Kant’s Diabolical Evil Reconsidered

1 Introduction

In RGV^[1] Kant uses the terms “diabolical” [teuflisch] or “devil” [Teufel] six times; two formulations, however, seem to capture the essence of Kant’s intent in this regard:

1. In order to find a ground of moral evil in the human being, sensibility entails too little; since it renders the human being, by removing the incentives which are rooted in freedom, to be a mere animal. However, a reason which is absolved from the moral law, a quasi [gleichsam] vicious [boshaft] reason (an evil will per se) entails too much, because thereby the collision [Widerstreit] with the law would be elevated to be the incentive (for without any incentive the power of choice [Willkür] cannot be determined) and this would make the subject a diabolical being. Neither of the two is applicable to the human being (RGV, 6: 35).^[2]

2. The viciousness [Bösartigkeit] of human nature is not so much malice [Bosheit], if one understands this word in its strict meaning, i.e. as a fundamental attitude [Gesinnung] (subjective principle of maxims), incorporating evil as evil into the maxim as the incentive (for this fundamental attitude is diabolical), but rather as a perversity of the heart, which is therefore to be termed an evil heart (RGV, 6: 37).

   In RGV, 6: 36 Kant addresses this same issue, however without explicitly mentioning the concept of “devil”:

3. The human being (even the basest [selbst der ärgste]) […] never renounces, as it were, the moral law in a rebellious manner (by the withdrawal of obedience).

Now it is clear that Kant is not so much interested in diabolical evil as such, but rather in human evil; the first designates the metaphysical horizon of the second and thus defines its intelligibility. Therefore, it would be mistaken to argue that diabolical evil is not a constitutive component of Kant’s moral theory and does not obtain a genuine philosophical value. Kant, indeed, did not have a philosophical interest in classic theological questions concerning diabolical beings (e.g. in the question of whether a being like the devil actually exists and tempts human agents to do evil, or in questions concerning the precise relation between the devil and God, or why God allows the existence of such a being to begin with and so on). Moreover: Kant’s usage of theological language and concepts in RGV deserves a separate clarification. Kant thinks that morality should provide the criterion for theology (e.g. the criterion by which the Bible is interpreted). Thus, when Kant raises the question of “whether morality must be interpreted according to the Bible or vice versa” (RGV, 110), his answer seems unequivocal: it is the Bible that should be interpreted according to morality (see also e.g. ST, 41: “reason is the highest interpreter of things related to religion [Religionssachen]”; this echoes Kant’s formulation in the first Critique, according to which he suggests “moral theology” and not “theological morality”; KrV B, 660). Concerning diabolical evil, this means that Kant is not committed to a specific Christian interpretation of this concept but rather conceptualizes this issue in light of his own independent moral framework. Thus, when Kant argues that it is impossible for humans to act on a specific (‘diabolical’) evil incentive, one ought first and foremost to elicit and clarify the metaphysical limitation on human freedom implied in this determination.

Before delving into details, let me first say something general about metaphysical differences in the context of Kant’s moral theory. Kant’s morality is based on the determination that human beings are “finite rational beings” (KpV, 5: 26), meaning, “rational beings in nature” (KU, 5: 435). These definitions imply that the human being is neither divine – it is “finite” and “in nature”, i.e. contrary to God it is not a “purely intelligible being” (Prol, 4: 344) and cannot be attributed “the utmost perfection” (KpV, 5: 41) – nor is the human being animalistic, for it is “rational”, i.e. it is defined as having access to the moral law.

In both cases the difference is metaphysical, i.e. it concerns the respective structure of the will, which cannot be altered by the agent itself according to Kant. For (i) the animalistic lack of moral incentive (of access to the moral law) is a matter of definition and not an outcome of a free act (an animal cannot decide to gain access to the law). In the same spirit (ii) a divine being cannot be ascribed sensual incentives by definition (see my discussion in Section 2). The same holds true (iii) for human beings, for neither their moral incentive nor their sensual incentive is an outcome of free choice; the first “forces itself [sich aufdringt]” (KpV, 5: 31) upon them, and the second concerns the fact that they find themselves interwoven in (an already existing) “nature”, so that it is impossible for humans to “eliminate” (RGV, 6: 28) one of these two constitutive incentives.

