1.1 Conceptual Contours

Many children like marshmallows. In Walter Mischel’s classic studies on delayed gratification, children were asked to resist the temptation to eat a marshmallow when the researcher left the room; they were told that those who resisted this temptation would receive two marshmallows instead of one.Footnote ¹ Self-control was crucial to realizing the goal of receiving two attractive treats. What is important for our purposes is that this study acknowledges the importance of prudential self-control, which presupposes more-or-less given goals.

Contemporary moral psychology has rightly acknowledged this instrumental kind of self-control, but this has led to the downplaying of another significant form of self-control: resisting the temptation to lower one’s moral standards when setting oneself goals.Footnote ² However, although ruthless dictators can be very self-controlled in pursuing their evil goals, the self-control they lack has to do with the temptation to settle for lower standards when adopting their goals. This example suggests that self-control also plays a crucial role in ethical goal-setting. This role may be highlighted by fully clarifying self-control as the ability to redirect attention. Numerous empirical studies confirm the close tie between self-control and attention. To my knowledge, no attempts to explain the relationship between the activity of end-setting and self-control, conceived of as redirected attention, can be found in the available literature on contemporary moral psychology.Footnote ³

In this Element, I address Kant’s conception of moral self-control by interpreting it in terms of abstraction, understood as the activity of refocusing attention. This interpretation finds support in textual sources. It covers prudential and instrumental self-control but primarily aims to provide a fuller account of the self-control characteristic of Kantian virtue. Instrumental and prudential self-control do not exhaust Kant’s notion of moral self-control. Kantian moral self-control should not be downgraded to a mere instrument for compelling ourselves to act on the basis of already adopted principles and ends. It is required not only for realizing predefined ends but also for setting ourselves moral ends. I will try to show that self-control, as a psychological condition, can be central to Kant’s conception of virtue – essentially conceived of as free self-constraint in end-setting. In order to do so, I will first try to situate Kant’s notion of self-control in his moral theory.

One commonly found, and to a considerable extent tenable picture of the relationship between Kant’s moral theory and his moral psychology, can be sketched as follows.Footnote ⁴ Try to imagine a moral theory according to which most of our psychological conditions, such as our instincts, natural feelings and desires, are obstacles we must overcome if we are to become virtuous. In this theory, moral agency is broadly conceived as a constant struggle with these psychological obstacles: acting morally requires that we prevent these subjective conditions from influencing our way of thinking. The underlying thought is that practical reason, when determining what is morally right or wrong, should be free from all sensible and personal influences. Only some psychological conditions, such as our cultivated sympathetic feelings, moral feelings, self-control and conscience, may be considered aids to morality, but even these subjective conditions are mere tools that help us to perform moral actions. Such conditions become useful only once we have discerned right from wrong and adopted moral principles – they can be means to observing maxims, but not their necessary conditions.

Self-control rightly plays a crucial role in this picture. Kantian moral agency is often conceived as a struggle with instincts, feelings and inclinations, and dealing with these psychological obstacles requires self-control. The capacity for controlling sensibility (CPrR 5: 159), or the capacity for mastering inclinations (MM 6: 383), is central to Kant’s account of virtue. He has been reported as saying that virtue means ‘strength in controlling oneself [Stärke in der Selbstbeherrschung]’ (L-Eth 27: 300). When explaining what virtue as moral strength involves, Kant writes that we must assume that all of us have ‘the capacity (facultas) to overcome all sensible impulses’, suggesting that this capacity is to be called strength if we think of it not as simply given but as acquired by us (MM 6: 397). As I will elaborate, he deems this capacity necessary for acquiring inner freedom, which is the constitutive basis of virtue (MM 6: 408). In the above described case, however, self-control as a psychological condition can only be relevant as a tool for performing morally correct actions.

Attention is usually paid to the context in which Kant mentions self-control (Selbstbeherrschung) at the beginning of the Groundwork of the Metaphysics of Morals (4: 394), where he argues that qualities such as self-control and the moderation of affects and passions can have only instrumental value. At first glance, Kant’s example of the cruel Roman dictator Sulla perfectly illustrates this view: Sulla can be said to have self-control because he steadfastly follows his bad maxims (Anth 7: 293). Although his ends and maxims are not morally acceptable, he is good at compelling himself to perform actions by which he realizes his evil goals and follows his evil maxims. When interpreted merely along these lines, the Kantian capacity for self-control is a tool without which we would be unable to compel ourselves to act as we think we should, whether for moral or other reasons. To the extent that we merely acknowledge self-control’s role in following already adopted maxims, we fail to capture the distinctive aspect of moral self-control identified by Kant – its role in the adoption of virtuous maxims and ethical end-setting fails to receive the attention it deserves. As a consequence, virtue as moral strength (as a proper exercise of self-control) and moral weakness (as a lack of such self-control) are also likely to be understood as expressed merely at the level of observing already adopted maxims.Footnote ⁵

