¹ Enchiridion, xxix. This work and all classical works referred to below (Diogenes Laertius's Lives of Eminent Philosophers, Cicero's De officiis, Plutarch's Parallel Lives, and Seneca's Epistulae morales) are cited by Kant in his anthropology lectures and in other works. See the Literaturverzeichnis to Kant's anthropology lectures in the German Academy of Sciences (GAS) edition of Kants gesammelte Schriften (Berlin: Walter de Gruyter, 1900—), 25: 1565–658Google Scholar.

² The rigorism charge is often levelled in conjunction with the formalism charge, which focuses on Kant's allegedly one-sided focus on the formal features of willing to the exclusion of the unique conditions and consequences of action. Perhaps the single most influential instance of the formalism charge is Max Weber's influential criticism of JudaeoChristian/Kantian Gesinnungsethik, or ‘ethic of pure conviction’, for its otherworldly ‘purity’ and irresponsible detachment from the scene of ethical decision. Weber formulates this criticism in the 1919 pamphlet Politik als Beruf (Politics as Vocation).

³ O'Neill, Onora, Constructions of Reason (Cambridge: Cambridge University Press, 1989), p. 130Google Scholar, depicts the rigorism charge in these terms.

⁴ Ameriks, Karl, ‘Kant on the good will’, in Hoffe, O. (ed.), Grundlegung zur Metaphysik der Sitten: Ein kooperativer Kommentar (Frankfurt a.M: Vittorio Klostermann, 1989), pp. 45–65Google Scholar, deals with the rigorism charge in a discussion of Kant's conception of the good will. The ‘inhuman standard’ charge has also been directed at Kant's notion of ‘imperfect duties’. The complaint is that fulfilling the two basic ‘ends that are also duties’, namely one's own perfection and the happiness of others, requires leading the life of a saint. Since one can never do enough to develop one's natural talents or increase the sum of happiness in the world, Kant's theory of imperfect duties fails to make room for the supererogatory in ethics. This charge has been met by pointing out that one's own happiness is on the same moral footing as the happiness of others, and so unremitting self-sacrifice, to the extent that it cancels out one's own chances for a good life, cannot be a requirement of moral virtue . Baron, Marcia, Kantian Ethics (Almost) without Apology (Ithaca, NY, and London: Cornell University Press, 1995), pp. 21–58Google Scholar, answers this charge in detail.

⁵ The term ‘rigorism’ (and its German counterpart, Rigorismus), signifying rigidity in principle or practice, gained usage in the early eighteenth century. Thus when Kant affirms the rigorism of his moral theory in Religion within the Bounds of Mere Reason, he is fully aware of the negative connotations of the term. Kant, Religion within the Boundaries of Mere Reason, trans . Giovanni, George di in Religion and Rational Theology (Cambridge Edition of the Works of Immanuel Kant), ed. Wood, A. and Giovanni, G. di (Cambridge: Cambridge University Press, 1996), 6: 36Google Scholar . Cited by marginal numbers (= volume and page numbers of the German Academy of Sciences [GAS] edition of Kants gesammelte Schriften (Berlin: Walter de Gruyter, 1900—)Google Scholar . All Kant citations refer to the volume and page numbers of the GAS edition (present in the margins of all Cambridge translations).

⁶ , O'NeillConstructions of Reason (1989)Google Scholar and Herman, Barbara, The Practice of Moral Judgement (Cambridge, MA, and London: Harvard University Press, 1993)Google Scholar.

⁷ The point is not made frequently enough that Aristotle's ethics, too, would become nonsense if we took him literally to mean that the immediate object of a courageous deed is an ‘activity of the soul expressing virtue’ (or happiness). Virtuous action (or happiness) is the ‘end’ of specific virtuous actions only at a higher order of reflection: one acts courageously because the action is appropriate, called for, because there is need to come to the defence of a friend or a cause – not first of all in order to cultivate or activate a virtue.

⁸ The exceptions to the rule are Ameriks, K., ‘Kant on the good will’Google Scholar, and Allison, Henry, Kant's Theory of Freedom (New York: Cambridge University Press, 1990), pp. 148–61CrossRefGoogle Scholar.

⁹ Engstrom, Stephen, ‘Conditioned autonomy’, Philosophy and Phenomenological Research, 48 (1988), 435–53CrossRefGoogle Scholar . Engstrom is not alone in calling attention to alleged unfortunate effects of combining rigorism and purism: see also O'Connor, Daniel, ‘Good and evil disposition’, Kant-Studien, 76 (1985), 288–302CrossRefGoogle Scholar . The concluding sentence of O'Connor's essay is worth citing: ‘Instead of an analysis which would permit us to appreciate the indefinite variety of strengths and weaknesses of moral character which human beings manifest, [Kant] gives us a view in which men and women are evil by nature and good, if they are good, only by the grace of God’ (302).

