his own candidates for the supreme principle of morality face elimination
on the grounds that they manifestly fail to meet it (section 8.2). At least
in the Groundwork, Kant suggests that a principle is the supreme norm for
the moral evaluation of action only if just those actions done in accordance
with the principle can have moral worth. But as I have contended, Kant is
rationally compelled to hold that actions done in accordance with an in-
deď¬nite number of different principles can have moral worth, as long as
the actions are done from duty. (An agent acts from duty just in case her
[in itself sufď¬cient] incentive for acting stems from the notion that a prin-
ciple, represented by her as a law, requires the action; she acts against the
background of conscientious reď¬‚ection; and she does her best to realize
her actionâ™s end.) Kant needs to modify criterion iv so that it demands that
the supreme principle of morality be such that actionsâ™ moral permissibility
and requiredness, but not their moral worth, be deď¬ned in terms of this
Let me now turn to the criteria for the supreme principle of morality that
go beyond those contained in Kantâ™s basic concept of this principle. Since
the end of Chapter 6, we have been operating with four further criteria.
Conclusion: Kantâ™s Candidates 189
The supreme principle of morality must be such that: (v) every case of
willing to conform to it because the principle requires it has moral worth;
(vi) the moral worth of willing to conform to the principle because the
principle requires it stems from its motive, not from its effects; (vii) an agentâ™s
representing the principle as a law â“ that is, a universally and unconditionally
binding principle â“ provides him with sufď¬cient incentive to conform to it;
and, ď¬nally, (viii) a plausible set of duties (relative to ordinary rational moral
cognition) can be derived from the principle. I will begin the discussion with
criterion viii and work my way to v.
Much of this chapter has been devoted to the question of whether Kantâ™s
candidates for the supreme principle of morality should be eliminated on
the basis of a manifest failure to fulď¬ll viii. This is, I think, an important
question, since I believe that viii is a criterion that Kant is correct to em-
brace. Given that Kant bases his Groundwork I derivation on reď¬‚ective moral
common sense, it would, I think, be unwarranted for him to endorse a can-
didate for the supreme principle of morality that generated prescriptions
totally unacceptable to common sense (and that thus violated viii). After all,
what grounds would Kant have for trusting ordinary moral consciousness
in its endorsement of several criteria for the supreme principle of morality,
yet ignoring its view of which duties would stem from a plausible candidate
for this principle?
Kant suggests two arguments in support of vii (see section 5.7). One
of the arguments turns on a dictum that Kant should abandon, namely
(a particular understanding of) âought implies canâ (see section 8.2). But
the other does not turn on this dictum. According to this argument, denying
vii would amount to holding that the supreme principle of morality must
be such that oneâ™s expectation of the effects of conforming to it necessarily
constitutes (at least part of) his incentive for conforming to it. Now consider
an agent who denied vii and embraced some principle as a viable candidate
for the supreme principle of morality. He would be committed to the view
that the worth of his conforming to the principle necessarily derived (at least
in part) from its results. For if he thought that conforming to the principle
was valuable in itself, then he would not hold that he necessarily needs to
look to the actionâ™s results to ď¬nd a sufď¬cient incentive to do so. But if
the agent ties inextricably the worth of his conforming to a principle to its
results, then he is rationally compelled to deny that his conforming to it
can have intrinsic worth. He is rationally compelled to deny a basic tenet of
reď¬‚ective moral common sense. Therefore, he must agree that we can hold
a principle to be the supreme principle of morality only if we can maintain
that our representing it as a law governing our actions gives us a sufď¬cient
incentive to conform to it.
It would be remiss not to acknowledge this argument to be controversial.
(Let me reiterate that the argument is not one Kant explicitly gives but rather
one that I believe he suggests.) One possible difď¬culty with it is the following.
Search for the Supreme Principle of Morality
The argument turns on the notion that, rationally speaking, an agent cannot
hold conforming to a certain principle to be intrinsically valuable, yet at the
same time hold that she needs necessarily to rely on the prospect of the
actionsâ™ effects in order to have sufď¬cient incentive to conform to it. But
this notion is disputable. Recall the principle EU: âAlways perform a right
action, one that you expect will yield as great a sum total of well-being as
would any alternative action available to you.â Suppose that according to
a particular agent, her conforming to EU because she expects that doing
so will maximize aggregate well-being is intrinsically valuable. The agent
holds further that in order to have sufď¬cient incentive to conform to EU,
she not only does but must rely on the prospect of maximizing aggregate
well being.40 It is not obvious that there is anything irrational in this agentâ™s
views. At the very least, the argument in question would need to be bolstered
in order to deal with cases such as this. In any event, if I am correct, Kant
does not have to rely on vii in order to eliminate rivals to his candidates for
the supreme principle of morality.
But he does need to rely on criteria v and vi. It almost goes without saying
that, in my view, v and vi are very plausible criteria for the supreme principle
of morality. I have discussed them in detail. So here I would like merely
to emphasize that v represents a modification of a criterion Kant advocates
in the Groundwork. According to Kant, the supreme principle of morality
must be such that all and only actions conforming to this principle because
the principle requires it (i.e., all and only actions done from duty) have
moral worth. I have advocated two signiď¬cant changes to this criterion.
First, Kant must acknowledge that some actions contrary to duty can be
from duty and can thus have moral worth (section 6.6). He needs to hold,
as criterion v contends, that the supreme principle of morality must be
such that every case of willing from duty to conform to it has moral worth.
Second, Kant offers no good reason for holding that only actions done
from duty have moral worth. He does not undermine the view that actions
done with an overriding commitment to morality but from other motives
(e.g., sympathy) have such worth (6.10). (Recall that on the account I have
sketched, an agent has an overriding commitment to morality just in case
he acts against the background of conscientious reď¬‚ection, and if after such
reď¬‚ection, he determines that an action is contrary to what he takes to be
morally required, he will for this reason refrain from performing it.) In my
view, Kant fails to establish that the supreme principle of morality must be
such that only instances of willing to conform to it because the principle
requires it have moral worth. Yet this failure has no signiď¬cant impact on
his derivations. With the help of other criteria, especially v and vi, he can
construct powerful arguments against many rivals to his candidates for the
supreme principle of morality.
It is one thing for Kant to show that rivals founder as candidates for
the supreme principle of morality; it is quite another for him to establish
Conclusion: Kantâ™s Candidates 191
that his formulas remain aď¬‚oat. Fulď¬lling criterion viii in particular poses a
formidable challenge. I have offered reasons for thinking that the Formula
of Universal Law fails to generate prescriptions acceptable to reď¬‚ective moral
common sense. And I believe it remains an open question whether the For-
mula of Humanity succeeds. It would be irresponsibly optimistic to claim
that Kant has demonstrated that if there is a supreme principle of morality,
then it is the Formula of Humanity.
A down-to-earth approach to the Kantian project in ethics emerges from
this book. To advance toward the ideal of showing that if there is a supreme
principle of morality, then it is some particular principle, it does not sufď¬ce
to rely merely on abstract premises such as that we are transcendentally free
rational agents or that a categorical imperative requires an unconditionally
good âground.â We need to enter concrete controversies regarding which
duties the principle would generate and whether these duties would be
acceptable to reď¬‚ective moral common sense. In searching for the supreme
principle of morality, we need to follow the twists and turns of everyday
moral experience. There is no royal road to a successful derivation.
Introduction: Derivation, Deduction, and the Supreme Principle of Morality
1. By Kantâ™s âcritical writingsâ in ethics, I mean simply the works in the ď¬eld he
published after the appearance of the Critique of Pure Reason.
2. In Groundwork II, Kant says that the âthe categorical imperative,â the principle
he takes to be the supreme principle of morality, is âthe canon of moral appraisal
of action in generalâ (GMS 424). On the next page (GMS 425), Kant says: âwe
have . . . set forth distinctively and as determined for every use the content of the
categorical imperative, which must contain the principle of all duty (if there is
such a thing at all).â
3. As Schopenhauer notes with irritation, Kant never seems to tire of reminding
us of this feature. See Arthur Schopenhauer, On the Basis of Morality, tr. E. F. J.
Payne (1841; Providence: Berghahn Books, 1995), 63.
4. Evidence that Kant holds this can be found in his Metaphysics of Morals discussion
of ends that are also duties (e.g., MS 386). Kant argues that an agent cannot
have an obligation to promote the end of his own happiness, since each agent
unavoidably has this end.
5. I discuss Kantâ™s notions of the will and its determining grounds in Chapter 1.
6. âCommon rational moral cognitionâ or, translated slightly differently, âordi-
nary rational knowledge of moralsâ is Kantâ™s starting point in the Groundwork.
He entitles Section I âTransition from common rational to philosophic moral
cognitionâ (GMS 393).
7. In the cited passage, Kant moves seemingly without argument from the no-
tion that a moral law must be absolutely necessary to the notion that it must
have wide universal validity. I think this move is problematic, as I explain in
8. Allison appears to employ this usage of âderivation.â See Henry E. Allison,
Idealism and Freedom (Cambridge: Cambridge University Press, 1996), 143â“144.
9. See Henry E. Allison, Kantâ™s Theory of Freedom (Cambridge: Cambridge University
Press, 1990), 214.
10. I have adapted this notion of moral particularism from Oâ™Neillâ™s helpful dis-
cussion. See Onora Oâ™Neill, Towards Justice and Virtue (Cambridge: Cambridge
University Press, 1996), 11â“13.
Notes to pp. 5â“10
11. For some interesting arguments against moral particularism, see ibid., especially
12. Rudiger Bittner, What Reason Demands (Cambridge: Cambridge University Press,
13. See also KpV 93 on this point in particular.
14. In my view, Groundwork III is one of the most enigmatic texts Kant published.
The argument is hard to follow, and philosophers differ signiď¬cantly in how they
interpret it. For reconstruction and criticism of Kantâ™s argument (or crucial as-
pects of it), see Allison, Kantâ™s Theory of Freedom, 214â“229; Karl Ameriks, âKantâ™s
Deduction of Freedom and Morality,â Journal of the History of Philosophy 19 (1981):
45â“65; RÂ¨ udiger Bittner, âWer frei ist, ist gebunden,â in Philosophiegeschichte und
Logische Analyse, ed. U. Meixner and A. Newen (Paderborn: Mentis, 2000), 209â“
221; Dieter Henrich, âDie Deduktion des Sittengesetzes,â in Denken im Schatten
des Nihilismus, ed. Alexander Schwan (Darmstadt: Wissentschaftliche Buchge-
sellschaft, 1975), 55â“112; Allen W. Wood, Kantâ™s Ethical Thought (Cambridge:
Cambridge University Press, 1999), 171â“182.
15. For Auneâ™s discussion of a gap in the Groundwork II derivation, see Bruce Aune,
Kantâ™s Theory of Morals (Princeton: Princeton University Press, 1979), 42â“43.
16. Ibid., 28â“29.
17. Ibid., 30.
18. Ibid., 31.
19. As Aune acknowledges (29), this statement of the Categorical Imperative, which
stems from Groundwork II (GMS 421), differs a bit from the one Kant actually
offers in Groundwork I. The statement Kant gives there is this: âI ought never to
act except in such a way that I could also will that my maxim should become a universal
law.â The differences between the two statements are not important here.
