1 The editions used and citation abbreviations of Kant's works are as follows (in all cases Akademie pagination is cited): Critique of Pure Reason, Smith, Norman Kemp trans. (New York: St. Martin's 1929)Google Scholar, with the two editions abbreviated as ‘A/B'; Religion within the Limits of Reason Alone, Greene, Theodore M. and Hudson, Hoyt H. trans. (New York: Harper and Row 1960)Google Scholar, abbreviated as ‘R'; Critique of Practical Reason, Beck, Lewis White trans. (New York: MacMillan 1993)Google Scholar, abbreviated as ‘Kp V’; The Metaphysics of Morals, Gregor, Mary ed. and trans. (Cambridge: Cambridge University Press 1996)Google Scholar, abbreviated as ‘MdS’; Groundwork of the Metaphysics of Morals, Gregor, Mary ed. and trans. (Cambridge: Cambridge University Press 1997),Google Scholar abbreviated as ‘G’; Lectures on Ethics, Heath, Peter trans. (Cambridge: Cambridge University Press 1997)Google Scholar.

2 Christine, Korsgaard ‘Kant's Formula of Universal Law,’ in Korsgaard, Christine Creating the Kingdom of Ends (Cambridge: Cambridge University Press 1996Google Scholar

3 Wood, Allen Kant's Ethical Thought (Cambridge: Cambridge University Press 1999), 91CrossRefGoogle Scholar

4 By contrast, Kant's two examples of violation of duties to others, non-beneficence and false promising, do require interpersonal universalization.

5 I insert ‘primarily’ here because one could argue that we must develop our talents so that we may help others. But then it is only a duty to oneself indirectly: it is a duty to others that involves oneself as the means.

6 Paton, HJ. The Categorical Imperative (London: Hutchinson 1947), 135Google Scholar

7 At the APA and in personal communication, Tom Hill has suggested a different way in which Kant may be utilizing universalization in his FLN argument against suicide. The idea would be that a teleological law, such as the teleological law that self-love has the purpose of self-preservation, can only be conceived of as a teleological law if everyone or most tended to follow that law. Accordingly, if upon universalizing my maxim it becomes universal law that everyone act contrary to the putative teleological law, e.g., such that everyone commit self-destruction from self-love, this would contradict the central idea of the teleological law, namely, that self-love has the purpose of self-preservation.

8 Note that imagining your maxim to hold at all times for the universalization procedure does not mean that you actually must will it at all times (thus violating the latitude of Kant's imperfect duties). It is, rather, analogous to interpersonal universalization: you imagine a hypothetical world in which everyone follows your maxim as if it were a law of nature, but, of course, this does mean that everyone actually does follow our maxims.

9 As an anonymous referee pointed out, if Kant in the Lectures literally means we ought to be obsessively regular in our actions (such as Kant's famous walks are reputed to be), this is a cartoonish and implausible view of morality. As I will explain below, however, I take Kant to mean merely that the moral law requires us to be consistent across time, such that temporal location alone is not a justification for altering one's behavior.

10 There are a few exceptions to this general trend in the literature. John Rawls explicitly recognizes the temporal dimension of Kant's universalizability procedure, Lectures on the History of Moral Philosophy, Herman, Barbara ed. (Cambridge, MA: Harvard University Press 2000), 171Google Scholar. His ‘perpetuity condition’ holds that the hypothetical world with the universalized maxim ‘is as if it always has existed, exists now, and always will exist.’ Two other recent discussions of Kant's universalization procedure that include a temporal component are Harbison, Warren ‘Self-Improvement, Beneficence, and the Law of Nature Formula,’ Kant-Studien 91 (2000) 17–24CrossRefGoogle Scholar, at 18, and McNair, Ted ‘Universal Necessity and Contradictions in Conception,’ Kant-Studien 91 (2000) 25–43Google Scholar, at 26. However, again, this attention to temporal universalization seems to be the exception rather than the rule.

11 Herman, Barbara The Practice of Moral Judgment (Cambridge, MA: Harvard University Press 1993)Google Scholar has a worry that closely mirrors this one, and which, in the end, leads her to reject Korsgaard's interpretation of the universalization procedure. I address Herman's worry in section V.

12 For more on this point, see O'Neill, Onora ‘Consistency in Action,’ in O'Neill, Onora Constructions of Reason: Explorations of Kant's Practical Ethics (Cambridge: Cambridge University Press 1989)Google Scholar.