Both human and animalistic beings are finite beings; thus, the metaphysical boundary is drawn within finitude (the first finite being obtains access to the moral law, the second – does not). In contrast, the difference between human beings and God concerns a distinction between a finite and a perfect being, which possesses “holy will” (KpV, 6: 32). Now how are we to understand the difference between a diabolical and a human being?^[3]

The answer is not clear at first glance. Surely this difference is not to be conflated with the difference between human and divine, for a diabolical being is not a perfect (holy) being, whose will is determined “of itself […] only by the representation of the good” (GMS, 4: 414). Nevertheless, the difference between human and diabolical is not to be identified with the difference between human and animalistic, since a diabolical being obtains per definition a relation to the moral law^[4] – it is defined as a being that rebels against it, thus essentially retaining a relation to the law.^[5] This renders the task of drawing a metaphysical boundary which concerns the formal structure of the will to be highly challenging. Why?

It is unclear at first glance why it is impossible on structural grounds to ascribe to the human being the incentive of “evil as evil” (or “rebellion against the law”, i.e. “collision with the law”), for ‘rebellion’ cannot mean that a diabolical being acts in light of an evil law; such a law is absolutely impossible in Kant’s philosophy (even for God). It is e.g. possible not to pay back a loan; however, it is utterly impossible to render the rule ‘you should not pay back a loan’ to be a universal law, for generalization here, according to Kant, leads to contradiction. This is why Kant says in MS that such a being (a devil) “rejects the authority of the law itself, whose validity he still cannot deny before his own reason” (MS, 6: 320; my emphasis), and this is the ground of the distinction Kant makes in RGV between corruption of reason itself, which is “absolutely impossible” (RGV, 6: 35), and the incentive “collision [or conflict; Widerstreit] with the law”, which is only impossible for human beings. So both human and diabolical beings acknowledge the validity of the moral law, and both can reject the law, however not in the name of a different – objectively valid – law; the difference is that it is claimed that a specific incentive (‘rebellion’, ‘evil as evil’) is essentially denied to an evil human agent.^[6] However, what can be the structural grounds for such a denial? An evil human agent can reject the law on many grounds (see the discussion in Section 3); eliminating one of them on principal reason must be accounted for structurally.

I will continue as follows: In Section (2) I will focus on central aspects of the structural difference between human and divine beings. In the subsequent two sections I will address the issue of the possible structural difference between diabolical and human wills in two steps: in Section (3), which also entails a discussion of the literature on the issue, I will first explain why the task of finding a structural difference between human and diabolical will is especially difficult; I will then present a possible new solution to the problem. After explaining why I think this solution eventually fails, I will present in Section (4) a second new possible solution that works. Section (5) entails a short concluding remark.

2 Human and Divine Will

A rational finite will consists of two components; the first concerns the “will of a human being, insofar as it is affected by nature” (GMS, 4: 387). The will here is merely affected by nature, but “not determined” (MS, 6: 213) by it. The second component concerns the “consciousness of the moral law” (KpV, 5: 121), which ought to determine our will. Notice that for the sake of the possibility of freedom of choice none of these components determine the will; the first affects the will, the second ought to determine it. These two components serve as a basis for two incentives: ‘self-love’ (see e.g. RGV, 6: 26–27)^[7] and “respect for the moral law” (see e.g. KpV, 5: 73; 75–76; 78; henceforth: ‘respect’).

In light of this, the ‘power of choice’ [Willkür] of a finite rational being has two tasks: first to elevate “both incentives in the same maxim” (RGV, 6: 36) and then to decide upon the inner “subordination” (RGV, 6: 36) within the maxim, meaning, to determine which incentive of the two will serve as the “highest maxim” (RGV, 6: 36). So rational finite beings – good or evil – always relate to both incentives (the moral and the sensual), and can “eliminate” (RGV, 6: 28) neither of them.

As a part of his discussion on “predisposition to personality” in RGV, 6: 27–28 Kant introduces a further subtlety, which will be crucial for my argumentation in the next section. According to Kant, since the incentive of ‘respect’ is ascribed to every human being, “even the basest” (RGV, 6: 36), then every human being must have receptivity for it. This receptivity has two senses:^[8] the first is “the receptivity of mere respect for the moral law” (RGV, 6: 27); the second is “the receptivity for the respect for the moral law as a sufficient incentive of the power of choice” (RGV, 6: 27). This distinction is important, since obtaining a mere receptivity for ‘respect’ is not sufficient for the possibility of a moral action in the full sense of the word, in which ‘respect’ must also serve as the sufficient (the subordinating) incentive of the power of choice.