20. Aune, Kantâ™s Theory of Morals, 30; see also 32 and 86â“87.
21. Ibid., 34.
22. See, for example, Allison, Idealism and Freedom, 144, 150; David Cummiskey,
Kantian Consequentialism (Oxford: Oxford University Press, 1996), 57; Thomas
E. Hill Jr., âThe Rationality of Moral Conduct,â in Dignity and Practical Reason
in Kantâ™s Moral Theory (Ithaca: Cornell University Press, 1992), 121â“122; Allen
W. Wood, Hegelâ™s Ethical Thought (Cambridge: Cambridge University Press, 1990),
164â“166; and Wood, Kantâ™s Ethical Thought, 47â“49, 81â“82.
23. Wood, Kantâ™s Ethical Thought, 81.
24. See Allison, Idealism and Freedom, 146, for a slightly more complex version of this
25. Ibid., 145.
26. I discuss the moral impermissibility of acting on this maxim of false promising
in Chapter 8.
27. I take it that this is the sense (or at least one of the senses) in which Kant em-
ploys the notion of a categorical imperative in Groundwork II before he ofď¬cially
introduces âthe categorical imperative.â See GMS 416. A principle that was a
categorical imperative in this sense, that is, an unconditionally and universally
binding principle, would presumably not manifest itself as an imperative to beings
incapable of violating it (e.g., God and angels).
28. In taking this to be a statement of the Formula of the Kingdom of Ends, I
am following Oâ™Neill. See Onora Oâ™Neill, Constructions of Reason (Cambridge:
Cambridge University Press, 1989), 127.
Notes to pp. 11â“18 195
29. I will also leave aside another formula Kant introduces, namely what has been
referred to as the Formula of Autonomy: â[C]hoose only in such a way that the
maxims of your choice are also included as universal law in the same volitionâ
(GMS 440). See Oâ™Neill, Constructions of Reason, 53. According to Allen Wood,
Kant presents the Formula of the Kingdom of Ends as a âmore intuitive versionâ
of the Formula of Autonomy. See Wood, Kantâ™s Ethical Thought, 163â“166.
30. Of course, Kant holds at the very least that both the Formula of Universal Law
and the Formula of Humanity each generate the duties (roughly) not to commit
suicide, not to make false promises, to develop oneâ™s natural talents, and to help
others in need. See GMS 421â“424, 429â“430.
Chapter 1: Fundamental Concepts in Kantâ™s Theory of Agency
1. MS 211â“213 gives one a taste of just how challenging Kantâ™s discussions of agency
2. âMaxime ist das subjective Prinzip zu handelnâ (GMS 421, note). See also GMS
3. The discussion of maxims in this section has been heavily inď¬‚uenced by Bittner.
Bittner disagrees, however, with the view I set out that, when fully described,
maxims incorporate descriptions of an agentâ™s end and incentive in acting.
See Rudiger Bittner, Doing Things for Reasons (Oxford: Oxford University Press,
2001), chap. 3.
4. This maxim is, of course, a slight variant of one that Kant employs in the Ground-
work. See, for example, GMS 422.
5. If nobody held M, then it would be a possible or potential maxim.
6. I am assuming here that Kant is justiď¬ed in his view that acting on M is forbidden
by the Categorical Imperative. See GMS 421â“422.
7. See, for example, Onora Oâ™Neill, Constructions of Reason (Cambridge:
Cambridge University Press, 1989), 151.
8. Of course, on Kantâ™s view all morally obligatory actions also count as morally
permissible ones. That Kant held all of an agentâ™s actions to be either morally
permissible or impermissible becomes clear in the Religion. In the main text,
Kant says: âIt is . . . of great consequence to ethics in general to avoid admitting,
so long as it is possible, of anything morally intermediate, whether in actions
(adiaphora) or in human charactersâ (Rel 22, English ed. 18). This suggests
that Kant does not want to admit there to be actions that are neither morally
permissible nor impermissible. And, as a matter of fact, central claims he makes
in the Religion (and elsewhere in his practical philosophy) prevent him from
admitting it. Kant holds that all human action is free and that free action is
action that does not result merely from natural laws (Rel. 41, English ed. 36).
But if there were human actions that were neither morally permissible nor
impermissible, claims Kant, then they would result merely from natural laws
(Rel 23, note; English ed. 18, note): they would not be free. Therefore, according
to him, there are no human actions that are neither morally permissible nor
9. On this point, see Nelson Potter, âMaxims in Kantâ™s Moral Philosophy,â
Philosophia 23 (1994): 62, 71.
10. The term âincentiveâ translates the German term Triebfeder. Translated literally
Triebfeder means something like âpush spring.â (A Triebfeder in a clock would be
Notes to p. 19
the spring that makes it run.) Of course, in contemporary English âincentiveâ
can mean something like object to be gained, as when we say that the childâ™s
incentive to clean her room is a trip to the amusement park.
11. At Rel 35 (English ed. 30), Kant says that âin the absence of all incentives the
will [WillkÂ¨ r] cannot be determined.â
12. For discussion of this point, see Potter, âMaxims in Kantâ™s Moral Philosophy,â
13. Suppose that, after much reď¬‚ection, an agent comes to believe that her incentive
in performing a particular action was the notion that it was morally required.
In the Groundwork, Kant claims that it is impossible for an agent to know that her
incentive in performing the action was the notion that it was morally required
(was âdutyâ), rather than some âcovert impulse of self-loveâ (GMS 407). When
coupled with the reading of maxims I have suggested, this claim implies that, in
cases where it appears to an agent that she has acted from duty, the agent cannot
know which maxim she acted on. For if the maxim were fully speciď¬ed, it would
include a description of her incentive. This implication would be problematic
if it turned out that in some cases, since an agent could not fully specify the
maxim of her action, she would not be able by using the Categorical Imperative
to determine whether her action was morally permissible. For if such cases
might arise, then it is hard to see how Kant could maintain as he does that the
Categorical Imperative is the canon of the moral estimation of all of our action
(GMS 423). Perhaps Kant holds that an agent is always capable of specifying
each of the elements in her maxim besides her incentive and that, through
specifying these elements, she would always get a sufď¬cient grasp on her maxim
to gain a reliable indication of its moral permissibility from the Categorical
14. As I indicated, the conclusion that acting on the maxim in question would be
impermissible is based on a traditional reading of the Categorical Imperative.
For a different reading of it, according to which acting on this maxim would not
turn out to be morally impermissible, see Thomas W. Pogge âThe Categorical
Imperative,â in Kantâ™s âGroundwork of the Metaphysics of Moralsâ: Critical Essays, ed.
Paul Guyer (Lanham, Md.: Rowman & Littleď¬eld, 1998), 189â“196. As I argue
in Chapter 8, this reading has problems of its own.
15. This account of what differentiates maxims from other rules of the same form
has been heavily inď¬‚uenced by Oâ™Neill who says: âThe maxim of an act is the
principle that governs the selection of ancillary principles of action that express
or implement the maxim in a way that is adjusted to the agentâ™s (perceived)
circumstances.â Oâ™Neill, Constructions of Reason, 129; see also 151â“152. Another
proposal regarding what distinguishes maxims from other rules of the same
form has been made by Bittner. He suggests that maxims are ârules of lifeâ
(Lebensregeln), whereas the others are not. (See RÂ¨ diger Bittner, âMaximen,â in
Akten des 4. Internationalen Kant-Kongresses, ed. G. Funke and J. Kopper [Berlin:
De Gruyter, 1974], 489.) An agentâ™s maxims express the kind of life he wants
to lead, the course that he intends his life as a whole to take. If Kant did indeed
hold maxims to be Lebensregeln, it is clear that he would deny status as a maxim
to the rule âFrom self-love, every Monday at 3 p.m. I take live karate lessons in
order to improve my endurance and ď¬‚exibility.â For surely this rule does not
express an agentâ™s conception of the direction he wants his life as a whole to
Notes to p. 20 197
take. Yet the more general rule of keeping oneself in shape by exercising during
oneâ™s free time does presumably express such a conception. However, there are,
I believe, fairly good grounds for rejecting the view that for Kant all maxims are
Lebensregeln. Consider, once again, the practical rule of false promising that Kant
discusses in the Groundwork: âWhen I believe myself to be in need of money, I
shall borrow money and promise to repay it, even though I know that this will
never happen.â This practical rule is not a Lebensregel. Granted, if we knew that an
agent acted on it, we would have a clue as to what his Lebensregeln might be like.
We might, for example, suspect that he has adopted one akin to: âWhenever
my happiness is threatened, if need be I will lie in order to secure it.â But
the false-promising rule does not, in itself, express the direction that an agent
who acts on it wants his life as a whole to take. (Allison makes this point. See
Henry E. Allison, Kantâ™s Theory of Freedom [Cambridge: Cambridge University
Press, 1990], 92â“93.) Kant refers explicitly to this rule as a maxim (GMS 422).
Since it is not a Lebensregel, it appears that Kant does not think of all maxims as
16. At KpV 32 Kant identiď¬es the will (Wille) of rational beings with âthe ability to
determine their causality by the representation of rules.â See also KpV 45, 55,
58â“59, 125. For examples of such usage of Wille in the Groundwork, see GMS
412, 427. Laberge offers an excellent discussion of Kantâ™s deď¬nition of the will
at GMS 412. See Pierre Laberge, âLa dÂ´ ď¬nition de la volontÂ´ comme facultÂ´
e e e
dâ™agir selon la reprÂ´ sentation des lois,â in Grundlegung zur Metaphysik der Sitten;
Ein kooperativer Kommentar, ed. Otfried HÂ¨ ffe (Frankfurt am Main: Klostermann,
17. In using the terms âexecutive Willeâ and âlegislative Wille,â I am following
(roughly) the suggestion of Beck. See Lewis White Beck, A Commentary on Kantâ™s
âCritique of Practical Reasonâ (Chicago: University of Chicago Press, 1960), 202.
18. For example, Greene and Hudson use this translation in their edition of the
Religion within the Limits of Reason Alone.
19. Some translators render WillkÂ¨ r as the â[power of] choice,â thus implying that
when one chooses to act, one exercises WillkÂ¨ r. (See, e.g., Gregorâ™s translation
of the Metaphysics of Morals, 650.) This rendering is potentially misleading. For
Kant, exercising WillkÂ¨ r involves acting on choice: it involves choosing to realize
an object and acting in the sense of trying to realize it. Some readers (myself
included) conceive of choosing to realize an object as distinct from trying to real-
ize it: one might do the former (e.g., choose to see a certain play) without doing
the latter (e.g., making any effort to see it). That Kant takes exercising WillkÂ¨ r to
involve trying to realize a chosen object is, I think, suggested in his (admittedly
dense and difď¬cult) deď¬nition of WillkÂ¨ r at MS 213. For a detailed discussion of
this deď¬nition, see Samuel J. Kerstein, âAction, Hedonism, and Practical Law:
An Essay on Kantâ (Ph.D. diss., Columbia University, 1995), 30â“34. Now I have
suggested that, for our purposes, it is safe to consider exercising WillkÂ¨ r to be
the same thing as exercising executive Wille (i.e., acting on self-given rules). But
one might think that it is possible for an agent to exercise WillkÂ¨ r (i.e., to act
on choice), without acting on any rule. However, as we have already seen, for
Kant all of our (rational agentsâ™) acting is acting on some rule, that is, some
maxim. Therefore, any time an agent exercises her WillkÂ¨ r, she also exercises
her executive Wille.