13 Here I adopt O'Neill's formula of what she calls ‘complete maxims’ (writing as Neil, Onora Acting on Principle: An Essay on Kantian Ethics [New York: Columbia University Press 1975], 35–8)Google Scholar. In these kinds of maxims, the purpose of the action is explicitly included. I utilize this way of formulating maxims here in order to draw out the difference a purpose can make in determining the moral permissibility of a maxim (which I do in order to show that it is other ingredients in one's maxim, such as purpose or circumstances, rather than temporal location, that determines the moral permissibility of one's maxim). Later I will use ‘incomplete maxims,’ which make no explicit reference to purpose, in order to simplify discussion.

14 Rawls, John A Theory of Justice rev. ed. (Cambridge, MA: Harvard University Press 1999), 259Google Scholar. Rawls justifies this approach on his contractarianism. He draws the general point from Sidgwick, and there are several contemporary consequentialists (e.g., Hurka, Thomas Perfectionism [New York: Oxford University Press 1993])Google Scholar who incorporate time-neutrality into their theories.

15 While it may be irrational to will differently at different times, a potential problem arises in that this is not always a mark of immorality. I take it, incidentally, that we would say the same thing about persons: maybe it is irrational to treat two people (say, strangers) differently, but surely this isn't always immoral. So if this is a problem for temporal universalization, it is also a problem for interpersonal universalization. The solution, then, is to find some way of distinguishing between morally relevant and non-morally relevant instances of irrationality. One such answer, which I want to adopt for the purposes of this paper, is that inconsistency is morally relevant only when it bears on treatment of humanity as an end in itself. This answer, while brief and rough, is that given by Timmons, Mark ‘Decision Procedures, Moral Criteria, and the Problem of Relevant Descriptions in Kant's Ethics,’ Jahrbuch fur Recht und Ethik 5 (1997) 389–417Google Scholar (see Herman, ch. 10, for a similar view). The idea is that the different formulations of the Categorical Imperative have different roles, and the role of the Formula of Humanity is (in part) to identify what is and is not morally relevant. Obviously, I cannot defend such an Interpretation here, but it seems a viable position to me.

16 I take it that when discussing Kant, we can grant everyone would be cognizant of this rational requirement, just as we are pre-theoretically aware of our moral duties (G, 403-4).

17 Kant will still have the problem that when we properly contextualize for one's psycho-motivational State, one can temporally universalize ‘to commit suicide when life is horrible’ consistently with the presence of a maxim ‘to preserve myself from self-love when life is good.’ However, this is a problem with the duty against suicide, not a problem about Kant falling to use universalization in the argument. Again, universalizing is consistent with proper contextualization of maxims. Thanks to Nelson Potter for discussion on this point.

18 Here, Kant's concept of duty (or Obligation) is ‘constraint to an end adopted reluctantly. Hence it is self-contradictory to say that he is under Obligation to promote his own happiness with all his powers.’

19 The clause ‘at some time or other’ is needed in order to fit this reading with Kant's claim that the duty to develop one's talents is an imperfect duty, allowing for latitude in how it is satisfied. (See n. 8.)

20 Wood (109) takes this to indicate that Kant must Supplement the universalizability tests with information derived from the other formulae of the Categorical Imperative.

21 A qualifier to the idea of temporal universalization is needed. As I understand it, just as we are to universalize to all rational beings, we are to universalize to all parts of our lives when we have rational capacities. Thus, just as Kant does not (directly) include non-rational animals in interpersonal universalization, so we should not include the non-rational parts of our lives.

22 I take it that this is the formulation that Wood (90) has in mind when he speaks of devoting one's ‘life entirely’ to pleasure and idleness. It also seems to be Feldman's, Fred (Introdudory Ethics [Englewood Cliffs, NJ: Prentice Hall 1978], 99)Google Scholar understanding of maxims. See Atwell, John E. Ends and Principles in Kant's Moral Thought (Dordrecht: Martinus|Nijhoff 1986), 48–9CrossRefGoogle Scholar, for further critique of this view.

23 There are maxims that hold for our character dispositions as well, for Kant. And he indicates (R, 25) that we have adopted a maxim for life about whether to privilege the moral law over inclination or vice versa. There is no principled reason why there cannot be these several kinds of maxim, but in the Groundwork argument that occupies us currently, he clearly wants to focus on maxims for individual actions.

24 For a spelling out of these problems, and the corollary that Kant's Tugendlehre arguments for the duties are more defensible, see Potter, Nelson ‘What is Wrong with Kant's Four Examples,’ Journal of Philosophical Research 18 (1993) 213–29, at 216-21)CrossRefGoogle Scholar.