¹⁰ I adopt the standard translation of Gesinnung as ‘disposition’ despite reservations I share with Munzel, G. Felicitas, Kant's Conception of Moral Character: The ‘Critical’ Link of Morality, Anthropology, and Reflective Judgement (Chicago: University of Chicago Press, 1999), p. xviiGoogle Scholar, that ‘disposition’ fails to convey the radical sense of spontaneity that Kant accords to Gesinnung. Kant strictly contrasts ‘Gesinnung’ and ‘Disposition’ in the Groundwork; the latter term denotes an inherited temperament not a moral commitment (4: 435). ‘Ethical attitude’ is closer to the sense in which Kant uses Gesinnung, but the constant combination of ‘ethical’ with ‘good’ or ‘evil’ becomes awkward. Finally, note that Luther translates Paul's phronema as gesinnet sein. See Romans 8: 6 (Lutherbibel): ‘Aber fleischlich gesinnet sein ist der Tod, und geistlich gesinnet sein ist Leben und Friede.’ By the time of the composition of Religion, Kant sets little in store by the Pauline distinction between carnal and spiritual mindedness, but the sense of one's ultimate priorities or basic orientation that Luther intends by (the passive verbal form of) Gesinnung is clearly preserved intact in Kant's usage.

¹¹ Engstrom uses ‘autonomy’, ‘conditioned autonomy’ and ‘heteronomy’ with reference to the principle of the agent's power of choice (Willkur); that is, he intends them ‘subjectively’ rather than ‘objectively’. Autonomy in Engstrom's usage applies only to persons of genuine moral virtue; it corresponds to what in the Doctrine of Virtue Kant calls ‘autocracy’, or more familiarly, virtue (6: 383). Heteronomy is ‘found in the wills of agents that act contrary to the moral law when such action is necessary to further personal ends’ (‘Conditioned autonomy’, 439).

¹² In somewhat clearer terms, moral character is the holistic property of a complex and dynamic ‘system’ of subjective moral endeavour and experience (comprising moral choices and initiatives, moral responses and perceptions, moral self-examinations and self-conceptions, as well as any higher-order reflections, use of rhetoric, and argumentation that have moral consequences).

¹³ We might ask where the line is to be drawn between idealized and realistic views of agency. Engstrom's notion of ‘conditioned autonomy’ implies that there are two higher-order principles, one or the other of which is operative in any circumstance. But to conceive matters this way is merely to take the ‘idealization’ down a notch. Further, granted that persons do not normally take themselves to be acting on maxims, Engstrom's criticism of Kant's tendency to idealization invites the question whether Kant's equation of choice and maxim construction is not also an implausible idealization.

¹⁴ D. O'Connor, ‘Good and evil disposition’, distorts matters in arguing that ‘there is no propensity to good’ in Kant's theory of incentives (297). Kant clearly does think that respect for law has an ‘inner dynamic that supplies motivation’ (denied by O'Connor). It is true that moral motivation operates by way of higher order reflection, but does that make it less vital or real as a motive? I do not see that it does. In the Doctrine of Virtue Kant states that the moral incentive does not merely operate negatively: if it did there would be no duties of virtue (6:383). Incidentally, Kant speaks explicitly of a ‘propensity to good’ in Anthropology (7:329).

¹⁵ The 1792 article that would become part one of Religion bore only the title ‘Concerning the radical evil in human nature’. Critics of Kant's agent rigorism would do well to bear in mind the subtitle that Kant added with the publication of the book; it implies a more subtle view of human character than is usually imputed to him. We note that Kant was familiar with Diogenes Laertius's Lives of Eminent Philosophers, and so with Laertius's influential depiction of the ancient debate between the Stoics and the Peripatetics in that work. The following passage might even be the source of Kant's famous talk of the ‘crooked timber’ of humanity : Lives of Eminent Philosophers, vol. 2 (Loeb Classical Library, with facing English translation by Hicks, R. D., London and Cambridge, MA: Harvard University Press, 1970), p. 231Google Scholar . One of the chief tasks Kant sets himself in the first two parts of Religion is to reconcile the Stoic and Peripatetic views, to work out a theory in which agent rigorism (‘there is nothing intermediate’) is consistent with making room for gradations of virtue (‘the state of moral improvement’).