Notes to pp. 21â“25
20. I am not here invoking the distinction Kant draws at GMS 427 between an
incentive and a motive (Bewegungsgrund ). Kant there writes: âThe subjective
ground of desire is an incentive ; the objective ground of volition is a motive ;
hence the distinction between subjective ends, which rest on incentives, and
objective ends, which depend on motives, which hold for every rational being.â
Kant himself does not appear to maintain this distinction. See, for example,
KpV 72, where he writes at length concerning the moral law as incentive.
21. Here I am following Allison. See Henry E. Allison, Idealism and Freedom
(Cambridge: Cambridge University Press, 1996), 131. I discuss this notion fur-
ther in section 2.2.
22. See Allison, Kantâ™s Theory of Freedom, 204. Also see Barbara Herman, The Practice
of Moral Judgment (Cambridge, Mass.: Harvard University Press, 1993), 217â“218.
23. Material in sections 1.6â“8 stems from Samuel J. Kerstein, âKantâ™s (Not So Radi-
cal) Hedonism,â in Kant und die Berliner Auf klÂ¨ rung. Akten des IX. Internationalen
Kant-Kongresses, vol. 3, ed. V. Gerhardt, R.-P. Horstmann, and R. Schumacher
(Berlin: Walter de Gruyter, 2001), 245â“253.
24. See, for example, T. H. Green, Lectures on the Philosophy of Kant, in Works of Thomas
Hill Green, vol. 2, ed. R. L. Nettleship (London: Longmans, Green, 1900), 141;
Terence Irwin, âKantâ™s Criticisms of Eudaemonism,â in Aristotle, Kant, and the
Stoics, ed. S. Engstrom and J. Whiting (Cambridge: Cambridge University Press,
1997), 74; A. Phillips Grifď¬ths, âKantâ™s Psychological Hedonism,â Philosophy 66
(1991): 211; Bernard Williams, Ethics and the Limits of Philosophy (Cambridge:
Cambridge University Press, 1985), 15, 64.
25. Grifď¬ths, âKantâ™s Psychological Hedonism,â 212.
26. See Andrews Reath, âHedonism, Heteronomy and Kantâ™s Principle of Happi-
ness,â Paciď¬c Philosophical Quarterly 70 (1989): 42â“72. Christine Korsgaard seems
to embrace Reathâ™s reading. See Christine M. Korsgaard Creating the Kingdom of
Ends (Cambridge: Cambridge University Press, 1996), 56.
27. See Reath, âHedonism, Heteronomy and Kantâ™s Principle of Happiness,â 46â“49.
28. Perhaps Allison suggests the alternative interpretation in a discussion of Kantâ™s
âphilanthropistâ (or âfriend of humanityâ). See Allison, Kantâ™s Theory of Free-
dom, 103. Allison says: âThus, whereas helping others in need provides [the
philanthropist] with satisfaction and he would not act in that way unless it did
so, the end he has in mind is nevertheless the improvement of the lot of oth-
ers and not the satisfaction of his own needs.â I am unsure whether Allison
is indeed urging the alternative interpretation, since the context of his re-
mark makes it doubtful whether by âsatisfactionâ he means pleasure, as the
alternative would require. But if he is suggesting the alternative interpreta-
tion, then my aim here could be construed as developing his suggestion by (a)
distinguishing it from Reathâ™s proposal and (b) shedding light on its textual
29. Beck treats the concept of the capacity of desire in a way that is typical for
contemporary scholars. In his classic commentary on the second Critique, he
employs the concept, but does not focus on explicating it. See, for example,
Beck, A Commentary on Kantâ™s âCritique of Practical Reason,â 94â“97.
30. When an agent exercises her capacity of desire, she tries to realize an object.
Kant, it is worth noting, does not exclude mental objects from the set of those
Notes to pp. 26â“28 199
an agent might try to realize in exercising this capacity. For example, guided by
my idea of remembering the name of my ď¬rst mathematics teacher, I might try
to remember it.
31. Of course, an agentâ™s pathological interest in something is not necessarily an inter-
est that stems from illness or an interest in something unhealthy. In this context,
âpathologicalâ does not connote disease. It means rather sensory, that is, having
to do with sensation.
32. In Kantâ™s vocabulary, the term âagreeablenessâ (Annehmlichkeit) designates a
kind of sensation. See, for example, KpV 22, where Kant speaks of the âsensation
of agreeableness.â To say that something is agreeable to a person is to say that
it enables him to experience the sensation of agreeableness. Furthermore, on
Kantâ™s view, to experience the sensation of agreeableness is to experience plea-
sure: agreeableness is a kind of pleasure. For the most part, Kant seems to use
the terms âpleasureâ (Lust) and âagreeablenessâ interchangeably. See, for ex-
ample, KpV 23. It appears, however, that âpleasureâ has a wider extension than
âagreeableness,â since Kant employs the term âmoral pleasureâ (e.g., at MS 378)
but, to my knowledge, not âmoral agreeableness.â
33. According to one common usage, the German sofern is equivalent to the German
im Fall, daĂ (in case that) or vorausgesetzt, daĂ (supposing that) (Gerhard Wahrig,
ed., Deutsches WÂ¨rterbuch [GÂ¨ tersloh: Bertelsmann, 1986], 1188). See also Jacob
Grimm and Wilhelm Grimm, eds., Deutsches WÂ¨rterbuch, vol. 10, pt. 1 (Leipzig:
Hirzel, 1905), 1402. Not surprisingly, we ď¬nd Kant using sofern in a way that
makes it seem equivalent to âonly if.â For example, at MS 223 Kant says: âAn
action is called a deed insofar [sofern] as it comes under obligatory laws.â But
sofern can also mean âto the extent that.â See Grimm and Grimm, 1402.
34. The conclusion that Kantâ™s claim regarding acting from inclination amounts to
this seems to be supported by the following assertion Kant makes at GMS 442,
note: âevery empirical interest promises to contribute to our well-being by the
agreeableness that something affords, whether this happens immediately and
without a view to advantage or with regard for it.â
35. Reath, âHedonism, Heteronomy and Kantâ™s Principle of Happiness,â 50.
36. In his Anthropology, Kant also deď¬nes inclinations as habitual desires. See Anth
251 and 265. In the Religion, Kant employs a different, narrower sense of incli-
nation. He states that to have an inclination, we must be acquainted with the
object of our desire. He contrasts inclination with instinct, âwhich is a felt want
to do or to enjoy something of which one has as yet no conception (such as
the constructive impulse in animals, or the sexual impulse)â (Rel 28â“29, note;
English ed. 24, note).
37. The question may arise of why Kant here writes of desire in the narrow sense:
what would be desire in the broad sense? In my view, Kant holds that both actions
done from duty and ones not done from duty involve an agentâ™s determining
his capacity of desire. To put the point roughly, both kinds of action involve the
agentâ™s choosing to realize some object. An agent has a desire in the broad sense
for an object when he has determined his capacity of desire with respect to this
object. Yet only if his ground for determining his capacity of desire included
the prospect of his own pleasure does the agent have a desire in the narrow
sense for the object. If the agentâ™s ground for choosing to realize some object
Notes to pp. 29â“33
did not include the prospect of his own pleasure â“ for example, if the agent
set himself to realize it solely because he thought doing so was commanded by
the Categorical Imperative â“ then the agent has a desire in the broad sense for
the object but not a desire in the narrow sense for it. Of course, this reading
of Kantâ™s notion of desire in the narrow sense rests on my particular analysis of
38. See Reath, âHedonism, Heteronomy and Kantâ™s Principle of Happiness,â 47.
39. At this point, a defender of Reathâ™s interpretation might make the following
argument: âGranted, the capacity of desire is not the capacity to have a desire.
Nevertheless, to say that an agentâ™s capacity of desire has been âdeterminedâ
is only to say that she has come to have a desire. It is not to say that she has,
through her representation of an object, chosen to realize it. Therefore, as
Reath argues, Kantâ™s Metaphysics of Morals account merely suggests that pleasure
plays a role in our coming to have inclinations.â In response to this argument,
let me say that I ď¬nd no evidence that Kant conceived of the determination
(Bestimmung) of the capacity of desire in this way. Actually, there is evidence
that he conceived of it in the way I suggest. In a note to the First Introduction
to the Critique of Judgment, Kant analyzes empty wishes and longings in terms of
the âdeterminationâ of the capacity of desire. An agent has an empty wish, he
suggests, just in case, through her representation of an object, she determines
her forces to realize this object, even though it is impossible for her to realize it.
He goes on to say: âIt is indeed a not unimportant problem for anthropology to
investigate why it is that nature has given us the predisposition to such fruitless
expenditure of our forces as [we see in] empty wishes and longings (which
certainly play a large role in human life). It seems to me that here, as in all
else, nature has made wise provisions. For if we had to assure ourselves that we
can in fact produce the object, before the representation of it could determine
us to apply our forces, our forces would presumably remain largely unusedâ
(KUE 231, note). Kant seems to suggest here that the determination of an agentâ™s
capacity of desire is equivalent to her actually setting herself to try to realize an
40. At GMS 400, Kant says that the will âmust be determined by the formal principle
of volition as such when an action is done from duty, where every material
principle has been withdrawn from it.â
41. Typically, Kant deď¬nes happiness (roughly) as the complete satisfaction of all
inclinations over a lifetime. See, for example, KrV A 806/B 834; GMS 399,
405; KpV 73, 124; Rel 58 (English ed. 51). Precisely how his hedonistic and
inclination-based deď¬nitions of happiness are supposed to cohere with one
another is a complex issue. For one attempt to resolve it, see Virginia Wike,
Kant on Happiness in Ethics (Albany: State University of New York Press, 1994),
42. Thanks to Michael Slote for this example.
Chapter 2: Transcendental Freedom and the Derivation
of the Formula of Universal Law
1. For example, in section 2 he aims to show that no âmaterialâ practical principle
could be the supreme principle of morality (KpV 21â“22).
Notes to pp. 35â“41 201
2. Henry E. Allison, Idealism and Freedom (Cambridge: Cambridge University Press,
3. Henry E. Allison, Kantâ™s Theory of Freedom (Cambridge: Cambridge University
Press, 1990), 204.
4. Allison, Idealism, 204.
5. Ibid., xviii.
7. Ibid., 131.
8. Allison acknowledges this point. See ibid., 130.
9. See, for example, ibid., 130â“131. At Idealism, 130, Allison says that âthe Incor-
poration Thesis is best seen as a general thesis about how motives function in
the case of ď¬nite rational agents.â He goes on (130â“131) to say that âalthough
a ď¬nite rational agent is sensuously or âpathologicallyâ affected, that is to say, it
ď¬nds itself with a set of given inclinations and desires, which provide possible
motives or reasons to act, it is not causally necessitated to act on the basis of any
10. See, for example, ibid., 131.
11. Ibid., 132.
12. See Allison, Kantâ™s Theory of Freedom, 207, and Idealism, 151â“152.
13. Allison, Idealism, 152. I am simply assuming here that Allisonâ™s account of tran-
scendental freedom is on target. One might argue that, at least in the Critique
of Pure Reason, transcendental freedom amounts to bare independence from
14. Allison credits Karl Ameriks and Paul Guyer with raising this sort of question.
See Idealism, 124.
