From radical evil to constitutive moral luck in Kant’s Religion

Abstract

The received view is that Kant denies all moral luck. But I show how Kant affirms constitutive moral luck in passages concerning radical evil from Religion within the Boundaries of Mere Reason. First, I explicate Kant’s claims about radical evil. It is a morally evil disposition that all human beings have necessarily, at least for the first part of their lives, and for which they are blameworthy. Second, since these properties about radical evil appear to contradict Kant’s even more famous claims about imputation, ‘ought implies can’, and free will, I unpack Henry Allison’s proof of radical evil and show how it is consistent with interpretations of Kant’s broader views about morality. Third, I define and illustrate the category of constitutive moral luck and argue that Kant embraces the existence of constitutive moral luck given Allison-style interpretations of radical evil. This provides a reason for philosophers to reject the received view, and it creates an occasion for Kantians and Kant scholars to check their reasons if they deny moral luck.

Moral luck occurs when factors outside of an agent’s control positively affect how much praise or blame she deserves (Hartman 2017, 23; cf. Nagel 1979, 26).¹ By ‘positively affect’, I mean that the luck at issue does not undermine praiseworthiness or blameworthiness and does make a positive difference in how much praise or blame is deserved. The source of the lack of control differentiates various kinds of moral luck. Thomas Nagel’s (1979, 28) taxonomy of sources includes consequences, circumstances, constitution, and causal determination.

The received view, in the literature on moral luck and elsewhere, is that Immanuel Kant’s moral philosophy rules out all kinds of moral luck. Consider some representative statements: ‘Kant’s moral philosophy is one of the grandest attempts in Western philosophy to banish any form of luck from the realm of morality’ (Vaida 2014, 124); ‘the Kantian [moral] position is presented as entirely incompatible with the possibility of luck’ (Athanassoulis 2005, 100; see also Kahn 2021, 366–368; Nagel 1979, 24; Williams 1981, 38).

Elsewhere, I have argued that there are plausible interpretations of Kant in which he embraces or does not deny various kinds of moral luck. Kant does not deny moral luck in results (Hartman 2019a; see also Kahn 2021; Moran 2019), and Kant embraces moral luck in actions influenced by lucky circumstances and psychological dispositions that affect the degree of difficulty in doing the right thing (Hartman 2024; see also Moran 2019).²

In this article, I expand the argument against the received view. Appealing to passages from Religion within the Boundaries of Mere Reason about radical evil, I argue that Kant embraces both moral luck in a disposition and in an action influenced by a lucky disposition. Ascribing moral luck in a disposition to Kant’s philosophy is new. Other pro-moral luck interpretations of Kant concern only moral luck in results and actions.

Radical evil is a morally evil disposition that human beings must have, at least for part of their lives,³ and for which they are blameworthy. On behalf of the claim about necessary possession, Kant asserts that the morally evil propensity is ‘innate’ (1996c, 6:22), ‘universal’ in human beings (1996c, 6:30), ‘woven into human nature’ (1996c, 6:30), and ‘entwined with humanity itself and, as it were, rooted in it’ (1996c, 6:32). On behalf of the claim about blameworthiness, Kant asserts that the evil propensity is ‘something that a human being can be held accountable for’ (1996c, 6:32), ‘we can well be responsible for the propensity to evil … imputed to the subject as itself guilty of it’ (1996c, 6:35), and it is something for which human beings have ‘innate guilt’ (1996c, 6:38). Radical evil seems to be a paradigm case of constitutive moral luck because the possession of this innate and universal disposition in human beings is at least partially outside of their control and yet they are blameworthy for this disposition.

Although these passages strongly indicate that Kant embraces constitutive moral luck, there is no consensus about how best to understand radical evil and its consistency with Kant’s famous view that moral goodness, badness, praiseworthiness, and blameworthiness must be traced back to exercises of free will. In view of these puzzles, I appeal to Henry Allison’s (1990, 2002, 2020) prominent⁴ interpretation of radical evil because it provides a general approach to radical evil that showcases its consistency with Kant’s moral ideas about imputation, ‘ought implies can’, and free will.