Now how are we to understand the structural difference between human and divine will? The following citation seems to capture the essence in this regard:

An absolute good will [….] can be determined of itself according to its subjective constitution [Beschaffenheit] only through the representation of the good. No imperatives can therefore be applied to divine will, and generally to holy will; the ‘ought’ is out of place here, because the will already and of itself necessarily corresponds with the law. Imperatives, thus, are only formulas which express the relation of objective laws of willing in general to the subjective imperfection of the will of this or that rational being, e.g. the human will (GMS, 4: 414).^[9]

The differences can be summarized in the following manner:

1. Divine will is determined only by one component (the representation of the good); human will obtains in contrast always two incentives (‘respect’ and ‘self-love’). Thus, no freedom of choice can be attributed to divine will.^[10] However, according to Kant’s celebrated determination that “a free will and a will under moral laws are one and the same [einerlei]” (GMS, 4: 447; see also KpV, 5: 29), ‘perfect freedom’ does not necessarily entail ‘freedom of choice’ (in fact, the first excludes the second), but rather: a ‘perfectly free’ will is one that is necessarily determined by reason. Put differently, the possibility of evil cannot be attributed to God.

2. Given that a divine will is determined only through the representation of the good, it follows that it has no relation whatsoever to sensuality;^[11] it is not “tempted [or misled; verleitet] by sensual inclinations” (MS, 6: 380).

3. The moral incentive appears for a finite rational being in the form of imperative, precisely because the moral incentive does not already determine of itself the will: “All imperatives […] designate the relation of an objective law of reason to a will […] which is not necessarily determined by this law” (GMS, 4: 413).

Notice that it would be misleading to argue that these differences refer solely to the quantity and not the quality of the incentives, as if a human being obtains two incentives (sensual and moral), whereas a divine being obtains only one incentive (the moral). Given that divine will is not to be ascribed any additional determining component apart from the moral one, it cannot be ascribed ‘respect for the moral law’ in the first place, for ‘respect’ is a predominant property of finitude:

It ought to be noticed that respect is an effect on feeling, and thus on the sensibility of a rational being; thus, it presupposes the sensuality and the finitude of such beings on whom the moral law imposes respect. Therefore respect cannot be attributed to a supreme being, or even to a being which is free from any relation to sensuality […] (KpV, 5: 76).

Thus, divine will is not determined by the same component which ought to determine the will of finite rational beings (‘respect’), but rather it is determined by a categorically different component (‘representation of the good’). This is precisely the reason why even the concept of ‘incentive’ cannot be attributed to divine will:

[O]ne can ascribe a divine will no incentives whatsoever […] [for] by incentive […] one understands the subjective determining ground of the will of a being whose reason does not already by its very nature correspond with the objective law (KpV, 5: 71–72).^[12]

To illustrate: consider a finite rational agent who throughout its entire moral life with no exception has always chosen the good. Can it be said that its will is holy (divine)? No, on structural grounds: this agent must have always subordinated anew (for it was each time confronted with the possibility of evil) the sensual incentive to the moral incentive. In contrast, God does not subordinate incentives, and cannot be ascribed the possibility of evil (or ‘freedom of choice’), and the moral component which already determines its will is of a different kind (‘representation of the good’ and not ‘respect’). Due to structural limitations it is impossible for human will to be divine.

3 Human and Diabolical Will: The Second Receptivity

Now a diabolical will is ascribed an incentive (“evil as evil as an incentive”, “collision with the law” etc.) which cannot possibly serve as an incentive for humans. This impossibility of the ascription of diabolical incentives to humans must be accounted for in structural terms.^[13]

Before explaining in detail why this challenge is highly unique, we can already reject positions which treat this issue empirically. Silber writes e.g. that Kant’s thesis ought to be dismissed because “man’s free power to reject the moral law in defiance is an ineradicable fact of human experience”;^[14] elsewhere he determines in the same spirit that “Kant’s ethics is inadequate to the understanding of Auschwitz”.^[15] These allegations are highly misleading; no historical (or future) event can change the formal structure of a finite rational being, just as the morally good deeds of the so called ‘Righteous Among the Nations’ in the same historical context which Silber mentions do not indicate that such peoples’ wills became divine. The same holds true for Pasternack’s position, according to which “the exclusion of diabolical evil may, however, seem naïve. Given the litany of horrors that humans have perpetrated, explanations that come down to mere self-interest may seem just too languid”.^[16] The possibility or impossibility of the ascription of an incentive to finite rational (human) beings concerns solely the structure of their will, and can never be decided upon on empirical grounds.

Due to similar reasons I think that Bernstein’s thesis, according to which Kant’s claim must be rejected because it is rooted in Kant’s “limited moral psychology”,^[17] is misleading, as well as Louden’s determination that Kant excluded diabolical evil from the realm of human freedom because he seeks to avoid romanticizing evil – that is, that according to Kant “we must not be seduced […] into thinking that some human beings, by virtue of their ‘strong’ and ‘potent’ personalities, are somehow above the rest of us, and are not to be judged by the same laws and principles that apply to ordinary human beings”.^[18] Kant seeks metaphysical criteria to designate human beings as (moral) species – independent of the question of whether a specific agent is strong or weak, beautiful or ugly, powerful or impotent.