15. Ibid., 123â“128.
16. Ibid., 138.
17. Ibid., 152; see also Kantâ™s Theory of Freedom, 207â“208.
18. See Thomas Nagel, âUniversality and the Reď¬‚ective Self,â in Christine
M. Korsgaard, The Sources of Normativity, ed. Onora Oâ™Neill (Cambridge:
Cambridge University Press, 1996), 201â“203.
19. Allison, Idealism, 152, and Kantâ™s Theory of Freedom, 208.
20. Allison, Kantâ™s Theory of Freedom, 208.
21. Ibid., 209. See also Idealism, 152.
22. Allison, Kantâ™s Theory of Freedom, 210.
23. That Allison agrees that each maxim contains, if only implicitly, a description
of an end becomes evident at Idealism, 119, and Kantâ™s Theory of Freedom, 90â“91.
24. Allison, Kantâ™s Theory of Freedom, 210.
25. Reath challenges Allisonâ™s argument on the grounds that it does not elimi-
nate the possibility that the âPrinciple of Happiness,â rather than the moral
law, could be legitimately adopted by a transcendentally free rational agent as
his fundamental principle. See Andrews Reath, âIntelligible Character and the
Reciprocity Thesis,â Inquiry 36 (1993): 427. Reathâ™s challenge is aimed at step
4 rather than step 3 of Allisonâ™s argument.
26. This sort of response is suggested by Allisonâ™s reply to the challenge of Andrews
Reath (mentioned in note 25). See Allison, Idealism, 117.
27. For an example of Kantâ™s making this move, see GMS 416: â[O]nly law brings
with it the concept of an unconditional and objective and hence universally valid
Notes to pp. 42â“53
necessityâ (emphasis added; Kantâ™s emphasis omitted). Kant, of course, makes
it abundantly clear that universally valid means binding on all rational beings
(see, e.g., KpV 32).
28. Allison, Kantâ™s Theory of Freedom, 212, and Idealism, 152.
29. Kant acknowledges that sometimes âreason employs the unity of the maxims
in general, a unity which is inherent in the moral law, merely to bestow upon
the incentives of inclination, under the name of happiness, a unity of maxims
which otherwise they cannot have. (For example, truthfulness, if adopted as a
basic principle, delivers us from the anxiety of making our lies agree with one
another and of not being entangled by their serpent coils)â (Rel 36â“37, English
30. Allison, Kantâ™s Theory of Freedom, 213.
31. Allison, Idealism, 152â“153. See also his Kantâ™s Theory of Freedom, 213.
Chapter 3: The Derivation of the Formula of Humanity
1. Here I am following Hill. See Thomas E. Hill Jr., Dignity and Practical Reason
in Kantâ™s Moral Theory (Ithaca: Cornell University Press, 1992), 38â“41. For a
slightly different account of what Kant means by âhumanity,â see Allen W. Wood,
âHumanity as End in Itself,â in Proceedings of the Eighth International Kant Congress,
vol. 1, ed. Hoke Robinson and Gordon Brittan (Milwaukee: Marquette University
Press, 1995), 306.
2. On my reading, it simply belongs to Kantâ™s concept of an agentâ™s having a
particular practical principle that he have a sufď¬cient ground (motive) to act
on it. See section 1.8.
3. He makes the claim at KpV 60â“61. By the time he makes it, Kant is obviously
operating not only with the view that he has offered a successful derivation of the
Formula of Universal Law (see KpV 41) but also that the Formula of Universal
Law is valid. At KpV 31, he states that consciousness of the moral law, that is, the
Formula of Universal Law, may be called a âfact of reason.â
4. Amartya Sen, âEvaluator Relativity and Consequential Evaluation,â Philosophy
and Public Affairs 12 (1983): 114.
5. One might be concerned that this distinction Kant makes between natural laws
and practical laws threatens his thesis of the unity of reason. According to this
thesis, practical and speculative reason are uniď¬ed in a common principle.
â[T]here can, in the end,â claims Kant, âbe only one and the same reason,
which must be distinguished merely in its applicationâ (GMS 391). Now Kant
seems to assert that speculative reason is constitutive of natural laws, laws of
what is, and practical reason is constitutive of practical laws, that is, laws of
what ought to be. Yet how could this be if practical and speculative reason were
uniď¬ed in a common principle? Recently, Neiman has claimed that for Kant
reason, whether speculative or practical, has the role of providing laws âthat
tell us what ought to happen, even if it never does, not laws of nature, which
tell us what does happenâ (Susan Neiman, The Unity of Reason [Oxford: Oxford
University Press, 1994], 108). Neiman suggests that on Kantâ™s considered view
speculative reason is not constitutive of natural laws but, when employed prop-
erly, is regulative. Speculative reason urges us toward the end of complete and
systematic knowledge of the realm of nature. See, for example, 125â“129. If, as
Neiman asserts, for Kant both speculative and practical reason are regulative
Notes to pp. 54â“58 203
in the sense of providing ends and standards for activity, then Kantâ™s unity of
reason thesis seems less problematic than it would if he held speculative reason
to be constitutive of natural laws.
6. Of course, Kant might, in the end, simply claim that it belongs to his concept
of a practical law that it contain âthe very same determining ground of the will in
all cases and for all rational beings.â If we embrace this concept, it does follow
that, if there is a practical law (categorical imperative), then each rational agent
must hold that some object is (or objects are) unconditionally good. But then
the question arises: why should we embrace this concept?
7. See the second full paragraph at GMS 428.
8. See Christine M. Korsgaard, The Sources of Normativity, ed. Onora Oâ™Neill
(Cambridge: Cambridge University Press, 1996), 122.
9. Parts of sections 3.4â“7 stem from Samuel J. Kerstein, âKorsgaardâ™s Kantian Ar-
guments for the Value of Humanity,â Canadian Journal of Philosophy 31 (March
10. For a brief criticism of the textual accuracy of the argument Korsgaard attributes
to Kant, see Berys Gaut, âThe Structure of Practical Reason,â in Ethics and Practi-
cal Reason, ed. Garrett Cullity and Berys Gaut (Oxford: Oxford University Press,
11. Korsgaard says: âIn the argument for the Formula of Humanity, as I under-
stand it, Kant uses the premise that when we act we take ourselves to be acting
reasonably and so we suppose that our end is, in his sense, objectively goodâ
(Christine M. Korsgaard, âKantâ™s Formula of Humanity,â in Creating the Kingdom
of Ends [Cambridge: Cambridge University Press, 1996], 116). Later Korsgaard
asks: âSuppose that you make a choice, and you believe what you have opted
for is a good thing. How can you justify it or account for its goodness?â (ibid.,
121). Korsgaard appears to use the terms âgood endâ and âobjectively good
endâ interchangeably. I employ the simpler term âgood end.â
12. Ibid., 115.
13. Ibid., 114.
14. Ibid., 115; see also 120, 122.
15. Korsgaard says: âIf oneâ™s end cannot be shared, and so cannot be an object of
the faculty of desire for everyone, it cannot be goodâ (ibid., 116).
16. Ibid., 122.
17. Christine M. Korsgaard, âAristotle and Kant on the Source of Value,â in Creating
the Kingdom of Ends (Cambridge: Cambridge University Press, 1996), 227. See
also Korsgaard, âTwo Distinctions in Goodness,â in Creating the Kingdom of Ends
(Cambridge: Cambridge University Press, 1996), 259.
18. Korsgaard, âHumanity,â 114â“119.
19. See Korsgaard, The Sources of Normativity, 122. Korsgaard there suggests that the
argument she attributes to Kant in âKantâ™s Formula of Humanity,â namely what
I call the âregressive argument,â supports the following conclusion. Unless an
agent takes humanity to be unconditionally valuable, he must embrace complete
practical skepticism. If the regressive argument is effective, it demonstrates that,
if an agent maintains that some of his ends are good, then he must hold rational
nature to be unconditionally good. In other words, an agent must either hold
humanity to be unconditionally good or give up the notion that he has good ends
in Korsgaardâ™s very robust sense. Korsgaard appears to believe that an agentâ™s
abandoning the notion that he has good ends in her sense would force him into
Notes to pp. 59â“62
complete practical skepticism. If an agent gives up this notion, then, rationally
speaking, he ď¬nds himself without reasons for his actions. Unless Korsgaard
holds this, it is totally unclear how the regressive argument is supposed to sup-
port her conclusion. She needs to block the possibility that it would be rational
to deny that one has good ends (and thereby deny the necessity of holding
humanity to be unconditionally good), yet to afď¬rm that one has reasons for
20. This example stems from Gaut, âThe Structure of Practical Reason,â 174.
21. In a context similar to that of the present discussion, Allen Wood recognizes that
one might raise an objection to the sort of argument Korsgaard (and presumably
Kant) want to make here. The objection is that âif y is something valuable and
x is its source, it does not in general follow that x is something valuable, still
less that it is objectively valuable or an end in itselfâ (Allen W. Wood, Kantâ™s
Ethical Thought [Cambridge: Cambridge University Press, 1999], 130). Wood
responds to this objection by claiming that it misunderstands the argument. In
the argument, says Wood, ârational nature is not being viewed as the source of
good things (i.e., of their existence), but instead as the source of the fact of their
goodnessâ (ibid., 130). I do not believe that Woodâ™s response here threatens
my counterexample. To use Woodâ™s terms, in my example rational disapproval
is being viewed as the source of the fact of the badness of bad things. And
that rational disapproval is being viewed as the source of this fact does not
entail that rational disapproval must itself be viewed to be bad at all, let alone
unconditionally bad. So the question remains: why must the source of the fact
of good thingsâ™ goodness be viewed to be good at all, let alone unconditionally
22. In The Sources of Normativity, 122, Korsgaard offers a âfancy new modelâ of the
regressive argument. For criticism of this new version of the argument â“ a version
that appeals to the notion of a âpractical identityâ â“ see Kerstein, âKorsgaardâ™s
Kantian Arguments for the Value of Humanity,â 42â“51.
23. Korsgaard, âHumanity,â 117.
24. Ibid., 110.
25. See ibid., 111, and Christine M. Korsgaard, âMotivation, Metaphysics, and the
Value of the Self,â Ethics 109 (1998): 55.
26. Korsgaard, âHumanity,â 110â“111.
27. See ibid., 111, 113.
28. For evidence that Korsgaard is using âhumanityâ as a synonym for rational na-
ture, see ibid., 110â“114. For evidence that she equates rational nature with the
power of rational choice, see ibid., 123. At 123 Korsgaard says that âhumanity is
the power of rational choice.â
29. See ibid., 119â“124. For a summary of the regressive argument, see Korsgaard,
âTwo Distinctions in Goodness,â 256â“262. See also Korsgaard, âAristotle and
Kant on the Source of Value,â 239â“243.
30. For evidence that this is the notion of a sufď¬cient condition that Korsgaard has
in mind, see Korsgaard, âHumanity,â 122.
31. Ibid., 121.
32. See GMS 393â“395. For a recent criticism of Kantâ™s argument, especially as a
response to the kind of value realism I have discussed, see Gaut, âThe Structure
of Practical Reason,â 165â“170.