I proceed as follows. First, I unpack Kant’s claims about radical evil. Second, I explicate Henry Allison’s proof of radical evil and show how the general move coheres with a plausible interpretation of Kant’s broader moral commitments. Third, I define and illustrate the nature of constitutive moral luck and show how Allison’s interpretation, and other relevantly similar interpretations, commit Kant to embracing two kinds of constitutive moral luck. This article provides another reason for philosophers to reject the received view. It also creates an occasion for Kantians and Kant scholars to check their reasons if they deny moral luck.

Kant on radical evil

According to Kant (1996c, 6:18), human persons are ‘radically evil’. Radical evil is a morally evil propensity, a corrupt character that inclines a human person to do what is morally wrong.⁵ The inclination is to prioritize self-love and self-conceit over respect for the moral law (Kant 1996c, 6:36). When human persons act on this corrupt prioritization, they do what is morally wrong, for example, by making an irrational exception for themselves or by using others as a mere means to advance their own happiness.

Kant’s claims that radical evil is ‘innate’ (1996c, 6:22) and ‘morally evil’ (1996c, 6:32) generate a puzzle. After all, these claims seem to contradict Kant’s claim about imputing moral evil:

The human being must make or have made himself into whatever he is or should become in a moral sense, good or evil. These two must be an effect of his free power of choice, for otherwise they could not be imputed to him and, consequently, he could be neither morally good nor evil (1996c, 6:44, italics in original).

In other words, actions and dispositions are morally good or evil only if they are the product of free choice. Since radical evil is a morally evil disposition (and a blameworthy disposition), it must be freely chosen. But how can an innate disposition be freely chosen?

Kant offers a solution. Human persons freely choose their innate fundamental moral orientation in a timeless noumenal free choice, an ‘intelligible deed cognizable through reason alone apart from any temporal condition’ (1996c, 6:31). The content of this timeless deed is to choose a ‘supreme maxim’ that incorporates and subordinates either the incentive of the moral law to the incentive of self-love (the morally evil propensity), or the incentive of self-love to the incentive of the moral law (the morally good propensity) (1996c, 6:31). Human beings do timelessly freely choose the morally evil propensity, and this propensity provides the dispositional basis for all free choices in time, which accounts for the innate character of the propensity. This intelligible deed, however, has no phenomenal deed as its counterpart (1996c, 6:31). All we observe in the phenomenal world is its effect, which is our fundamental evil moral orientation. So, the timeless free choice is what explains the innate character of radical evil.

But Kant’s appeal to a timeless choice generates yet another puzzle because Kant also claims that radical evil is ‘universal’ (1996c, 6:30), ‘woven into human nature’ (1996c, 6:30), and ‘entwined with humanity itself and, as it were, rooted in it’ (1996c, 6:32). Why would all human persons make the same free choice to subordinate the incentive of the moral law to the incentive of self-love as their fundamental moral orientation? And what does that free choice have to do with human nature? Kant, in my view, fails to answer these questions and merely points to an inductive generalization as evidence for radical evil: ‘we can spare ourselves the formal proof that there must be such a corrupt propensity rooted in the human being, in view of the multitude of woeful examples that the experience of human deeds parades before us’ (1996c, 6:32–33; italics in original). But an empirical generalization is inadequate to explain the universality of radical evil.

Allison’s proof of radical evil

Henry Allison (1990, 2002, 2020) provides a solution to this puzzle on Kant’s behalf that has the potential to explain the universality of radical evil and its relevance to human nature. In brief, the proof is that human beings must be either good or evil in their fundamental moral orientation; human nature rules out the fundamental good moral orientation; thus, the fundamental moral orientation must be evil. So, human nature ensures that human beings timelessly choose the morally evil propensity.

The first premise is explicit in Kant’s writing. In a timeless deed, human beings must choose either a morally good or evil propensity (1996c, 6:22–25).⁶ Human nature furnishes human beings with two incentives, and only two incentives, to action that must be incorporated into their supreme maxim that itself constitutes their fundamental moral orientation. These two incentives are self-love and respect for the moral law. If human beings timelessly choose to prioritize respect for the moral law over self-love, the meta-maxim, and fundamental moral orientation, is good. If human beings fail to prioritize respect for the moral law, they prioritize self-love instead. This choice is morally evil due to its turning away from the moral law. There is no morally indifferent option because the only possible meta-maxims involve a morally good or evil priority; there is also no mixed character option because either choice amounts to a global priority. Thus, human beings must choose either a morally good or evil propensity.