The main problem in such readings, I think, lies in the implied assumption that the meaning of ‘in defiance’ is already known, and on account of this preceding understanding, Kant’s position is held to be naïve. But Kant’s stance must be accounted for in Kant’s own terms, i.e. as a structural account according to which an agent cannot by definition reject the moral law in defiance, regardless of historical circumstances. Thus, the interpretative task is to decipher what ‘in defiance’ for Kant structurally means, so that it would entail ‘Auschwitz’ as well. And as I will show, such an account exists.

Now what does the (exclusively diabolical) incentive ‘evil as evil’ (or ‘collision with the law’), in which the validity of the moral law is nevertheless acknowledged, mean? A first possible explanation concerns the ‘predisposition to personality’ which was mentioned in Section 2. Notice that according to Kant, ‘predispositions’ constitute the structure of human will (“they are original, since they belong to the possibility of human [moral] nature”; RGV, 6: 28); thus, if one wants to find a structural explanation for the exclusion of diabolical incentive from the realm of human evil, ‘predispositions’ is a good place to start.

Recall that I mentioned that in RGV (but not in KpV or KU), as a part of his discussion on predispositions to the good, Kant introduces two senses of the receptivity for the moral law: the receptivity of mere respect for the moral law, and the receptivity for the respect for the moral law as a sufficient incentive of the power of choice. In fact, Kant describes here two different kinds (or levels) of relation (access) to the moral law. Let us raise now the following speculation, bearing in mind that ‘rebellion’ must still retain by definition a relation to the moral law: A diabolical being might obtain only the first but not the second kind of receptivity for the moral law. The reason for this possible state of affairs might be twofold: either (i) a diabolical being intentionally (freely) gave up the second relation to the law,^[19] or (ii) a diabolical being is defined as having ab ovo only the first and not the second kind of access to the moral law.

1. If renunciation as an outcome of a free accountable decision is the case, it must first be explained how such an (evil) act of renunciation can be thought of in Kant’s philosophy.

   To grasp this, it is crucial to distinguish between form and ground of an evil act. Here is a typical citation:

   The human being […] is evil only because it reverses the moral order of the incentives when incorporating them into its maxim: the moral law is elevated indeed to the maxim alongside the law of self-love, however when it becomes aware that the one cannot subsist alongside the other, but rather that one must be subordinated to the other as its supreme condition, it renders the incentive of self-love and its inclinations to be the condition of compliance with the moral law […] (RGV, 6: 36).

Notice that Kant suggests here merely a description of evil action, but not an account of the determining ground which generated this reversed order. The description seems clear enough: an evil action is an action in which the incentive of ‘self-love’ subordinates the incentive of ‘respect’ (whereas a morally good action is an action in which ‘respect’ is the subordinating and not the subordinated incentive). However the ground for why an agent brought about the evil (reversed) order instead of the morally good order can be diverse; one agent might decide to reverse the right (moral) order of the incentive due to boredom, a second agent out of curiosity to experience what an evil action feels like (since evil fascinates), a third due to weakness of the will, a fourth due to habit, laziness, spite,^[20] and so on.^[21]

Thus, when Kant asks “which of the two [incentives] does he [er; meaning, the free agent] make the condition of the other” (RGV, 6: 36), the word ‘he’ plays a crucial role; it means that neither ‘respect’ nor ‘self-love’ alone serves as the determining ground of the will. Rather, the determining ground of the will is the free decision of the agent to elevate ‘self-love’ or ‘respect’ to be the subordinating incentive. The ground for the decision concerning the order of the incentives is not to be identified with the two incentives themselves, or else the human agent would be a mere marionette, whose will is determined sometimes by ‘self-love’, sometimes by ‘respect’, but never through its own free decision.^[22]

Now we can better conjecture what a (possible) diabolical renunciation of the second kind of receptivity for the moral law might mean. To grasp this, notice first that it is absolutely possible for a rational finite being to want to change the structure of its will, since no structural ground prevents it from wanting it. (No structural ground prevents us e.g. from wanting or longing to be divine, or from longing to be devoid of inclinations;^[23] structural grounds only prevent us from becoming divine, or from becoming devoid of inclinations).