Self-control is a psychological condition (L-Eth 27: 360) that plays a crucial role in Kant’s ethics, but this is not to say that self-control is the only subjective condition of Kantian virtue. Kant does not list it together with the four subjective conditions of our moral receptivity – probably because self-control is meant to have a more ‘active’ status than moral feeling, conscience, love of human beings and self-respect (MM 6: 399–402). It is due to these four psychological conditions that we ‘can be put under obligation’ (MM 6: 399), and it is through self-control that we put ourselves under obligation.Footnote ⁶

Despite the abundant textual evidence for and important clarificatory advantages of appealing to the notion of abstraction to explain self-control, Kantian self-control has rarely been viewed through this lens. A thorough analysis of the terms Kant uses to explain the phenomenon of moral self-control, such as ‘control (Beherrschung)’, ‘rule (Herrschaft)’, ‘autocracy’, ‘government (Regierung)’, ‘self-composure’, ‘inner freedom’, ‘apathy’ and ‘free self-constraint’, along with his notions of abstraction, cultivation, affect and passion, shows why we should not commit ourselves to the reductive model of self-control suggested by the Groundwork passage. This analysis also shows that Kant provides a novel account of moral self-control by offering an alternative not only to the Aristotelian views of continence and the moderation of affects and passions, but also to the Humean model of self-control, conceived as a battle between different types of passions. Even more, if Kantian self-control is more than a mere instrument in the performance of morally good actions, and if we, in this light, rethink its relation to some other important psychological conditions, then we have good reason to buttress the relevance of Kant’s moral psychology.

1.2 Further Details

The prevalence of the above picture, to which a purely instrumental reading of self-control is tailored, is not without reason. Kant gives us grounds for subscribing to this reconstruction. I first explain why this picture is justified by sketching the contrast between some basic ideas in Kant’s moral theory, which hinge on the necessary purity of reason and empirical psychology. I then situate Kant’s doctrine of virtue within this pure–impure divide and outline the relevance of certain psychological conditions in this doctrine.

2.1 Different Interpretations of Autocracy

Some scholars claim that autocracy is not a capacity. They argue that it is a matter of being good at exercising different capacities and techniques. Others disagree. On their view, autocracy is the capacity for controlling inclinations. A more substantial disagreement arises when it comes to explaining autocracy’s role in acting virtuously in relation to the autonomy–autocracy distinction. Roughly speaking, there are two possibilities. One is to claim that autocracy is only needed to remove obstacles once we are tempted to fail to act in accordance with our already adopted morally correct maxims. The other is to claim that autocracy must also reach deeper than our ability to adhere to our established maxims.

According to Henry Allison’s influential reading, autocracy and autonomy are two different aspects of the same capacity. Autocracy is ‘actual strength of character or self-control’, and autonomy is ‘the mere capacity (Vermögen) for it’ (Reference AllisonAllison, 1990: 164). The autonomy–autocracy distinction is cashed out in terms of a capacity and its actualization. Self-control ‘is merely a necessary and not also a sufficient condition of virtue’, and virtue is a form of self-control that is based on a principle of inner freedom (Reference AllisonAllison, 1990: 164). Viewed from this perspective, the vicious Roman dictator Sulla, who is very good at following his morally impermissible maxims, can be said to satisfy a necessary but not also a sufficient condition for virtue. Or, he can be said to lack the form of self-control that characterizes virtue; his self-control is not based on inner freedom and this explains his bad maxims. However, from this we do not yet know whether Sulla should use his capacity for self-control to compel himself to adopt morally permissible maxims and how such self-constraining activity relates to the principle of inner freedom.

A very different explanation of the autonomy–autocracy distinction can be found in Baxley’s extensive work on this topic. Reference BaxleyBaxley (2003: 17–18) finds Allison’s reading unsatisfying because it incorrectly ‘equates autonomy with a capacity for self-control’ and  ‘fails to capture the legislative-executive thrust of the distinction’. On her view, Allison’s reading blurs the distinction between two separate capacities: autonomy as a legislative power and autocracy as an executive power. Autocracy is required for compliance with the norms prescribed by the legislative power of the will, which, as Reference BaxleyBaxley (2003: 18, Reference Baxley, Denis and Sensen2015: 229) adds, involves the notion of self-determination. Accordingly, weakness of the will is explained as a case in which the will is autonomous but has failed to achieve autocracy (Reference BaxleyBaxley, 2010: 60, 81). Reference BaxleyBaxley (2003: 15, 16) relatedly claims that autonomy concerns ‘motivational independence’, whereas autocracy concerns ‘temptation independence’.