¹⁶ Herman, B., ‘Making room for character’, in Engstrom, S. and Whiting, J., Aristotle, Kant and the Stoics (Cambridge: Cambridge University Press, 1996), p. 36.Google Scholar

¹⁷ A good precedent for taking ‘intelligible character’ in a non-noumenal way is provided by Kant himself, in the sense-change of his usage of Denkungsart (best translated as ‘conduct of thought': see , MunzelKant's Conception of Moral Character, p. xvi)Google Scholar . In the first Critique, Denkungsart is strictly equated with noumenal causality, as contrasted with Sinnesart or ‘empirical character’. From the second Critique onward Kant uses Denkungsart in a sense that exactly corresponds to ‘empirical character’ as used in the first Critique (5: 99; compare A549–50/B577–8). One point to bear in mind is that ‘empirical character’ in the first Critique is still a character of an agent's rational choice (understood in light of all those factors that make the choice inevitable). That is, it does not exclude voluntary choice in the conventional (‘psychological’) sense, but on the contrary, it marks the compatibilist feature of Kant's thinking (a sense in which Kant affirms compatibilism). But if Denkungsart can so easily migrate from the noumenal to the phenomenal, then nothing prevents us from construing Kant's appeal to an ‘intelligible deed’ in Religion in non-noumenal (compatibilist) terms.

¹⁸ Kantian moral psychology is developmental: evil precedes virtue because self-love takes root in us before we acquire the capacity for moral reflection: ‘for us – dependent as we are on objects of the senses –happiness is by nature the first that we desire and desire unconditionally’ (6: 46 note). Social comparisons and antagonisms step in to ensure that the propensity to evil develops without hindrance ‘into the corrupted practical attitude that Kant calls ‘depravity’.

¹⁹ But note that it would not make sense to begin in our ethical training from the presupposition that some of us are incurably ‘depraved’ from childhood. Kant's reference early in Religion to our ‘immutable disposition’ as being what it is ‘from birth’ (6: 22) – an unfortunate rehearsal of talk in the second Critique of ‘born villains’ manifesting ‘early wickedness from childhood’ an d ‘quite incapable of improvement’ (5: 99) – is an isolated statement in Religion, inconsistent with what else he has to say in the book.

²⁰ So Allison, H., Kant's Theory of Freedom (1990), pp. 159–60Google Scholar, 169, goes too far in denying that frailty is compatible with virtue. Frailty is incompatible with virtue only if one makes no efforts to reduce the extent of one's frailty, or if these ‘efforts’ prove wholly ineffectual (and so insincere - but then we are witnessing the second stage, impurity). Virtue is distinguished from what Kant calls ‘mere lack of vice’ not by an absence of frailty, as Allison suggests, but by steady progress in (never completely) overcoming it.

²¹ It might seem inconsistent of Kant to accord frailty, which involves actual violations of duty, a higher moral status than impurity, which apparently involves consistently lawful action. The appearance of discrepancy is annulled, however, if we note that the ‘empirical character of virtue’ is simply virtue as appearance; how an action appears can change depending on wh o is looking at it and what they are looking for. Th e greater degree of self-deception involved in impurity cashes ou t in a smaller list of infringements owned up to. Frailty involves less self-deception than impurity; one has been sufficiently conscientious in one's self-examinations to recognize one's moral failings. Note that Kant identifies impurity and falsehood or mendacity (see for example 8: 267). Impurity ‘lies deep in what is hidden, where the huma n being knows how to distort even inner declarations before his own conscience’ (8: 270).

²² In the spirit of a dynamic view of character and a developmental moral psychology, I note that the transition from impurity to depravity is continuous, not hard and fast. The misrepresentation of one's own attitude characteristic of impurity is helped along by practised avoidance of the principles expressed in one's action. But depravity involves a denial that the moral requirement has anything to do with subjective principles of one's actions. Thus we have to deal with stages or degrees of a basic tendency to downplay the question of the principles on which one acts.

²³ Ameriks, Karl, Kant and the Fate of Autonomy: Problems in the Appropriation of the Critical Philosophy (Cambridge: Cambridge University Press, 2000), p. 337.CrossRefGoogle Scholar

²⁴ This sense of character also occurs in the second Critique, when Kant defines character as ‘consistent conduct of thought in accordance with unchangeable maxims’ (5: 152). The word translated as ‘consistent’, konsequent, has the senses of single-mindedness, rigour, and resolve lacking in the English term.

²⁵ The Life of Sulla in Plutarch's Lives (Loeb Classical Library, with facing English translation by Perrin, B., London and Cambridge, MA: Harvard University Press, 1968), vi. 7Google Scholar. A memorable and gruesome example of Sulla's firmness in holding to his maxims is his commanding of the immediate slaughter of 6,000 men in the circus at Rome before proceeding to address the senators in the (adjacent) city hall.