Now consider a being which is originally ascribed the same structural relation to the moral law like us – it obtains the first kind of receptivity for the moral law as well as the second kind and thus possesses the full possibility to obey the law. At some point it wants to release itself from the troublesome burden of the demands of the moral law which it loathes, i.e. it wants to alter the structure of its will. It still wants, however, to retain a vital relation to the moral law, for it wants to perpetuate and to steadily recall so-to-say its rejection of the law (that is, it does not want to become merely animalistic). This can be done by renouncing the second but not the first kind of receptivity for the law, wherefore this being still retains a relation to the law but renounces the possibility to render the moral incentive to be the sufficient incentive of its will. This is one possible structural meaning of ‘rebellion’ against or ‘collision’ with the law, which yields a genuine structural difference between human and diabolical will: Human will obtains two kinds of receptivity for the law; diabolical will originally obtained these two kinds of receptivity, but after renouncing the second, now obtains only the first kind.

However, the possibility of altering structures renders the very term ‘structure’ in the Kantian context empty. The terminology Kant uses time and again in this regard attests to it – it is impossible for a human evil agent to obtain a diabolical incentive, impossible for a human good agent to be divine, impossible for divine to be human (for the ‘possibility of evil’ cannot even be ascribed to God), impossible for a human agent to become animalistic (i.e. to renounce the consciousness of the moral law) and so on. The possibility of altering structures cannot be seen as a structural component; rather, it negates the very concept of ‘structure’ in the Kantian context.

1. If a diabolical being is defined as having only the first and not the second kind of access to the moral law, then the original impossibility to act morally yields the structural meaning of ‘rebellion’. In this case a diabolical rebellion against the law is lawful, i.e. it serves by definition as a determining ground of a diabolical will.^[24] Now ‘lawful’ is not to be confused with ‘always’; if rebellion defines the structure of a diabolical will, then refraining from rebellion must be considered impossible. In this case, however, we can ascribe to diabolical will neither ‘freedom of choice’ (human freedom), nor ‘perfect freedom’ (divine freedom). In Kant’s philosophy there is only one case in which a will is considered free despite the fact that it is defined as already being determined by a certain component: a holy (divine) will. But if the devil cannot be attributed any sense of freedom whatsoever, what makes it evil? Such a view contradicts^[25] the very concept of diabolical evil.^[26]

4 Human and Diabolical Will: The Question Regarding Rationalization

In order to see if there is nevertheless a way to maintain the structural difference between human and diabolical evil without excluding ‘diabolical evil’ from the realm of freedom,^[27] it might be useful to systematically examine sections in which Kant speaks of the primordial reaction of human beings as a moral species to the moral law;^[28] for even if the concept of diabolical evil is not mentioned in these sections explicitly, it pertains directly to the issue at stake.

One revealing formulation in this regard can be found in Kant’s discussion on the “origin of evil in human nature” (RGV, 6: 39ff.); notice that Kant stresses time and again that he is referring to the human “considered in its species” (RGV, 6: 32; see also RGV, 6: 20; 21; 25; 29). In his account on this issue Kant at some point describes the “way of presenting [Vorstellungsart]” this issue in the ‘Scripture’ (RGV, 6: 41).^[29] Before delving into this issue, let me say something about the systematic status of this presentation. Kant says that the Scripture describes the “beginning of the origin of evil in the human species” (RGV, 6: 41). By determining this, Kant is neither (i) interested in analyzing the falling into evil of a specific agent (the First Adam), nor (ii) does he raise a causal claim (as if the fall of the First Adam into evil caused the evil of all other free agents – each agent accounts for its own deeds according to Kant). Rather, the Scripture presents how finite agents considered as a species react to the moral law. Considered as a species does not mean (iii) that the species itself is reified and treated as an accountable agent; it means that Kant thinks that all agents freely react to the moral law in this primordial context in the same manner (radical evil), i.e. that “the grounds that justify ascribing it [freely acquiring an evil rather than a good fundamental attitude] to one person are constituted in a way that there is no reason to exclude any other human being from it” (RGV, 6: 25, my emphasis). So according to Kant the reaction to the law, though rooted in freedom (i.e. essentially contingent), can nevertheless be proven^[30] to be attributed to the entire species, meaning, to every agent without exception.^[31] Now how is this primordial reaction to the law (radical evil) described by Kant?

The point of departure must be thought of as the “stage of innocence [Unschuld]. The moral law preceded, as a prohibition, the human being, which is not a pure being but a being which is tempted [versucht] by inclinations” (RGV, 6: 42). So no determination of the will has taken place yet. What happens next? How does the agent react to the demand of the law? Kant writes (I have marked the three relevant verbs):

Instead of simply to follow this law as sufficient incentive (which is alone unconditionally good, thus there is no place for further deliberations [or qualm; Bedenken], the human being [i] has looked around [umsehen] for other incentives […] which can be good only conditionally […] and has made it – if one thinks of the action as arising consciously from freedom – a maxim to follow the law of duty not out of duty but at best [allenfalls] out of consideration for other aims. As a consequence the human being began to doubt the strictness of the command which excludes the influence of any other incentive, and afterwards [ii] to rationalize-in-order-to-downgrade [or to quibble; herabvernünfteln] the obedience to be a mere (under the principle of self-love) conditional obedience as a means, and eventually the preponderance of the sensual impulses over the incentive rooted in the law [iii] was elevated [aufgenommen] in the maxim and thus sin came about […] (RGV, 6: 42).