Second, Allison’s (2002, 342) innovation is to argue that human nature rules out choosing the morally good fundamental orientation, or the morally good propensity. To see how human nature rules it out, consider what the morally good propensity would be. It would be a supreme maxim to subordinate the incentive of self-love to the incentive of respect for the moral law. That meta-maxim amounts to a ‘spontaneous preference for the impersonal requirements of morality over an agent’s need as a sensuous being’ (Allison 2002, 342). A fundamentally morally good human person, then, could not be tempted to transgress duty and the moral law would not even function as a constraint for them; such a human being would be beyond duty because their automatic disposition is to do what is permissible (Allison 2002, 342).

But human nature rules out this spontaneous preference for morality. Human nature includes what Kant calls ‘a predisposition to humanity’ that provides the incentive to self-love that must be incorporated in some way into the supreme maxim (1996c, 6:27; cf. 6:26–28). This predisposition is the capacity to use reason to promote self-love; it is a necessary end of human rationality to pursue happiness: ‘it is unavoidable for human nature to wish for and seek happiness, that is, satisfaction with one’s states …’ (1996e, 6:387). This unavoidable end of happiness requires agents at least first to consider the claims of happiness when deliberating about what to do (Allison 2002, 342). This is not to say that human beings cannot choose to prioritize respect for the moral law over self-love on a particular occasion but that such a choice can occur only after happiness first gets its due consideration. Such prior consideration of happiness is precisely what the morally good propensity rules out. That is, if human beings make the timeless free choice for the meta-maxim that subordinates self-love to respect for the moral law, they would have a spontaneous preference for morality and would automatically fulfil duty from duty without first considering the claims of happiness. So then, given the necessary end for happiness that is a part of human nature in the predisposition to humanity, human beings cannot choose a fundamentally good moral orientation.

Third, we can now complete Allison’s proof of radical evil. Because human beings must have either a fundamentally good or evil orientation and they cannot have a fundamentally morally good orientation, they must have a fundamentally morally evil orientation. So then, that human beings must choose the evil propensity explains the universality of radical evil.

There are important questions about the plausibility of Allison’s proof, just as there are questions about the plausibility of any interpretation of Kant’s vexing claims about radical evil, but it is not my aim to explore them in this article.⁷ Instead, I lean on Allison’s general approach to show how any interpretation of the choice to be radically evil due to a necessary feature of human nature can be consistent with Kant’s views on free will, duty, and imputation, which highlights a general approach to Kant’s embracing constitutive moral luck from radical evil. I accomplish this aim by considering and responding to two objections.

Objection 1: Kant’s incompatibilism rules out a necessary choice’s being a free choice. Incompatibilism is the view that an action’s being causally determined by factors external to one’s self is incompatible with its being a free action. Free choice has the property of ‘absolute spontaneity’ (1996c, 6:24). This spontaneity may be described negatively as ‘independence from the determining causes of the world of sense’ (1996b, 4:452) and positively as the initiation of new causal chains that begin ‘a series of consequences entirely from himself’ (1998, A555/B583).

Reply 1: If Kant is a leeway incompatibilist, a necessitated choice cannot be free. According to leeway incompatibilism, a choice is free only if the agent has alternative possibilities at the moment of choice. But necessitation rules out alternative possibilities. Thus, the choice to be radically evil cannot be necessitated and free if Kant is a leeway incompatibilist.