The shrieks of such a multitude, who were being massacred in a narrow space, filled the air, of course, and the senators were dumbfounded; but Sulla, with the calm and unmoved countenance with which he had begun to speak, ordered them to listen to his words and not concern themselves with what was going on outside, for it was only that some criminals were being admonished by his orders, (xxx. 2–3)

I note that Cicero, who survived Sulla's regime, could observe that Sulla's cause, unlike Caesar's, was ‘a just one’, adding that his victory proved appalling (De officiis vii. 26). Critics who insist that Kant's reflections on evil would have been very different had he lived through the Holocaust must not have read Thucydides or Plutarch.

²⁶ In this light we can make sense of Kant's statement in an early Reflexion, ‘first form any character at all, then a good one’ (R 1162, 17: 514). Kant could not say this if he saw any danger that one might go about forming an evil character.

²⁷ The point is that the line distinguishing ‘diabolical’ and ‘perverse’ wills blurs on analysis. If the moral law is the criterion or principle of the ‘supreme good’, then to prefer any other principle to it is inconsistent with choosing the supreme good; it is to choose the recognized worse over the recognized better of the two principles. Consequently, choosing self-love over duty at a high level of abstraction amounts to an ungrounded preference for the worse of two principles, and an ungrounded rejection of the better. That is as good as choosing evil for its own sake. All noumenal preference for self-love over morality is diabolical, or at least ‘incomprehensible’. As Kant writes in the Opus postumum, ‘The evil principle would be a subjective practical principle without a principle (subjectiver practiscber Grundsatz ohne Princip) – to act against all principle, indeed; so it is a contradictio in adjecto' (204-5). I read this as a rejection of the coherence of noumenal evil.

²⁸ Relevant here is a note in the Doctrine of Right in which Kant invokes the idea of a criminal whose ‘maxim is diametrically opposed to the law, in the sense of being contradictory to it (hostile to it, so to speak)’ (6: 322 note). He adds: ‘As far as we can see, it is impossible for a human being to commit a crime of this kind, a formally evil (wholly pointless) crime; and yet it is not to be ignored in a system of morals (although it is only the idea of the most extreme evil)’ (ibid.).

²⁹ If we lived in the idyllic isolation that Rousseau provides for his noble savage, our disposition would be good. In a primeval forest the propensity to evil would never progress beyond a certain susceptibility to temptation were our circumstances to change. In fact, Kant's discussion of Adam's original ‘innocence’ indicates his acceptance of Rousseau's view that the pre-historical human being lacked a propensity to evil. Kant's primary ‘objection’ to Rousseau is that the self-sufficient, reclusive homme is a fiction: the propensity to evil is ‘woven into human nature’ because the human being is ‘by nature’ social. The truly solitary, ‘natural’ human being would be much like the Stoic sage, attaining autarkeia and a tranquil flow of life in wishing the things that happen to be as they are. Kant recognizes the similarity between Rousseau's natural human being and Seneca's sapiens (6: 20). Kant probably learned the insight that evil is catching from Seneca before he heard it confirmed by Rousseau. Seneca writes to Lucilius:

When a mind is impressionable and has none too firm a hold on what is right, it must be rescued from the crowd: it is so easy for it to go over to the majority. A Socrates, a Cato or a Laelius might have been shaken in his principles by a multitude of people different from himself: such is the measure of the inability of any of us, even as we perfect our personality's adjustment, to withstand the onset of vices when they come with such a mighty following. A single example of extravagance or greed does a lot of harm. (Letters to Lucilius vii: 10)

Seneca, Letters from a Stoic, trans. Robin Campbell (London and New York: Penguin Books, 1969), p. 71. Allen Wood, Kant's Ethical Thought (Cambridge: Cambridge University Press, 1999), pp. 291-6, clarifies Kant's proximity to and reliance on Rousseau in forming his conception of evil. But we should not underestimate the influence of Hellenistic philosophy on both Kant and Rousseau in just this regard.

³⁰ Kant goes on in the same note to distinguish choice from execution: ‘Those inclinations only make more difficult the execution of the good maxims opposing them; whereas genuine evil consists in our will not to resist the inclinations when they invite transgression, and this disposition is really the true enemy’ (6: 59 note). Kant's appeal to grace is supposed to cover the problem of the execution of the good maxims. But I reject this distinction between ‘will’ and ‘execution’, which merely invokes the separateness (chorismos) of a noumenal choice; my discussion of resolution should show why we must dispense with the distinction.