The three verbs Kant uses here correspond with the two components of an evil act; the third verb (to elevate [aufnehmen] ‘self-love’ over ‘respect’) is a description of an evil act. The first two constitute the ground of the evil act.^[32]

RGV is not the first work in which Kant uses the verb ‘vernünfteln’ to describe a similar mechanism; In GMS, 4: 405 for instance, one can find an almost identical formulation (however, without distinguishing ‘looking for’ as preceding ‘rationalizing’). Kant speaks there of a “natural dialectic”, i.e. of the

propensity [Hang] to quibble [vernünfteln] against these strict laws of duty and to cast doubt upon their validity, or at least upon their pureness and strictness and to make them, where possible, more suitable to our wishes and inclinations, that is, to corrupt them in their ground […].

It seems that the mechanism Kant describes (‘rationalization’) has something to do with self-deception;^[33] self-deception, however, is not that easy to conceptually account for, as Kant well knew (see MS, 6: 430).

In the last few years there has been increasing interest in this issue in the literature.^[34] To give a few representative examples: Green thinks that all three kinds of “natural propensity of humans to evil”, and not only the third kind, represent a form of self-deception.^[35] In contrast, Rukgaber argues that the first kind, i.e. “Kantian frailty [Gebrechlichkeit] is a form of practical irrationality without self-deception”.^[36] Welsch, as well, rejects Green’s position and holds that only the third kind of original evil can be identified with lying to oneself.^[37] Papish says that this original act cannot be fully identified with self-deception, but nevertheless a “dissimulation” is involved.^[38] Callanan says that the natural dialectic mentioned in GMS is “a kind of self-deception whereby the subject undermines morality while still paying lip service to it”^[39] and concludes that according to Kant, and contrary to Rousseau’s position, reason is not responsible for this self deception.

Now the question concerning the adequate positive understanding of rationalization in this context is of no systematic relevance to the issue at stake; much more important is the negative aspect of Kant’s description. For whatever rationalization means, it does not concern a direct rejection of the law; rather, what the agent does is a bit different: facing the demand of the law, the agent “has made it […] a maxim to follow the law of duty [however] not out of duty” (I am interested here solely in the outcome of the process Kant describes in the long citation above from RGV, 6: 42, and not in the stages leading to it, such as beginning “to doubt the strictness of the command” etc.). From this we learn that the primordial confrontation with the law consists of two possibilities: the agent either (i) follows the law (the moral possibility) or (ii) begins a process of rationalization.

Notice that these two possibilities of the primordial reaction to the moral law described by Kant are not constituent components of the structure of the will; it is not the structure which is at stake here, but the free (contingent) usage of it. However, the formal structure of the will and possible free usages of it shed light on each other. That we can infer from the structure of the will to possible usages of it is clear enough, I think; for instance, based on the structure of the will of a finite rational being as obtaining per definition two incentives (sensual and moral) we know that neither a divine nor an animalistic usage of the will is possible for human agents. But a structural conclusion can be inferred from possible (or impossible) usages of the will as well. If the primordial reaction to the moral law – besides fully adhering to it – is rationalization and not rejection, it must be rooted in an essential limitation that is attributed to us. Which limitation might Kant be implying here?

The answer lies in the manner in which the law is given to humans. Kant’s systematical position is that in the realm of morality explanations concerning emergence count as transgressing the limits of cognition. This is why Kant terms the consciousness of the moral law an “inexplicable fact [unerklärliches Factum]” (KpV, 5: 43, my emphasis). Similarly, Kant says that reason “imposes itself” (KpV, 5: 85) upon us, and determines further: “how the consciousness of the moral law […] is possible cannot be further explained” (KpV, 5: 46; see also KpV, 5: 72). Elsewhere Kant maintains that “the law finds on its own entrance [or access; Eingang] into the mind [Gemüth]” (KpV, 5: 86; my emphasis). Thus, any explanations concerning the emergence of moral consciousness are impossible; the awareness of the moral law without knowing its origin is an original irreducible fact of our existence as rational beings and serves as an ultimate point of departure in this regard. The point of departure of morality for finite rational beings entails therefore an essential epistemic limitation.