If viewed in this light, Sulla appears to have a problem with achieving ‘motivational independence’. He follows his maxims, and this makes him different from Baxley’s weak-willed person who fails to attain ‘temptation independence’. Baxley might also argue that Sulla lacks autocracy, understood as a specific kind of self-control that involves following morally correct maxims, that is, a kind of self-control that presupposes autonomy. In any case, according to her ‘two separate powers’ account, Sulla seems to fail to act virtuously for a completely different reason than a lack of self-control – he somehow, independently of his executive power of self-control, seems to misuse his legislative power.

Reference GuyerPaul Guyer (2005) ascribes a similar role to autocracy, but autocracy is not a specific capacity in his view. He argues that the process of achieving autocracy is the empirical realization of autonomy. For Guyer, autonomy is an ideal, and autocracy cannot be the capacity for autonomy when it is realized – it is a state we achieve via the cultivation of different capacities and practices. This cultivation includes strengthening our moral feelings and developing certain techniques, such as controlling our imagination and becoming better at postponing our judgements. Reference GuyerGuyer (2005: 143–44) suggests that these different ways of cultivating are ‘simply the means’ by which we ‘implement’ maxims or realize moral ends.

Reference WilsonEric Entrican Wilson (2015) finds Guyer’s explanation of the autonomy–autocracy distinction unsatisfying because autonomy is a property of the will that we all have. As a result, autocracy cannot be a matter of implementing the ideal of autonomy. Autocracy or ‘self-command is a condition or state achieved by those agents who become proficient at solving problems presented by the passions’ (Reference WilsonWilson, 2015: 256). Self-command is a matter of being good at exercising different capacities and skills. Such proficiency makes one able to ‘stick to the results of self-legislation over time’ and to ‘extend the results of self-legislation over time’ (Reference WilsonWilson, 2015: 256, 271). As a specific kind of moral self-control, self-command presupposes the results of self-legislation.Footnote ²⁴ The person with self-command is capable of staying committed to ‘the results of his own activity of moral deliberation’ (Reference WilsonWilson, 2015: 260).

2.2 Autocracy Reconsidered

Indeed, Kant does not make it easy to settle the question of whether autocracy is a capacity. In the Metaphysics of Morals (6: 383), he argues that autocracy involves consciousness of the capacity to master our inclinations. Elsewhere, he refers to autocracy by using the term ‘capacity’ (PM 20: 295), and the lecture notes support both the notion that autocracy is to be explained in relation to the exercise of a number of capacities and the notion that autocracy itself is the executive power of self-control. The fact that Kant has been reported as presenting autocracy in relation to other powers, such as imagination and judgement, does not necessarily support the conclusion that autocracy cannot be more closely tied to a particular capacity. The lecture notes suggest that autocracy is not simply about ensuring that we control our inclinations, but also about ensuring that we control all our faculties (L-Eth 27: 364–8; 24: 1496–98). Accordingly, self-control has been described as ‘the faculty for freely disposing over the free use of all one’s powers’ (L-Met 28: 589–90).

What is important for our present purposes, however, and consistent throughout the texts, is that autocracy is inseparable from the capacity for self-control and essential to Kant’s conception of virtue. Kant speaks of virtue as an actualized capacity for self-control, but he also holds that we must assume that we always have a mere capacity for self-control (MM 6: 397).Footnote ²⁵ Autocracy can then involve consciousness of either our realized or our yet-to-be-realized capacity for self-control. In the lecture notes, autocracy is sometimes used to refer to states in which human beings actually have their inclinations and capacities under control (e.g. L-Eth 29: 626). But it is very unlikely that what Kant had in mind in the passage on autocracy from the Metaphysics of Morals (6: 383) was virtue as acquired moral strength. Although there is no doubt that fulfilling duties of virtue requires exercising the capacity for self-control, the doctrine of virtue need not presuppose that we really have our inclinations under control in a given moment, or that we have autocracy as an actual strength. Rather, Kant’s suggestion seems to be that autocracy, as a kind of consciousness of the mere capacity for self-control, must be taken into account whenever we want to explain how we can fulfil the duties of virtue.