But Allison (2002, 343) plausibly interprets Kant as a sourcehood incompatibilist: ‘Kant characterizes freedom in terms of a causality of reason rather than a general capacity to do otherwise’ (see also Pereboom 2006, 542–543). According to sourcehood incompatibilism, an action is free only if the agent’s rationality and will are the proper source of the action. Causal determination by a factor external to one’s self, such as from the external world, rules out being the proper source of the action. But the choice for radical evil is different. It is necessitated by a feature internal to the agent: human rationality. As Allen Wood (1984, 82) puts it, ‘Kant holds that a holy will is free even though its acts are necessitated, because they are necessitated from within reason rather than by sensuous impulses that are foreign to our rational nature.’ According to the sourcehood view, necessity is compatible with freedom if the necessity proceeds from reason. So, the choice to be radically evil can be a free choice even though it is necessitated precisely because Kant is a sourcehood incompatibilist rather than a leeway incompatibilist.

Objection 2: The necessary free choice to be radically evil involves an inconsistency if there is a duty to choose the morally good propensity. Here is the inconsistency:

1. (1) The morally good propensity cannot be chosen.

2. (2) Ought implies can.

3. (3) There is a duty to choose the morally good propensity.

(1) follows from Allison’s proof and other approaches to radical evil that appeal to a feature of human nature to explain why human beings universally choose the evil propensity; thus, (1) is in good standing. (2) is an obvious feature of Kant’s moral philosophy: ‘we ought to conform to it, and therefore we must also be able to’ (1996c, 6:62) and ‘So considered, a rational being can now righty say of every unlawful action he performed that he could have omitted it …’ (1996a, 5:98). That is, if a person had a moral duty to x and she failed to x, she must have been able to x, and thus she must have had alternative possibilities at the moment of choice in that circumstance.⁸ Although leeway is not essential to freedom in general, it is essential to free choices that violate duty. Thus, (2) is plausible. So, the inconsistent triad depends on a defence of (3). Although I have not seen a defence for it, one might defend it by analogy to one of Kant’s other duties. Kant embraces an obligation to be holy: ‘Here the command is “be holy”’ (1996e, 6:446; cf. 1996c, 6:66), which amounts to a duty to acquire a character above reproach that is well-pleasing to God.

Reply 2: But appeal to the duty to be holy is not a good reason to believe (3) because there is an important difference between these duties. The command to be holy requires God’s grace to fulfil, but a person choosing their fundamental moral character in the original circumstance cannot be a recipient of divine grace.

Consider some background on Kant’s view of divine grace.⁹ Divine grace cannot improve human moral character without human free will: ‘To expect an effect of grace means, however, the very contrary, namely that the good (the morally good) is not of our doing, but that of another being – that we, therefore, can only come by it by doing nothing, and this contradicts itself’ (1996c, 6:53, italics in original). Two movements in human free will are essential for receiving divine grace. First, before receiving divine grace, human beings must have freely made themselves ‘worthy of this assistance’ (1996c, 6:52). To make themselves worthy, they must have fulfilled the duty to be holy to the extent that they are able:

Reason says that whoever does, in a disposition of true devotion to duty, as much lies within his power to satisfy his obligation (at least in a steady approximation toward complete conformity to the law), can legitimately hope that what lies outside his power will be supplemented by the supreme wisdom in some way or other (which can render permanent the disposition to this steady approximation) (1996c, 6:171, italics in original).

[T]here is no other means (nor can there be any) by which to become worthy of heavenly assistance, except the earnest endeavor to improve his moral nature in all possible ways, thereby making himself capable of receiving a nature fully fit – as is not in his power – for divine approval (1996c, 6:192).

Second, after the divine assistance comes, the human being must freely accept it:

Granted that some supernatural cooperation is also needed to his becoming good or better, whether this cooperation only consists in the diminution of obstacles or be also a positive assistance … he must accept this help (which is no small matter), i.e. he must incorporate this positive increase of force into his maxim: in this way alone is it possible that the good be imputed to him, and that he be acknowledged a good human being (1996c, 6:44, italics in original).

The original circumstance in which a person timelessly chooses a fundamental moral orientation rules out the possibility of divine grace. In that circumstance, human beings make their first free choice where ‘first’ refers to logical or sequential priority rather than a temporal priority. It follows, then, that they cannot have performed more fundamental free actions to make them worthy of divine assistance nor can they have performed earlier free actions to accept divine grace. But God provides divine assistance to make human beings morally good only if they have antecedently performed free actions that make them worthy of it and they have accepted it. Thus, since human persons cannot be worthy of divine assistance in the original circumstance and they must be worthy of it to receive divine assistance, human beings cannot receive divine assistance when choosing their fundamental moral orientation. Without such assistance, they are unable to choose the morally good propensity due to the limitations of human nature.