Notes to pp. 63â“70 205
33. Korsgaard, âHumanity,â 121.
35. Ibid., 122.
39. Ibid., 123.
42. For this conception of happiness in Kant, see, for example, GMS 399, 405.
43. It is not logically impossible for everyone to be happy, even on Kantâ™s desire-
satisfaction account of happiness. For we can coherently imagine a world in
which every person always gets what he wants. Of course, in this imagined world,
no one personâ™s satisfying any of his desires would preclude any other person
from satisfying any of her desires.
44. Kant says that âan impartial rational spectator can take no delight in seeing the
uninterrupted prosperity of a being graced with no feature of a pure and good
willâ (GMS 393).
45. This reading of âgood willâ would have to be broadened to accommodate Kantâ™s
view that perfectly rational beings such as God cannot act from duty. To them
the âoughtâ of duty does not apply, since their willing is necessarily in accord
with the law. See GMS 414. We might attribute to Kant the view that these beings
have a good will (engage in unconditionally good willing) just in case they act
âfor the sake of the law.â Presumably such beings are capable of doing this. And
Kant does not seem averse to the idea that acting from duty is a species of acting
for the sake of the law.
46. See Karl Ameriks, âKant on the Good Will,â in Grundlegung zur Metaphysik
der Sitten; Ein kooperativer Kommentar, ed. Otfried HÂ¨ ffe (Frankfurt am Main:
Klostermann, 1989), 54â“59.
47. See Christine M. Korsgaard, âKantâ™s Analysis of Obligation: The Argument of
Groundwork I,â in Creating the Kingdom of Ends (Cambridge: Cambridge University
Press, 1996), 60â“61.
48. I am here following Ameriks, âKant on the Good Will,â 53.
49. Ibid., 51â“54.
50. In focusing on the possibility of a value realist posing this sort of question, I am
following Gaut. See Berys Gaut, âThe Structure of Practical Reason,â 176. I do
not wish to suggest that Gaut defends the environmentalist position I have men-
tioned. He does support a version of value realism but not environmentalism. In
The Sources of Normativity, Christine Korsgaard offers an argument against value
realism. I critically discuss this argument in âKorsgaardâ™s Kantian Arguments
for the Value of Humanity.â 44â“51.
51. See Gaut, âThe Structure of Practical Reason,â 167. My criticism of Kantâ™s claim
that nothing but the good will is good without qualiď¬cation is indebted to Gautâ™s
treatment. See also H. J. Paton, The Categorical Imperative (New York: Harper &
Row, 1967), 38.
52. For a forceful challenge to Kantâ™s view that courage, cleverness, and knowledge
are not unconditionally good, see Gaut, âThe Structure of Practical Reason,â
Notes to pp. 72â“77
53. This point derives from David Cummiskey, Kantian Consequentialism (New York:
Oxford University Press, 1996), chap. 5.
Chapter 4: The Derivation of the Formula of Universal Law:
A Criterial Reading
1. Parts of this chapter have been adapted from Berys Gaut and Samuel Kerstein,
âThe Derivation without the Gap: Rethinking Groundwork I,â Kantian Review 3
2. Korsgaardâ™s reconstruction can be found in Christine M. Korsgaard, âKantâ™s
Analysis of Obligation: The Argument of Groundwork I,â in Creating the Kingdom
of Ends (Cambridge: Cambridge University Press, 1996), 43â“76. For evidence
that in Korsgaardâ™s view the derivation succeeds, see her last paragraph on
3. The numbering of the steps here is not Korsgaardâ™s.
4. Korsgaard, âKantâ™s Analysis,â 60 (emphasis omitted).
5. Ibid., 61.
7. Ibid., 63 (emphasis omitted).
8. Ibid. (emphasis omitted).
9. Ibid., 61â“62.
10. Korsgaard in her interpretation appears to be trying to exploit some of the con-
siderations on which an account of autonomy might draw to make the derivation
work. But autonomy is not mentioned once in the derivation in Groundwork I.
Its deployment belongs to the second section of the Groundwork.
11. Korsgaard, âKantâ™s Analysis,â 62.
12. There is an additional ground I have for rejecting the notion that the derivation
as interpreted by Korsgaard succeeds. Korsgaard endorses the view expressed
in step i, namely that the reason why a good-willed person does an action and
the reason why the action is right are the same. Here the assumption is that
if an action expresses good will, for example, if it is done from duty, then it is
right. But I will argue in Chapter 6 that, actually, Kant is rationally compelled
to acknowledge that an action can express good will even if it is not right.
13. Kantâ™s main task in Groundwork II also seems to be to derive the supreme prin-
ciple of morality â“ in all the complexity of its various formulas. Of course, in
neither of the ď¬rst two sections of the Groundwork does Kant claim to show that
there is a supreme principle of morality. For as he explicitly acknowledges he
has not there eliminated the possibility that our view that we are bound by moral
requirements is a âphantom of the brainâ (see section I.3).
14. For a similar point, see Marcia W. Baron, Kantian Ethics Almost without Apology
(Ithaca: Cornell University Press, 1995), 181â“182.
15. In the roughly two pages that remain in Groundwork I, Kant tries to show the
need to engage in a more rigorously philosophical discussion of the supreme
principle of morality. This discussion takes place in Groundwork II.
16. Strictly speaking, the principle is a preliminary version of the Formula of Univer-
sal Law. Kant sets out the canonical version of this formula later, in Groundwork
II (GMS 421). The preliminary version is this: âI ought never to act except in such a
way that I could also will that my maxim should become a universal lawâ (GMS 402).
Notes to pp. 79â“90 207
17. This reconstruction is based on Bruce Aune, Kantâ™s Theory of Morals (Princeton:
Princeton University Press, 1979), 32â“33.
18. At GMS 397, Kant says that he is going to explicate the concept of a good will
through exploring the concept of duty, âwhich contains that of a good will
though under certain subjective limitations and hindrances.â It is clear that
these subjective limitations and hindrances are inclinations the fulď¬llment of
which would involve acting contrary to what morality requires.
19. For Kantâ™s suggestion that moral worth is unconditional worth, see GMS 400.
Kant refers to the âunqualiď¬ed worthâ of actions at GMS 411.
20. See, for example, Barbara Herman, The Practice of Moral Judgment (Cambridge,
Mass.: Harvard University Press, 1993), 1; Henry E. Allison, Kantâ™s Theory of
Freedom (Cambridge: Cambridge University Press, 1990), 107, and Baron,
Kantian Ethics Almost without Apology, 28, n. 19. That Kant held that only ac-
tions from duty have moral worth is strongly suggested in his discussion of cases
from GMS 397â“399. In the second Critique, Kant conď¬rms that this is his view:
â[M]oral worth must be placed solely in this: that the action takes place from
duty, that is, for the sake of the law aloneâ (KpV 81). As Thomas E. Hill Jr.
has pointed out to me, Kant does not in the Groundwork explicitly state that
all actions from duty have moral worth. However, I think Kant strongly implies
that he holds this near the end of GMS 399: âThe second proposition is this:
an action from duty has its moral worth not in the purpose to be attained by it but
in the maxim in accordance with which it is decided uponâ (emphasis added,
Kantâ™s emphasis omitted).
21. At GMS 400, Kant suggests that in an action done from duty, the will is deter-
mined by the law.
22. For Kantâ™s usage of the term âimpulse,â see GMS 434, 444.
23. See also the end of GMS 427 and the beginning of GMS 428.
24. Yet, as we noted in section 1.8, this is not the end of the story. Under Theorem II,
Kant states that all material principles place the determining ground of the will
in the pleasure or displeasure to be received from an object. On Kantâ™s concep-
tion, an agent has sufď¬cient motive to conform to a material practical principle
only if he expects that doing so will result in the realization of some object he
desires (e.g., his visiting Grantâ™s tomb) and he expects that realizing this object
will have a hedonic payoff. Whenever an agent acts on a material practical prin-
ciple, that is, follows the principleâ™s prescription for trying to realize an object,
she has some hedonic motivation.
25. For what might be a different way Kant has of distinguishing between material
and formal principles, see the end of GMS 427 and the beginning of GMS 428.
Allison offers a helpful discussion of some difď¬culties in Kantâ™s use of the term
âformal.â See Henry E. Allison, Idealism and Freedom (Cambridge: Cambridge
University Press, 1996), 150.
26. See Henry E. Allison, Kantâ™s Transcendental Idealism (New Haven: Yale University
Press, 1983), 78.
27. With regard to theoretical rules, Kant says the following: â[E]xperience never
confers on its judgments true or strict, but only assumed and comparative
universality, through induction. We can properly only say, therefore, that, so far
as we have hitherto observed, there is no exception to this or that rule. If, then,
a judgment is thought with strict universality, that is, in such a manner that
Notes to pp. 91â“99
no exception is allowed as possible, it is not derived from experience, but is
valid absolutely a priori. Empirical universality is only an arbitrary extension of a
validity holding in most cases to one which holds in allâ (KrV B 3â“4). Just as
empirically based universality is insufď¬cient to ground theoretical rules that al-
low of no possible exception, so it is insufď¬cient to ground practical rules that
allow of none. See KpV 26.
28. Kant claims to prove the validity of the principle, â[A]ct on no other maxim
than that which can also have as object itself as a universal lawâ (GMS 447).
Chapter 5: Criteria for the Supreme Principle of Morality
1. Chapter 6 explores objections to this ď¬rst criterion.
2. This section has been adapted from my paper âThe Kantian Moral Worth
of Actions Contrary to Duty,â Zeitschrift fÂ¨ r Philosophische Forschung 53 (1999):
3. This interpretation is not wholly unproblematic. In the Preface to the Ground-
work, Kant remarks that moral laws require âa power of judgment sharpened by
experience, partly in order to distinguish in which cases they are applicable and
partly to provide them with access to the will of the human being and efď¬cacy
for his fulď¬llment of themâ (GMS 389). The question is: according to Kant,
does everyone, that is, every rational agent, acquire this power of judgment in
the course of maturing? Of course, if Kant would answer afď¬rmatively, then the
interpretation at hand is not really threatened by this remark.
4. Allison, for example, says: âStarting with the assumption, itself questionable,
that actions performed from duty cannot, objectively speaking, be contrary to
duty, he proposes to limit his consideration to actions that are at least in agree-
ment with duty (pflichtmÂ¨ Ăig).â See Henry E. Allison, Kantâ™s Theory of Freedom
(Cambridge: Cambridge University Press, 1990), 109.
5. Curzer adopts this interpretation. See Howard Curzer, âFrom Duty, Moral Worth,
Good Will,â Dialogue 36 (1997): 290â“291.
