ Critique of Judgement, §40.

⁶⁴ Ibid., §40, 294.

⁶⁵ Critique of Pure Reason, A738/B766.

⁶⁶ Critique of Judgement, §40.

⁶⁷ Ibid.

⁶⁸ Ibid.

⁶⁹ A821/B850.

⁷⁰ This claim is not an isolated instance in Kant's writings. This is how moral life works.

⁷¹ See Metaphysics of Morals, where Kant expresses the injunction to ‘Know thyself’ in terms of erforschen and ergründen. Neiman, Susan, The Unity of Reason (Cambridge, MA: Harvard University Press, 1994)Google Scholar, p. 134, contains a short discussion of this issue.

⁷² Foundations of the Metaphysics of Morals, p. 393

⁷³ Ibid., p. 433.

⁷⁴ Ibid., p. 431–2.