Now let us reconstruct a possible hypothetical being which is, just like a finite human being, confronted with the same law, but is epistemically superior in comparison to humans in the sense that it can cognize the way of emergence of the moral law – i.e. it sees where the law comes from so-to-say, its ultimate origin (God; see my discussion in Section 5). Such a being, if it decides to reject the law, cannot rationalize (in the sense mentioned above), for we can rationalize and deceive ourselves only concerning something which we cannot tell its ultimate origin; only a being which is epistemically inferior in this regard might try to convince itself that the law cannot come into force without its subjective agreement. For this is what the mechanism of rationalization is actually all about; it relates the unconditional to one’s own subjective will, i.e. it interprets the demand of the law by means of one’s subjective inclinations. The essential lack of knowledge concerning the origin of the law enables such a mechanism. A being which is epistemically superior in this sense can only reject the law directly (call it ‘collision’ or ‘rebellion’) and not by means of rationalization and quibbling. The impossibility to attribute diabolical evil to humans is rooted, thus, in the essential epistemic superiority^[40] of a diabolical being.^[41]

Such a hypothetical being can also be ascribed freedom of choice, but it is categorically different from humans. This holds true for the (i) primordial decision for evil, i.e. before evil took place; both kinds of beings are confronted with the law; the human can either follow the law (then it is morally good), or set up a rationalization process which ends up in the reversal of incentives (then it is evil). The diabolical, due to its superior epistemic status concerning the origin of law, can either rebel against it (in this case it is diabolical), or adhere to it (then it is a good angel). But no less important is that (ii) this holds true for the possibility of restoration of the good as well. For both kinds of beings restoration is possible, but categorically distinct. As for humans, they are “beings-in-the-world [Weltwesen]”, thus subjected to “continually becoming” (RGV, 6: 74–5). This is why the departure from original evil for humans concerns an endless, ongoing process. A diabolical being on the other hand has a direct relation to the law; for such a being, thus, rehabilitating the good (if it freely decides to alter its rebellious relation to the law) does not involve a process at all; rather, its goal is immediately achieved, with no need of mediating links whatsoever.^[42]

Let me address five possible misconceptions.

Firstly (i), the human agent recognizes the authority of the law as well as the validity of the law independently of the question concerning its origin. (For instance, the moral law binds atheists as well,^[43] and the validity of the moral law concerns the concept of contradiction and not knowledge of its origin). Not knowing the origin of the law is simply another aspect of human finitude (along with possessing sensual incentive next to the moral incentive), an aspect which changes nothing regarding Kant’s familiar account of moral action: a morally good action still means subordinating the sensual to the moral incentive without knowing the last metaphysical origin of moral consciousness, and a morally evil action still means subordinating the moral to the sensual incentive without knowing the last metaphysical origin of moral consciousness.

Secondly (ii), it is not claimed here that a human agent would be morally better or worse if it knows the origin of the law. Knowing the origin does not render a being better or worse, but categorially different; to wit, it enables either a rejection of the law in a way which is impossible for humans, or an acceptance of the law in a way which is impossible for humans. Put differently, humans would not earn or lose moral points if they obtained cognition of the origin of the law; they would simply be different kinds of beings.

Thirdly (iii), my account explains what ‘rebellion’ (or rejection of the law ‘in defiance’) in Kantian terms means. Rejection of the law while knowing its origin (the diabolical case) is ‘rebellion’ according to Kant (this, and not our preconception of this concept). Rejection of the law while not knowing its origin (the human case) is ‘rationalization’ (this, and not our preconception of it). Now it is clear why assertions such as Silber’s (“Kant’s ethics is inadequate to the understanding of Auschwitz”) or Pasternack’s (“the exclusion of diabolical evil may, however, seem naïve, given the litany of horrors that humans have perpetrated”) and other positions I have cited are misleading. In Kantian terms, Auschwitz changes nothing in the definition of ‘rebellion’ against the law, for no one would seriously argue that Goebbels for instance^[44] was not a finite human agent, i.e. that he possessed knowledge of the origin of the law. Kant could not possibly have known many things about future developments (the industrialization of death like that which took place in Auschwitz, for instance, was surely beyond his intellectual and scientific horizons); however, this has nothing to do with Kant’s structural claim concerning the epistemic limitations of finite beings as such.

Fourthly (iv), one might claim that of all things, knowing the origin might render ‘rationalization’ easier rather than impossible. But this cannot be the case here, for two reasons.

The first reason (a) concerns the inability of a diabolical being to deceive itself by rationalization; a diabolical being is defined by the knowledge of the divine origin of the law, i.e. it is defined by not being able to ignore it (and by ignoring it – to downgrade it); if a diabolical being starts ignoring the origin of the law, this strictly amounts to altering the very structure of its will. Put differently, knowing the origin of the law is not a temporal achievement or an accidental property of a diabolical being; it is a constitutive component of its being. Thus, it obtains by definition no leeway which can pave the way for ‘rationalization’.