6. That Kant would hold this is not as obvious as it might at ď¬rst appear. Kant
asserts that duties cannot conď¬‚ict: if at the same time two moral rules prescribe
different actions, then it cannot be a duty to act in accordance with both. See
MS 224. But suppose that disagreeing with Kant on this score, an agent holds
that duties can conď¬‚ict. Further suppose that here and now she believes that
she has conď¬‚icting duties, a duty to keep a promise to meet a student and
a duty to take her mother-in-law to the airport. It seems that in performing
the latter action, she could both believe that she was acting contrary to duty
(i.e., her duty not to break her promises), yet be acting from duty. Moreover,
Kant might acknowledge this possibility, although he would insist that the agent
was mistaken in her belief that duties can conď¬‚ict. I think it is clear that if
an agent believes along with Kant that duties cannot conď¬‚ict, then she could
not, in doing something she believed to be contrary to duty, be acting from
7. It is striking that in a work where Kant is at pains to explore the concept of duty,
he mentions not one example of an action done from duty but which conď¬‚icts
8. At KpV 81, Kant says that moral worth âmust be placed solely in this: that the
action takes place from duty, that is, for the sake of the law alone.â
Notes to pp. 100â“101 209
9. At KpV 79, Kant says: âA maxim is . . . morally genuine only if it rests solely on
the interest one takes in compliance with the law.â
10. Here one might suggest the following. Supererogatory actions are not morally
required; they are optional in the sense of beyond what duty requires. Never-
theless, in performing one (e.g., by jumping on a live grenade to save oneâ™s
comrades), one might put aside entirely the inď¬‚uence of inclination. So there
are some actions that are neither morally required nor at all inď¬‚uenced by in-
clination. In response, note ď¬rst that Kant does not recognize the category of
supererogatory actions. For discussion of why, see Marcia W. Baron, Kantian
Ethics Almost without Apology (Ithaca: Cornell University Press, 1995), chap. 1.
Moreover, even if Kant did recognize this category, a supererogatory action
would not be done from duty in his sense. In an action from duty, an agentâ™s
will is determined by her representation of the law in itself (GMS 401). But in
a supererogatory action, her will is presumably not determined simply by her
representation of the law, since she is aiming âabove and beyondâ what the law
11. See GMS 402 for evidence that Kant recognizes this point.
12. For discussion of Kantâ™s notion of respect, see Allison, Kantâ™s Theory of Freedom,
120â“128, and Ralph C. S. Walker, âAchtung in the Grundlegung,â in Grundlegung
zur Metaphysic der Sitten: Ein kooperativer Kommentar, ed. Otfried HÂ¨ ffe (Frankfurt
am Main: Klostermann, 1989), 97â“116.
13. See Barbara Herman, The Practice of Moral Judgment (Cambridge, Mass.: Harvard
University Press, 1993), 13â“17. See also Baron, Kantian Ethics Almost without
14. See Herman, Practice, 16.
15. It is striking that Herman does not cite a single case of Kantâ™s mentioning that
an agent performs a particular action from duty as a âlimiting conditionâ or
16. This point is made by H. J. Paton, The Categorical Imperative (New York: Harper
& Row, 1967), 47â“50; Herman, Practice, 12; Allison, Kantâ™s Theory of Freedom,
110â“111; and Baron, Kantian Ethics, 150. Both Allison and Baron note that Kant
implies early in his Groundwork I discussion of duty that an agentâ™s having an
inclination to do something does not in itself preclude him from doing it from
duty. Kant suggests that to determine whether an action is from duty is difď¬cult
in cases âwhen an action conforms with duty and the subject has, besides, an
immediate inclination to itâ (GMS 397). But why, Allison and Baron rightly ask,
would he suggest this to be difď¬cult if he adopted the view that having an
immediate inclination to do something itself precluded one from doing it from
duty? If he adopted this view, then, it seems he would simply say that if one has
an immediate inclination to an action, then it is thereby impossible for him to
do it from duty. (I discuss Kantâ™s distinction between mediate and immediate
inclination in section 5.5.) For an opposing interpretation of Kant, see Noa
Latham, âCausally Irrelevant Reasons and Action Solely from the Motive of
Duty,â Journal of Philosophy 91 (1994): 599â“618.
17. Of course, Kant does hold that it is impossible for me to be certain that I have
kept the promise from this motive. See GMS 407.
18. Allison puts the point in this way. See Allison, Kantâ™s Theory of Freedom, 111.
19. In addition to Baron, Kantian Ethics, 146â“187, see, for example, Paul Benson,
âMoral Worth,â Philosophical Studies 51 (1987): 365â“382; Paul Guyer, Kant on
Notes to pp. 103â“115
Freedom, Law, and Happiness (Cambridge: Cambridge University Press, 2000),
287â“303; Richard G. Henson, âWhat Kant Might Have Said: Moral Worth and
the Overdetermination of Dutiful Action,â Philosophical Review 88 (1979):
39â“54; Tom Sorell, âKantâ™s Good Will and Our Good Nature,â Kant-Studien
78 (1987): 87â“101.
20. This paragraph has been inď¬‚uenced by Henry Allison. See Allison, Kantâ™s Theory
of Freedom, 116â“118.
21. Typically, Kant deď¬nes happiness (roughly) as the complete satisfaction of all
inclinations over a lifetime â“ thus seemingly not in purely hedonistic terms. See,
for example, KrV A 806/B 834; GMS 399, 405; KpV 73, 124; Rel 58 (English
22. Here I am following Marcia Baronâ™s initial account of overdetermined actions.
See Kantian Ethics, 150. (She offers a more detailed account on 156â“157). In con-
cluding that for Kant there can be no overdetermined actions done from duty
and from inclination, I am also following Baron. She writes: âOverdetermined
actions involving duty and inclination do not merely lack moral worth; they are
not intelligibleâ (161). However, the grounds Baron offers for this conclusion
differ signiď¬cantly from the ones I have given (see Kantian Ethics, 156â“161).
23. Kant implies this, for example, in his âsecond propositionâ: âan action from duty
has its moral worth not in the purpose to be attained by it, but in the maxim in
accordance with which it is decided uponâ (GMS 399; emphasis added, Kantâ™s
24. I have been inď¬‚uenced here by Guyerâ™s discussion of this Kantian notion. See
Paul Guyer, Kant and the Experience of Freedom (Cambridge: Cambridge University
Press, 1993), 344â“351. See also Guyer, Kant on Freedom, Law, and Happiness,
25. See also Rel 30â“31 (English ed. 26): âFor when incentives other than the law
itself (such as ambition, self-love in general, yes, even a kindly instinct such as
sympathy) are necessary to determine the will [WillkÂ¨ r] to conduct conformable
to the law, it is merely accidental that these causes coincide with the law, for they
could equally well incite its violation.â
26. This example is based on one offered by Barbara Herman. See Herman, Practice,
27. Here I am following ibid., 5â“6.
28. See ibid., 5.
29. I offer some criticism of this argument in section 8.10.
30. Of course, Kant expresses this view not only in the Groundwork, but in the second
Critique as well. See, for example, KpV 20.
31. Kant seems to think that we must interpret this principle as a material one. But
I do not see why we must. See section 7.2.
32. Ultimately, I do not believe that Kant should rely on this argument. For reasons
I crystallize in section 8.2, I think Kant should abandon one of the argumentâ™s
key premises, namely that ought implies can.
Chapter 6: Duty and Moral Worth
1. I do not wish to imply that Williamsâ™s and Stockerâ™s criticisms are meant to
apply to Kantian morality exclusively. They both aim their criticisms at utilitarian
theories as well, just to name one additional target.
Notes to pp. 116â“120 211
2. Both my understanding of this objection and my response to it have been heavily
inď¬‚uenced by Baronâ™s work. See Marcia W. Baron, Kantian Ethics Almost without
Apology (Ithaca: Cornell University Press, 1995), 118â“135.
3. Michael Stocker, âThe Schizophrenia of Modern Ethical Theories,â Journal of
Philosophy 73 (1976): 462.
4. At MS 456, Kant states that we, human beings, have a duty to use our capacity
to have sympathetic feelings as a means to promote âactive and rational benev-
olence.â Perhaps the person Kant describes in the Groundwork is lacking in this
capacity. Kant suggests that ânature had put little sympathyâ in his heart (GMS
5. Oakley appears to defend this position. See Justin Oakley, Morality and the
Emotions (New York: Routledge, 1992), 57â“63, 83.
6. See Bernard Williams, Moral Luck: Philosophical Papers, 1973â“1980, (Cambridge:
Cambridge University Press, 1981), 18. Williams discusses a case put forth by
Charles Fried. The case I discuss is inspired by, but differs from, the one Williams
7. I have been inď¬‚uenced in my thinking regarding cases of this kind by Baron,
Kantian Ethics, 136â“140; Paul Guyer, Kant and the Experience of Freedom
(Cambridge: Cambridge University Press, 1993), 392â“393; and Barbara
Herman, The Practice of Moral Judgment (Cambridge, Mass.: Harvard University
Press, 1993), 41â“42.
8. My treatment of this thesis (sections 6.5â“9) mirrors my discussion in Samuel
Kerstein, âThe Kantian Moral Worth of Actions Contrary to Duty,â Zeitschrift fÂ¨ r u
Philosophische Forschung 53 (1999): 45â“66.
9. Some commentators ignore (what I take to be) Kantâ™s Groundwork denial of
moral worth to all morally impermissible actions. See, for example,
H. J. Paton, The Categorical Imperative, 5th ed. (New York: Harper & Row, 1965),
46â“50. Others take note of this denial but do not explore in any detail Kantâ™s
grounds for it or its plausibility. See, for example, Henry E. Allison, Kantâ™s
Theory of Freedom (Cambridge: Cambridge University Press, 1990), 109. Roger
Sullivan offers some helpful references to remarks Kant makes relating to the
sources of error in moral judgment. See Roger J. Sullivan, Immanuel Kantâ™s
Moral Theory (Cambridge: Cambridge University Press, 1989), 57â“60. How-
ever, Sullivan does not directly address the issue I discuss here, namely that of
whether the logic of Kantâ™s own moral theory rationally compels him to
embrace the possibility that morally impermissible actions can have moral
10. Kant sums up what he takes to have accomplished in Groundwork I when he says:
âThus, then, we have arrived, within the moral cognition of common human reason,
at its principleâ (GMS 403, emphasis added).
11. By writing of an actionâ™s (potentially) expressing good will rather than being
identical to it, I am implicitly invoking the âwhole characterâ understanding of
a good will. The point at issue would not be altered if I instead invoked the
âparticular actionâ understanding of a good will. For the distinction between
these two notions of a good will, see section 3.7.
12. Thomas Nagel focuses on this aspect of Kantâ™s doctrine in the Groundwork.
See Thomas Nagel, âMoral Luck,â in Mortal Questions (Cambridge: Cambridge
University Press, 1979), 24â“25. See also Thomas E. Hill Jr., Respect, Pluralism,
and Justice (Oxford: Oxford University Press, 2000), 159â“160.
Notes to pp. 121â“124
13. Recall Kantâ™s dualistic view of the ultimate grounds of human action. Either
we act from inclination, or we act from duty (section 1.6). Since he denies that
actions from duty can conď¬‚ict with duty, Kant must hold that actions that conď¬‚ict
with duty are done from inclination.