So then, even if there is a duty to be holy, it does not follow that there is also a duty to choose the morally good propensity precisely because human beings cannot be raised beyond the powers of human nature via divine grace to choose the fundamental good moral orientation. If there is no other reason to embrace (3), we should conclude that (3) is false. After all, we have strong reasons to embrace (1) and (2), and they are inconsistent with (3).¹⁰ Thus, ‘ought implies can’ provides no reason to doubt that human beings can freely choose the morally evil propensity precisely because there is no duty to choose the morally good propensity.

Therefore, the choice to be radically evil ensured by a necessary feature of human nature is consistent with Kant’s broader views on imputation, freedom, and the ‘ought implies can’ link between duty and freedom.

From radical evil to constitutive moral luck

Constitutive moral luck occurs when an agent possesses dispositions at least in part due to factors beyond her control and those dispositions positively affect her praiseworthiness or blameworthiness.¹¹ David Enoch and Andrei Marmor (2007, 426) helpfully highlight an ambiguity in this definition concerning whether the object of moral responsibility is a disposition or an action. We can disambiguate it according to these two objects of moral responsibility. First, direct constitutive moral luck occurs when a person’s possession of a disposition is influenced by factors beyond her control and those factors positively affect her praiseworthiness or blameworthiness for the disposition. For example, a person raised in an irascible household possesses irascible tendencies through habituation that is at least in part beyond her control. If she is blameworthy for being irascible, it is a case of direct constitutive moral luck. Second, indirect constitutive moral luck occurs when a person’s possession of a disposition is influenced by factors beyond her control and that disposition positively affects her praiseworthiness or blameworthiness for an action. For example, if a person is blameworthy for performing wrathful actions that are motivated by an irascible disposition that is possessed due at least partially to factors outside of her control, it is a case of indirect constitutive moral luck.

The Allison-style interpretation of radical evil definitely commits Kant to indirect constitutive moral luck. After all, a necessary feature of human nature beyond the control of each human being ensures that in the original choice scenario, each human being freely chooses the evil propensity. And it must be a choice for which a person is blameworthy given that human beings are blameworthy for its effect – namely, the morally evil propensity. Thus, interpretations of radical evil that explain its universality and innateness through a timeless free choice ensured by a feature of human nature are committed to indirect constitutive moral luck.¹²

The Allison-style interpretation of radical evil plausibly also commits Kant to the existence of direct constitutive moral luck. It is beyond a person’s control to some extent that they possess the morally evil propensity because human nature has so confined the options for free choice in the original circumstance to guarantee that each human possesses the morally evil propensity. So, even though it is a free choice to adopt the morally evil propensity, the choice is so shaped by luck that possession of the morally evil propensity is due at least partially to factors beyond the agent’s control. And human beings are also blameworthy for the morally evil propensity: it is ‘something that a human being can be held accountable for’ (1996c: 6:32), ‘we can well be responsible for the propensity to evil … imputed to the subject as itself guilty of it’ (1996c: 6:35), and it is something for which we have ‘innate guilt’ (1996c: 6:38). So then, the possession of the morally evil propensity is significantly influenced by factors beyond their control (namely, human nature) in a way that positively affects blameworthiness for it. So, Kant embraces at least some degree of direct constitutive moral luck in every human person. To my knowledge, no one else has argued that Kant embraces direct constitutive moral luck, and so these passages point to a completely new source of moral luck in Kant’s moral philosophy.

One might reasonably wonder about the coherence between Kant’s views of freedom and the morally evil propensity, on the one hand, and the contemporary category of luck, on the other. I consider and reply to two more objections.

Objection 3: It is incoherent to assert that a free action is lucky. Intuitively, an action is free only if its occurrence is not a matter of luck (Mele 2006; van Inwagen 2002). Radical evil is chosen as a matter of constitutive luck. Thus, the choice for radical evil cannot be free.