The second reason (b) concerns the inability of a diabolical being to deceive God by rationalization. To illustrate: Say that my wife is worried about our not saving enough for retirement and thus opposes my buying a new car, and say that I know the origin of her anxiety in this regard (her family was poor). Knowing the origin of her anxiety is precisely what enables me to rationalize buying a new car – say, by pointing out that we have enough savings, or by deceiving her about our financial state. However, Kant’s point about diabolical beings and God is categorically different, for it does not concern the relation between two finite beings; a diabolical being cannot deceive God, for God knows by definition that the devil knows the origin of the law (omniscience is a divine attribute; see e.g. KU, 5: 444), and knows further that the devil cannot ignore this knowledge. The distinct metaphysical context makes all the difference.

Given these two points, it is clear why on Kant’s account ‘rationalization’ can be ascribed only to humans.

Fifthly (v), one might ask what is the ground according to which diabolical beings act. To wit: A human rejection of morality involves reference to some end that the will aims to promote (rejecting ‘respect for the law’ in order to promote ‘self-love’). However, what is the end which a diabolical being, defined purely negatively (in Kant’s formulations, “evil as evil as an incentive”, “rebellion against the law”, “collision with the law” etc.), wishes to promote by rejecting morality? Recall that, as I have indicated at the outset, diabolical evil is according to Kant “an unreachable ideal” (V-MS/Vigil, 27: 691), i.e. the “idea of extreme evil” (MS, 6: 321–322); and ‘ideas’ in Kant’s philosophy are not incoherent, but only unreachable for human beings. Thus, if rejection of morality involves reference to some end, this must be applied to a diabolical rejection as well.

Notice that unlike human beings, a diabolical being does not obtain a sensual aspect (see footnote 26) – meaning, it is not a being which is already embedded per definition in a sensual ‘worldly’ context (it is not Weltwesen; see e.g. KU, 447), a context in which, among other things, other humans play a constitutive role as well (“comparing [vergleichende] self-love”, RGV, 27). Therefore, whatever the non moral aim a diabolical being wishes to promote, it cannot be thought of as preferring ‘self-love’ in the human (sensual or inter-subjective) sense; a diabolical rejection of morality must be thought of as devoid of reference to human ‘self-love’ interests of any kind. It follows that the utter destruction of morality – an aspiration that can be ascribed only to a being which is confronted with the origin of the law – serves as its ultimate incentive.

5 Concluding Remark

The challenge of drawing a metaphysical boundary between human and diabolical evil discloses, I think, a very deep feature of Kant’s moral theory which is often ignored or misunderstood. Kant’s moral theory is usually praised (or condemned) for its formalism, which is understood, among other things, as a claim that morality is essentially independent of the concept of God. However, the precise sense of this independence is not always clear. The fact that according to Kant finite rational beings are expected to obey the moral law independently of their belief in God stands beyond any doubt, just as Kant’s claim that the touchstone of whether a certain practical rule might serve as law is not the question concerning its conjectural divine origin, but the possibility of its universalization without contradiction. However, taking the general framework of Kant’s morality into account and the metaphysical differences constituting it sheds a more subtle light on the inherent relation between morality and divinity.

In a somewhat paradoxical manner, morality as an independent (autonomous) project of finite rational beings is only possible given an essential cognitive limitation, namely that knowledge of the ultimate origin of the point of departure of morality (the consciousness of the moral law) is not accessible to them. This limitation constitutes the horizon to which we are unavoidably led from morality itself, as Kant well knew: “morality leads inescapably [unumgänglich] to religion” (RGV, 6: 6; my emphasis) (or: “the moral law [leads] […] to religion”; KpV, 5: 129), i.e. to “observing [Beobachtung] all human [moral] duties as divine command [Gebot]” (RGV, 6: 84). This determination is not meant to designate contingent, weak agents, who are in need of some relation to divinity in order to grasp (or to implement) their moral duty; rather, this is an essential contention about the situatedness of morality itself, i.e. about the inescapable context in which morality sprouts. In this sense morality is not just about formal duties, but rather about the “inescapable” transition from ‘morality’ to ‘religion’,^[45] i.e. from the consciousness of the moral law to its (divine) origin, to which we can only practically (endlessly) approach but never theoretically cognize. More than the structural differences between human and animalistic will and human and divine will, which are relatively easy to grasp, it is the unfolding of the intricate structural background that renders diabolical evil impossible for us which ultimately reveals the essential framework of Kant’s moral theory as a whole.