14. One might wonder whether, in Kantâ™s view, an agentâ™s failure to perform a
morally permissible action is always a failure of will. For it seems that when a
person is drunk, he can perform a morally impermissible action that does not
stem from a failure of will â“ if only because he is so intoxicated that, at that
moment, he has no will to fail. In response, I think Kant would argue that the
drunk personâ™s morally impermissible actions really do stem ultimately from a
failure of will, namely a failure to suppress the inclination to drink in the ď¬rst
place. In the Lectures on Ethics, Kant says: âeverything is imputable that pertains
to freedom, even though it may not have arisen directly through freedom, but
indirectly nevertheless. E.g., what a person has done in a state of drunkenness
may well not be imputed; but he can be held accountable for having got drunkâ
15. Strictly speaking, Kant suggests that we use for moral appraisal âthe universal
formula of the categorical imperative: act in accordance with a maxim that can at
the same time make itself a universal lawâ (GMS 436â“437). In the example, Colonel
Mikavitch takes âthe universal formula of the categorical imperativeâ to be a
version of the Formula of Universal Law. Wood has argued recently that it is
actually a version of the Formula of Autonomy; see Allen W. Wood, Kantâ™s Ethical
Thought (Cambridge: Cambridge University Press, 1999), 188â“189.
16. Admittedly, Kant sometimes comes out against the permissibility of suicide in
any circumstances. For example, in the Lectures on Ethics he says: âThere are
many conditions under which life has to be sacriď¬ced; if I cannot preserve it
other than by violating the duties to myself, then I am bound to sacriď¬ce it,
rather than violate these duties; yet on the other hand, suicide is not permitted
under any conditionâ (LE 372).
17. An orthodox Act Utilitarian would deď¬ne the good as (something like) the
maximum happiness of all sentient beings. He would then conceive an actionâ™s
moral value purely in terms of the degree to which it promotes the good thus
deď¬ned. For the Act Utilitarian, if Stramâ™s action diminishes the general hap-
piness, then it has no moral value â“ unless, perhaps, all of the actions open
to him would diminish the general happiness, but this action is the one that
would diminish it least. As an orthodox Act Utilitarian, Stram would himself
hold the moral value of his lying to the politician to be contingent on its effects.
Of course, we need not agree with Stramâ™s own take on when his actions have
moral worth. For a far more detailed discussion of utilitarianism and moral
worth, see sections 7.3â“5.
18. Actually, there seems to be a third kind of case of an actionâ™s being morally
impermissible, yet having moral worth. In this kind of case, an agent does
his best to adopt the correct moral principle but fails. He also does his best
to apply properly the principle he has adopted but fails at that as well. In
this kind of case, there is a failure both of principle choice and of principle
19. For an interesting discussion of the Kantian conception of conscience, as well
as two other conceptions of it, see Hill, Respect, Pluralism, and Justice, 260â“274.
Notes to pp. 126â“128 213
20. In the Anthropology from a Pragmatic Point of View, one might argue, Kant ac-
cepts that a person who adopted a standard of moral judgment other than the
Categorical Imperative might perform morally impermissible actions, which
nonetheless had moral worth. Consider his discussion of âcharacterâ (Charakter).
In much the same language he uses in Groundwork I to describe a good will, he
tells us that character has intrinsic worth (inneren Wert) and is beyond all price.
A few sentences earlier, he says: âto have character relates to that property of
the will by which the subject has bound himself to certain practical principles
which he has unalterably prescribed for himself by his own reason. Although
these principles may sometimes indeed be false or defective, nevertheless the
formal element of the will as such, which is determined to act according to ď¬rm
principles (not shifting hither and yon like a swarm of gnats), has something
precious and admirable to it, which is also something rareâ (Anth 292). Here,
one might argue, Kant is advocating the idea that, even if a person has adopted
false or defective practical principles, she can have character. Since Kant believes
character to have intrinsic worth, he also holds that actions expressing character
have a special worth. Therefore, concludes the argument, Kant thinks that a
person could perform a morally impermissible action that had moral worth: an
action of obeying a self-given, yet false, principle because she believed obeying
it was the right thing to do. In response, note that Kant does not really em-
brace the idea that a person who has adopted false practical principles can have
character. Shortly after the cited passage, Kant says: âCharacter requires max-
ims, which proceed from reason and from ethicopractical principlesâ (Anth
293). Kant then lists principles that, he suggests, a person of character would
have to live by, including those of not speaking an untruth intentionally, not
dissembling, and not breaking oneâ™s (legitimate) promise (Anth 294). These
are, of course, just the sort of principles that Kant believes to stem from the
Categorical Imperative. At the very least, if Kant believed someone living by
false principles could have character, among her false principles could be none
that prevented her from also embracing those on Kantâ™s list. In addition, Kant
suggests that a person of character would act on principles that are valid (gelten)
for everybody (Anth 293). Kant would hardly claim that false principles would
be valid for everybody! Despite initial appearances, Kant is not suggesting in
the Anthropology that a person who lives by false principles can have character,
nor, by extension, that the personâ™s action on such principles can have moral
worth. He is praising the quality of sticking to oneâ™s principles much in the same
way that, in Groundwork I, he acknowledges the value of self-control and calm
deliberation. These qualities, he says, are rightly held in high esteem, but they
have no absolute (i.e., moral) worth (see GMS 394).
21. Some of this disagreement is manifest in our discussion of the Formula of
Universal Law in Chapter 8. At any rate, Korsgaard discusses three ways of
interpreting the Formula of Universal Law, each one of which is found in
the literature: Christine M. Korsgaard, âKantâ™s Formula of Universal Law,â in
Creating the Kingdom of Ends (Cambridge: Cambridge University Press, 1996),
77â“105. For another way of reading this formulation, see Thomas W. Pogge,
âThe Categorical Imperative,â in Kantâ™s âGroundwork of the Metaphysics of Moralsâ:
Critical Essays, ed. Paul Guyer (Lanham, Md.: Rowman & Littleď¬eld, 1998),
Notes to pp. 129â“141
22. At GMS 411, for example, Kant says: âFrom what has been said it is clear that
all moral concepts have their seat and origin completely a priori in reason, and
indeed in the most common human reason just as in reason that is speculative
in the highest degree.â
23. Kant, of course, holds that it is only by virtue of having its source in reason alone
that the Categorical Imperative could be valid.
24. Michael Slote has suggested this sort of objection, using the example of a con-
scientious Nazi prison guard. See Michael Slote, Goods and Virtues (Oxford:
Oxford University Press, 1983), 63. See also Jonathan Bennett, âThe Conscience
of Huckleberry Finn,â Philosophy 49 (1974): 123â“143.
25. Hannah Arendt, Eichmann in Jerusalem (New York: Penguin, 1994), 135â“137.
26. As Thomas Pogge has suggested to me, one wonders how, in Kantâ™s view, an
inquisitor can be sure that it is right to spare a defendant who, in the inquisitorâ™s
view, has violated divine doctrine.
27. Kant himself holds that âit is a basic moral principle, which requires no proof,
that one ought to hazard nothing that may be wrongâ (Rel 185, English ed.
173). Hazarding nothing that may be wrong involves being sure that what one
proposes to do is right. For Kant goes on to say that âconcerning the act which
I propose to perform I must be sure that it is not wrong; and this requirement is
a postulate of conscienceâ (Rel 186, English ed. 174). I think it is questionable
whether Kant has highlighted here a basic moral principle that requires no
proof. Why doesnâ™t it require any proof ? Also, one might wonder what relation
Kant takes to hold between this principle and the Categorical Imperative. Does
the Formula of Universal Law (or perhaps the Formula of Humanity) entail this
principle? How, precisely, does it do so?
28. See Lawrence Blum, Friendship, Altruism and Morality (London: Routledge and
Kegan Paul, 1980), 12â“15. For an account of emotions in general that is in the
spirit of Blumâ™s account of sympathy, see Oakley, Morality and the Emotions, 7â“16.
29. This discussion has been inď¬‚uenced by Barbara Hermanâ™s distinction between
the motive of an action and its object. See Herman, Practice, 25.
30. Oakley, Morality and the Emotions, 83â“84.
31. See Michael Slote, Morals from Motives (New York: Oxford University Press,
32. See Oakley, Morality and the Emotions, 101â“102.
33. A defender of the notion that all actions from sympathy have moral value might
argue that a soldier who displayed such a lack of sympathy toward ethnic mi-
norities would, contrary to what has been suggested here, be incapable of acting
from sympathy at all, even toward his fellow soldier. The idea would be that an
utter lack of sympathy toward one group is incompatible with genuine sympathy
toward another group. I suppose that this is possible, but I see no good reason
to believe it.
Chapter 7: Eliminating Rivals to the Categorical Imperative
1. See also Berys Gaut and Samuel Kerstein, âThe Derivation without the Gap:
Rethinking Groundwork I,â Kantian Review 3 (1999): 18â“40.
2. In the Groundwork version of this argument (GMS 444, especially the lower
half of the page), Kant does not use the term âmaterial principle,â although
Notes to pp. 143â“153 215
he does employ it earlier in the text (GMS 400). As we will see, he instead
writes of âheteronomy of the will.â Nor does Kant in the Groundwork version
explicitly invoke the notion that material principles are conditional for their
motivational force on the agentâ™s expectation of a hedonic payoff. At GMS 444,
however, he does suggest that the motivational force of such principles depends
on an âimpulseâ that the representation of an object exerts on the will.
3. Here I am following David Cummiskey who argues that not all consequentialist
principles must be considered to be âmaterialâ ones. See David Cummiskey,
Kantian Consequentialism (New York: Oxford University Press, 1996), 46â“48.
4. At KpV 34, Kant writes: âThus, the happiness of other beings can be the object of
the will of a rational being. But if it were the determining ground of the maxim,
one would have to presuppose that we ď¬nd not only a natural satisfaction in
the well-being of others but also a need, such as a sympathetic sensibility brings
with it in human beings.â
5. Amartya Sen, âUtilitarianism and Welfarism,â Journal of Philosophy 76 (1979):
6. Ibid., 464.
7. This is not an unusual conception of states of affairs. See Sen, âUtilitarianism,â
464â“465; âEvaluator Relativity and Consequential Evaluation,â Philosophy and
Public Affairs 12 (1983): 128â“129; and âWell-Being, Agency, and Freedom: The
Dewey Lectures 1984,â Journal of Philosophy 82 (1985): 181â“182. See also Bernard
Williams, âA Critique of Utilitarianism,â in J. J. C. Smart and B. Williams, Utili-
tarianism For and Against (Cambridge: Cambridge University Press, 1973), 83.
8. Thanks to Thomas Pogge and Michael Slote for pushing me on this point.
9. An agentâ™s conforming to EU amounts to her performing an action that she
expects will yield as great a sum total of well-being as would any alternative action
available to her. Suppose that from duty an agent wills to conform to EU. From
duty she wills to do what she expects will yield as great a sum total of well-being as
would any alternative action available to her. From the perspective of a Kantian
conception of acting from duty, it is hard to see how, in this case, she could fail
to do what she expects will yield this result. It seems that she could only fail if
she indulged her inclinations. But since she has acted from duty, she has not
10. Actually, as we will see in section 8.2, arguments such as the one summarized
in this paragraph will require a modiď¬cation of one of Kantâ™s criteria, namely
11. For a concise criticism of Act Utilitarianism on the grounds that it generates a
set of duties that clashes with common sense, see Richard B. Brandt, âToward
a Credible Form of Utilitarianism,â in Contemporary Utilitarianism, ed. M. Bayles
(New York: Doubleday, 1968), 146â“147.
12. At least for advocates of the Formula of Universal Law, this seems to be a some-
what dangerous argument to make. For, as I contend in Chapter 8, it is very
doubtful whether this formula generates a set of duties that squares with ordi-
nary moral thinking.