25 Harrison, Jonathan ‘Kant's Examples of the First Formation of the Categorical Imperative,’ in Kant: A Collection of Critical Essays, Wolff, Robert Paul ed. (Notre Dame: Notre Dame University Press 1967), 235Google Scholar

26 Kemp, J. ‘Kant's Examples of the Categorical Imperative,’ in Kant: A Collection of Critical Essays, Wolff, Robert Paul ed. (Notre Dame: Notre Dame University Press 1967), 252–3Google Scholar

27 See Mark, Timmons ‘Contradictions and the Categorical Imperative,’ Archiv fur Geschishte der Philosophie 66 (1984) 294–312Google Scholar; Galvin, Richard F. ‘Ethical Formalism: The Contradiction in Conception Test,’ History of Philosophy Quarterly 8 (1991) 387–408Google Scholar; Barbara Herman; and Christine Korsgaard (among others), for different interpretations, discussions, and mapping of the contemporary philosophical territory on this issue.

28 Since what I will have to say is intended as a general response (and not just a response to the ‘tennis’ case), it is worth noting that all of the coordination maxims Herman uses to derive false negatives for the practical contradiction Interpretation explicitly rely on some particular temporal location.

29 Baron, Marcia W. ‘Kantian Ethics,’ in Baron, Marcia W. Pettit, Philip and Slote, Michael eds., Three Methods of Ethics (Maiden, MA: Blackwell 1997), 72–3Google Scholar, also suggests replacing ‘Sunday at 10’ with ‘when the courts are open.’ Rather than justifying this switch on the basis of temporal universalizability, Baron grounds it on the idea that it ‘seems more truly to be the agent's maxim.’ As Baron remarks in her evaluation of this move, however, there may be some agents for whom ‘Sunday at 10’ is integral to their maxim, in which case the switch becomes question-begging. Regarding the interpretation presented here, I discuss the question-begging worry below.

30 But see her comment (on 140) that ‘The practical interpretation draws strong intuitive support when it seems to pick out free riding on the activities of others as the morally salient fallure of impermissible actions.’

31 There are other potential weaknesses facing the practical interpretation, which must be addressed case-by-case. One set of putative problem cases concern a subset of ‘natural,’ rather than ‘conventional,’ actions (i.e., actions that rely on no Convention in order to succeed). Herman's case (118) of convenience killing and Korsgaard's case of revenge killing (100) illustrate the point. Here, one may universalize killing for revenge (or convenience) without frustrating one's own killing. It might appear, on first glance, that temporal universalizability would be of help; one might think that if I must temporally universalize my killing for revenge (or convenience), then if someone in the past had cause to kill me out of revenge (or for convenience), they would have done so, making my current killing impossible. While this kind of argument may certainly work for many people, it would not apply to a person who was never the target of vengeance or who never got in the way of another's plans (and so would not inconvenience the agent). While such persons are perhaps rare, they are possible, and as such revenge (or convenience) killing seems to be legitimate for them on the practical contradiction interpretation. Accordingly, that interpretation seems to still generate false positives, even if temporal universalizability renders it immune to false negatives. Herman's (eh. 6) own generation of a deliberative presumption against killing seems strained, insofar as it generates a contradiction in the will, rather than a contradiction in conception; as such, it would appear that on Standard Kantian readings, agents have only an imperfect duty not to kill, allowing for latitude in how, when, and to what extent they will comply with the duty not to kill. (However, the ‘how, when, and to what extent’ reading, while Standard, is controversial. See Cummiskey, David Kantian Consequentialism [New York: Oxford University Press 1996]CrossRefGoogle Scholar, eh. 6, for an opposing view.) See Calhoun, Cheshire ‘Kant and Compliance with Conventionalized Injustice,’ The Southern Journal of Philosophy 32 (1994) 135–59CrossRefGoogle Scholar, for a different solution to the problem of natural actions. This solution rests on conceiving the victims of either natural or conventional actions as ‘fully free, equal, and rational,’ and therefore capable of effective resistance to ‘the conditions created by universalizing a maxim’ (148). As such, convenience or revenge killing becomes impossible, as do other problematic practices, such as some forms of slavery.

32 An earlier version of part of this paper was presented at the 2001 meeting of the APA Pacific Division. Thanks to the participants, as well as Lori Alward, for many helpful comments. Helpful comments were also supplied by anonymous referees for the Canadian Journal of Philosophy. Finally, thanks to Mark Timmons, Nelson Potter, and Tom Hill for their constructive suggestions.