Reply 3: I reject the premise that ‘radical evil is chosen as a matter of constitutive luck’ – at least if ‘choosing as a matter of luck’ includes the normative property of ruling out freedom. According to Allison, if the decision proceeds from constitutive features of her reason and will in the right way – and the choice for radical evil does – the choice is up to the agent in the sense relevant to be a free action. Thus, the choice for the evil propensity must not occur as ‘a matter of luck’. But how, then, does luck figure into the choice? There is a plausible distinction between an event’s being a matter of luck and its being influenced by luck. A choice’s being influenced by luck means that factors beyond the agent’s control shape the manner in which the choice is up to the agent, for example, by influencing the choice options, incentives, and difficulties. In this way, the free choice to be radically evil is significantly influenced by luck but is not a matter of luck.

Objection 4: It is incoherent to assert that a necessary event is lucky. Intuitively, an event is lucky only if it could easily have failed to occur (Pritchard 2019) or if its occurrence is improbable (Rescher 2019). The necessary choice of radical evil could not easily have failed to occur and it is not improbable. Thus, the choice of radical evil is not lucky. If the choice of radical evil is not lucky, the choice of radical evil is not morally lucky. Thus, the choice of radical evil is not morally lucky.

Reply 4: I deny the premise that ‘if the choice of radical evil is not lucky, the choice of radical evil is not morally lucky’. That premise assumes that moral luck is a species of luck. But it is not. The phrase ‘moral luck’ is a misnomer; it is not about luck per se (Hartman 2017, 23–31; see also Anderson 2019; Statman 2019). Most philosophers writing on moral luck employ the lack of control conception of luck with the clear understanding that it is an implausible definition of luck itself.¹³ But lack of control is a paradigmatic feature of lucky events. And it is this paradigmatic feature of luck that is necessary and sufficient to generate the tension in our moral thinking that Bernard Williams (1981, 21–22) coined ‘moral luck’ to describe. The tension is in a person’s being praiseworthy or blameworthy for something that is beyond her control in some way. We can substitute ‘moral lack of control’ for ‘moral luck’, and nothing substantive would change about this debate. But if the absence of control is the feature of luck highlighted in a definition of moral luck, the problem dissolves; a lucky event can be necessary if what makes it necessary is beyond one’s control (Hartman 2019b, 3186). So then, the necessary choice to be radically evil is morally lucky because it is features of human nature about reason beyond their control that ensure the choice.

Conclusion

The received view about Kant is that his moral philosophy rules out all moral luck. But Kant’s claims about radical evil highlight previously unrecognized sources of indirect and direct constitutive moral luck in his moral philosophy.

Radical evil is a disposition that a person has necessarily and innately but for which she is blameworthy. If human beings are blameworthy for radical evil, they must have freely chosen it, but the free choice must be timeless for radical evil to be innate. Following Henry Allison, I argued that certain kinds of necessitated choices can be free because Kant is a sourcehood incompatibilist; alternative possibilities at the moment of choice are not necessary for a choice to be free. Additionally, I argued that a timeless, necessary, and free choice for the morally evil propensity does not violate ‘ought implies can’ because there is no duty to choose the morally good propensity in the original circumstance. Furthermore, I argued also that the free choice to have the morally evil propensity should be classified as ‘influenced by luck’ rather than occurring as ‘a matter of luck’ and that an essential feature of human nature can be a suitable source of ‘luck’ in ‘moral luck’ because the relevant definition of ‘luck’ is the absence of control.

Allison-style interpretations of radical evil commit Kant to the existence of two kinds of constitutive moral luck. The blameworthy free choice for the morally evil propensity shaped by constitutive features of human nature is a case of indirect constitutive moral luck. The blameworthy disposition (the morally evil propensity) possessed at least in part due to factors beyond the agent’s control is a case of direct constitutive moral luck.

This article strengthens the case against the received view that Kant’s moral philosophy precludes all moral luck by appealing to passages from the Religion. Thus, philosophers have yet another reason to reject the received view, which may also provide Kantians and Kant scholars an occasion to check their reasons if they deny moral luck.