13. For a defense of this sort of principle, see Thomas Hurka, Perfectionism (New
York: Oxford University Press, 1993), especially chaps. 2 and 4.
14. This point can also be illustrated with reference to rational perfection. An
increase in an agentâ™s rational perfection might not result from an agentâ™s
Notes to pp. 154â“167
willing to develop his rational capacities. For example, in an attempt to develop
these capacities, the agent might take an experimental âbrain-enhancingâ drug
that ends up diminishing them.
15. For Cummiskeyâ™s claim that the âď¬rst propositionâ is consistent with conse-
quentialism, see Kantian Consequentialism, 27; for his claim that the âsecond
propositionâ is also consistent with it, see 39.
16. This is nearly a direct quotation from ibid., 99. Cummiskeyâ™s detailed statement
of his principle spans four paragraphs, 98 to 99.
17. Cummiskey denies that this requirement entails that we ought to maximize the
number of rational beings. See ibid., 91.
18. Ibid., 150.
19. Cummiskey suggested this sort of reading of his principle in a paper presented at
the annual Paciď¬c Division meeting of the American Philosophical Association,
Berkeley, California, March 1997.
20. See Cummiksey, Kantian Consequentialism, 4, 79, 156.
21. See ibid., 6, 16.
22. See ibid., 11â“12.
23. Cummiskey discusses this argument at length in chapter 4 of his book. See ibid.,
24. As did we, Cummiskey examines Kantâ™s argument as reconstructed by
Korsgaard. See ibid., 81, n. 11.
25. See ibid., 156.
26. TC would have to be rephrased to accommodate Kantâ™s view that a viable can-
didate for the supreme principle of morality must be capable of being bind-
ing on all rational agents, including perfectly rational ones such as angels.
Instead of âYou ought to honor your father and mother; you ought not to
kill; you ought not to commit adultery . . . ,â the principle would have to read
something like âHonor your father and mother; do not kill; do not commit
adultery. . . . â Presumably, angels would, by virtue of their perfect rationality,
necessarily act in accordance with whatever principle was the supreme principle
of morality. Therefore, with respect to angels, the âoughtâ in TC would be out of
Chapter 8: Conclusion: Kantâ™s Candidates for the Supreme
Principle of Morality
1. This is an argument Kant suggests in his discussion of his âsecond propositionâ
at GMS 399â“400.
2. Recalling an argument we discussed (and criticized) in Chapter 3, one might
suspect that Kant himself implies that the principles could not fulď¬ll the cri-
terion. To summarize this argument, if an agent holds there to be a supreme
principle of morality, then she must also hold there to be something uncon-
ditionally good, claims Kant. For if she did not take there to be something
unconditionally good, then she might ď¬nd herself without sufď¬cient motive to
conform to the principle and thus, for reasons that require no repeating here,
the principle could not be the supreme principle of morality. In light of this ar-
gument, it might appear that Kant (perhaps without being aware of it) commits
Notes to pp. 168â“169 217
himself to the view that the representation of a principle as a law does not pro-
vide an agent with sufď¬cient incentive to abide by it. After all, according to the
argument, to have sufď¬cient incentive, an agent must (at least in some cases?)
take conforming to the principle to promote or secure something uncondition-
ally valuable. In response, note that for Kant it is an agentâ™s representing a
principle as universally and unconditionally binding that gives rise to her con-
ception of the good. So ultimately her incentive for acting lies nevertheless in
3. Kant, of course, actually tests the maxim of false promising using the Formula of
the Law of Nature, stated in the third full paragraph at GMS 421.
4. Christine M. Korsgaard, âKantâ™s Formula of Universal Law,â in Creating the Kingdom
of Ends (Cambridge: Cambridge University Press, 1996), 80, 92â“93. Baron joins
Korsgaard in holding the Practical Contradiction Interpretation to offer the most
plausible account of how, precisely, a maxim such as that of false promising fails
the Formula of Universal Law test. See Marcia W. Baron, âKantian Ethics,â in
Marcia W. Baron, Philip Petit, and Michael Slote, Three Methods of Ethics (Oxford:
Blackwell, 1997), 69â“70.
5. According to Korsgaard, on the Practical Contradiction Interpretation, âthe con-
tradiction is that your maxim would be self-defeating if universalized: your action
would become ineffectual for the achievement of your purpose if everyone (tried
to) use it for that purposeâ; see Korsgaard, âKantâ™s Formula of Universal Law,â 78
6. Here one might object that at an early stage in the imagined world, that is,
before everyone has realized that money borrowed simply on a promise will not be
repaid, an agent acting on FPM may well attain her end. Perhaps Korsgaard would
respond to this objection by saying that, nevertheless, in willing the imagined
world, the agent would be willing a world in which her chances of getting money
on a false promise were severely diminished. And that would be enough to render
practically irrational willing the imagined world at the same time as acting on
7. As Korsgaard acknowledges, one might also read the Formula of Universal Law
to land a person acting on such a maxim of false promising in a logical con-
tradiction. According to (what Korsgaard calls) the Logical Contradiction Inter-
pretation, the universalization of the maxim would be as follows: from self-love,
when anyone believes himself to be in need of money, he borrows (rather than
tries to borrow) money on a promise to repay it, even though he knows that this
will never happen. In order to be able to will this world, an agent needs to be
able to conceive of it. However, she cannot really conceive of the world, suggests
Kant. For not everyone in ď¬nancial need could get a loan based simply on a
promise if no such person ever repaid a loan she received in this way. Credi-
tors would not part with their money. In other words, the practice of lending
money to those in need based simply on their promise to repay would cease
to exist if none of them ever repaid their loans. There simply is no world in
which when each and every agent ď¬nds herself in ď¬nancial need, she gets money
through false promising. The agent considering the false promising maxim has
been forced into a logical contradiction. The Formula of Universal Law requires
her to hold that she can conceive of the world of her universalized maxim, since
Notes to pp. 170â“174
it requires her to will this world and (as she must acknowledge) she could not
will this world without being able to conceive of it. However, she concludes that
she cannot conceive of this world. There is obviously a logical contradiction in
holding, as the agent would (presumably) have to, that something both is and
is not conceivable. Korsgaard grants that this interpretation is well supported
by Kantâ™s text. (See Korsgaard, âKantâ™s Formula of Universal Law,â 81â“82.) It is
worth mentioning that a philosophical difď¬culty seems to arise in connection
with the Logical Contradiction Interpretation. It is not obvious that the world
of the universalized maxim is inconceivable. Granted, it is very unlikely that
people would continue to lend money simply on a promise that they would be
repaid, even though whenever they did lend it, they were not repaid. But this
unlikely scenario is (arguably) not inconceivable. Thanks to Thomas Pogge for
8. See Korsgaard, âKantâ™s Formula of Universal Law,â 93, and Barbara Herman, The
Practice of Moral Judgment (Cambridge, Mass.: Harvard University Press, 1993),
9. This maxim and my analysis of it stem from Herman, Practice, 138â“139.
10. See Baron, âKantian Ethics,â 73.
11. This maxim stems from Pogge. See Thomas W. Pogge, âThe Categorical Imper-
ative,â in Kantâ™s âGroundwork of the Metaphysics of Moralsâ: Critical Essays, ed. Paul
Guyer (Lanham, Md.: Rowman & Littleď¬eld, 1998), 190.
12. Moreover, in this world it is questionable whether anyone would be earning a
comfortable living (as Jack conceives of one). With everyone trying to become
a professor to earn a living, who would do the work necessary to sustain an
economy in which it is possible for (many) people to have their own houses,
cars, and computers?
13. See Pogge, âThe Categorical Imperative,â 189â“196.
14. Ibid., 190.
15. See ibid., 191.
17. Ibid., 192.
19. Ibid., 191.
20. As Korsgaard acknowledges, the Practical Contradiction Interpretation also
faces difď¬culties regarding maxims of violence. See Korsgaard, âKantâ™s Formula
of Universal Law,â 100. On this interpretation, it is not obvious that acting on
the one we have been discussing â“ that of killing for revenge â“ turns out to be
morally impermissible. And a maxim such as âIn order to release my anger, I
will punch anyone who offends meâ seems to sail through.
21. See Pogge, âThe Categorical Imperative,â 196.
22. We have already noted Korsgaardâ™s reservations as to whether, on the inter-
pretation she champions, the Formula of Universal Law generates adequate
results regarding certain maxims of violence (see Korsgaard, âKantâ™s Formula
of Universal Law,â 100). Pogge, it appears, does not think that the Formula of
Universal Law itself produces an adequate set of duties; for, in his view, it fails to
generate a duty of beneď¬cence (see Pogge âThe Categorical Imperative,â 196).
According to Herman, if we read the Formula of Universal Law as âa method
Notes to pp. 175â“184 219
of judgment to be used by agents in determining the permissibility of their own
maximsâ (and we have read the formula in this way), then it is not effective. See
Herman, Practice, 143. Herman (147â“157) offers an innovative account of what
the Formula of Law procedure might actually accomplish, but I do not discuss
23. Herman, Practice, 143.
24. Here I am following Allen W. Wood, âHumanity as End in Itself,â in Proceedings
of the Eighth International Kant Congress, vol. 1, ed. Hoke Robinson and Gordon
Brittan (Milwaukee: Marquette University Press, 1995), 317, n. 2.
25. I am assuming here (and I take Kant to hold) that there is no way of treating
humanity such that one is treating it neither as a means nor as an end.
26. Here I am following Thomas E. Hill Jr., Dignity and Practical Reason in Kantâ™s
Moral Theory (Ithaca: Cornell University Press, 1992), 41â“42.
27. That for Kant in the Formula of Humanity âendâ is equivalent to âend in itselfâ
is clearly implied at GMS 428.
28. See Hill, Dignity and Practical Reason, 47â“49.
29. See Thomas W. Pogge, âKant on Ends and the Meaning of Life,â in Reclaiming
the History of Ethics, ed. Andrews Reath, Barbara Herman, and Christine M.
Korsgaard (Cambridge: Cambridge University Press, 1997), 361â“362.
30. Here âreasonâ refers to a Kantian motivating reason.
31. In the Metaphysics of Morals, Kant seems to derive the duty of beneď¬cence from
the Formula of Universal Law. See MS 393, 453.
32. According to Kant, we can safely presuppose that each human agent has this
end by a necessity of nature (GMS 415).
33. For discussion of how demanding a duty of beneď¬cence Kant endorses (or is
compelled by his own views to endorse), see Marcia W. Baron, Kantian Ethics
Almost without Apology (Ithaca: Cornell University Press, 1995), chaps. 1â“3;
Cummiskey, Kantian Consequentialism, chap. 6; Hill, Dignity and Practical Reason,
34. Allen W. Wood, Kantâ™s Ethical Thought (Cambridge: Cambridge University Press,
1999), 153 (emphasis added).
35. There is something very odd about setting oneself to pursue the end of being
deceived. It would seem that pursuing the end would likely involve knowing (at
least roughly) in what circumstances one is to be deceived (e.g., by a card shark
in Las Vegas). But if one knows (even roughly) in which circumstances one is
to be deceived, then there is a sense in which one is not entirely deceived.
36. For far more detailed discussion of some of the practical implications of the