³¹ Herman, B., The Practice of Moral Judgement, p. 36.Google Scholar

³² Though its focus is English Methodism rather than German Pietism, William James's discussion of ‘conversion’ in chapters 9 and 10 of The Varieties of Religious Experience (London: Longmans Green, 1902Google Scholar throws intriguing light on the pietistic context of Kant's formulations of moral regeneration.

³³ Korsgaard, Christine, Creating the Kingdom of Ends (Cambridge: Cambridge University Press, 1996), pp. 180–1CrossRefGoogle Scholar, writes: ‘You cannot, just by making a resolution, acquire a virtue or recover from a vice. Or better still, we will say that you can, because you are free, but then we must say that only what happens in the future establishes whether you have really made the resolution or not.’ Robert Louden, ‘Kant's Virtue Ethics’, in Immanuel Kant (vol. 2), Practical Philosophy, ed. Klemme, H. F. and Kuehn, M. (Aldershot: Dartmouth and Ashgate, 1999), pp. 473–89Google Scholar, makes a similar point: ‘One cannot decide at noon on Monday to be courageous and saintly, and then suddenly become so by Tuesday’ (480).

³⁴ On this point I concur with Albrecht, Michael, ‘Kants Maximenethik’, Kant-Studien, 85 (1994), 129–46CrossRefGoogle Scholar . Albrecht overstates his case, however, in claiming that maxim adoption is incompatible with any deviations from the maxim. Albrecht also goes too far in claiming that ‘in Kant the maxim is from the beginning a) a deliberate decision that is b) executed and applied by very few persons’ (135). Albrecht argues, even more, that Kant denied that maxims could be followed unconsciously and that Kant thought it absurd t o suppose that all (rational) human beings act on maxims. All of Albrecht's claims are contradicted in one fell swoop by Kant's attribution t o the whole human species of a propensity to evil rooted in a freely chosen maxim. Moreover, a statement from Kant's 1791 lecture in anthropology proves that Kant could affirm in the same breath that maxims are products of freedom and that we are often unconscious of them: ‘Beings who act freely must act in such a way that maxims are always the basis of their actions. If these maxims are abiding, then we call them character. We are not always conscious of our maxims, but act according to them.’ Die philosophische Hauptvorlesungen Immanuel Kants (according to a recently discovered notebook from Count Heinrich zu Dohna-Wundlacken), ed. Kowalewski, A. (Hildesheim: Georg Olms Verlagsbuchhandlung, 1965), p. 324Google Scholar.

³⁵ There is some confusion with respect to what ‘holiness’ designates. In Religion ‘holiness’ no longer designates a disposition devoid of sensible incentives, but merely one in which duty has priority over all other incentives. So it is technically possible for a sensible being to possess a holy disposition (we can conceive of a species of sensible being that is not radically evil).

³⁶ Kant makes these observations in a short work, On the Common Saying: That May Be Correct in Theory, But It Is of No Use in Practice, published in the same year as Religion (1793).

³⁷ In the first Critique Kant states that we can speak of an infinite progression only in cases when the relevant whole is empirically given 511/B 539). He argues that in cases when the totality is a mere ideal, we must speak of an indefinite progression. Holiness is not and cannot be empirically given; accordingly, it makes best sense to treat holiness not as a (noumenal) property but as a practical principle of the greatest possible continuation and extension of our moral effort, one that allows no achieved degree of moral perfection to hold as absolute.

³⁸ If we recognize how widespread is our endeavour to obscure our actual practical attitude by, perversely, taking a narrow, legalistic view on our actions, isolating them from our conduct as a whole, then the practical function of the dramatic scenario of judgement day becomes perspicuous (6: 77). In this light the ethical force of Seneca's proto-Christian account of a last reckoning is striking:

Anyway, here's what I do: I imagine to myself that the testing time is drawing near, that the day that is going to see judgement pronounced on the whole of my past life has actually arrived. I'm going to leave it to death to settle what progress I've made. Without anxiety, then, I'm making ready for the day when the tricks and disguises will be put away and I shall come to a verdict on myself, determining whether the courageous attitudes I adopt are really felt or just so many words, and whether or not the defiant challenges I've hurled at fortune have been mere pretence and pantomime. Away with the world's opinion of you -it's always unsettled and divided... It's only when you're breathing your last that the way you've spent your time will become apparent. (Letters to Lucilia xxvi. 8)

The distinction between the moral and legal senses of virtue in just the sense in which Kant intends it is present here. The right implication to draw here is that the distinction between noumenon and phenomenon to which Kant appeals in contrasting moral and legal senses of virtue resolves into a proto-existentialist distinction between authentic and self-deceptive modes of self-presentation, between authenticité and mauvaise foi.