At first glance, this claim about autocracy as consciousness of a yet-to-be-actualized capacity for self-control might seem unimportant, or even pointless. But it is not. Kant thought that it is through this consciousness that we become aware of our freedom. Near the end of the second Critique (5: 159), he describes how we become aware of our freedom to put the influence of all sensible impressions aside, that is, how we become aware that we can do what the moral law demands of us. This is, as Kant illustrates, ‘as it were, to raise oneself altogether above the sensible world, and this consciousness of the law also as an incentive is inseparably combined with consciousness of a power of ruling over sensibility [die Sinnlichkeit beherrschenden Vermögens], even if not always with effect’ (CPrR 5: 159) [italic in the original].Footnote ²⁶

An apt illustration of how we become aware of our power to master inclinations is Kant’s well-known example in which a prince forces a man to choose between giving a false testimony or being executed. This man ‘would perhaps not venture to assert whether he would’ really choose to be executed or not, but he ‘must admit without hesitation that it is possible for him’ to overcome his love of life or to control one of his strongest inclinations (CPrR 5: 30). As soon as he starts forming his maxims, this man becomes conscious of the moral law, and then he also becomes aware that it is possible for him to choose to do what the moral law demands. Through his consciousness of his capacity for self-control, he becomes aware that his choice is capable of producing certain objects (MM 6: 213; Anth 7: 251). This seems to be why Kant argues that the capacity for self-control ‘can and must be simply presupposed in man on account of his freedom’ (MM 6: 397).

Even more, the claim that autocracy can be attributed to our will at all times, not only in those moments in which we actually have our inclinations under control, opens up the possibility of self-control’s coming into the picture even during the very process of maxim adoption. Although the man in the above example might be said to have ‘a merely theoretical cognition of a possible determination of choice’ (MM 6: 218), this knowledge has not yet fed back into his decision regarding how to act in that situation because he has not yet exercised his capacity for self-control by actually refusing to give priority to love of life in his maxim. Drawing on Kant’s idea that actual self-determination requires the inclusion of an incentive in lawgiving (MM 6: 218), we may conclude that this man has not yet actually determined his choice. Rather, he has merely realized that it is possible for him to put aside his love of life – he has only become aware of the freedom of his choice.

In line with this, Kant does not seem to draw a clear-cut distinction between autonomy and autocracy by claiming that the latter is only needed for sticking to the results of self-legislation, or for sticking to our already made moral decisions. He claims that the autonomy of practical reason is simultaneously its autocracy (MM 6: 383; PM 20: 295). There are further reasons why I am more willing to accept Allison’s explanatory framework than those that present autocracy and autonomy as two completely separate powers. First, it does not seem to rest on a one-sided understanding of Kant’s notion of temptation. In the passage from the Metaphysics of Morals (6: 383), temptation is mentioned as the reason why autocracy is needed in the case of human beings. Autocracy can be said to be needed because we are tempted not to follow our already adopted, morally good maxims; unlike holy beings who gladly do everything that is in accordance with the moral law, we must compel ourselves to perform morally correct actions. This is acceptable as long as we remember that Kant also holds that purely rational beings are incapable of morally unacceptable maxims (CPrR 5: 32; 5: 79 and G 4: 439) and that we, by contrast, are tempted to adopt such maxims. Acting virtuously involves compelling ourselves to adopt maxims that guide actions that are both morally correct and morally worthy. Given Kant’s overall emphasis on the maxims of actions and his insistence that virtue necessarily involves maxims of ends, the main reason why autocracy is needed is rather our temptation to adopt morally impermissible maxims.

Second, the claim that autocracy and autonomy are two separate powers that perform two completely different functions, may find support in the Collins notes (1784–1785) but not in the Mrongovius notes (1785). In both cases, autocracy is presented as the executive power and equated with moral feeling (L-Eth 27: 361–2, 29: 626), but this is glossed in different ways. According to the Collins lecture notes, autocracy is ‘the authority to compel the mind’, which ‘involves mastery over oneself, and not merely the power to direct’ (L-Eth 27: 362). This power to direct is a forerunner of autonomy: It corresponds to the principle of appraisal of obligation, which should not be confused with the principle of performance or execution (L-Eth 27: 274–5). The executive power ‘can compel us, in spite of all impediments, to produce certain effects’ (L-Eth 27: 362). This claim can be taken to suggest that autocracy, as the executive power, is only needed to remove obstacles once we are tempted to fail to act in accordance with our previously made decisions. Compelling ourselves ‘to produce certain effects’ in spite of all sensible obstacles would then have to be understood as bare compelling ourselves to perform certain physical actions, or as mere disciplining ourselves to obey given rules. If conceived in this way, the capacity for self-control may seem completely different from the power that provides us with norms. However, this conclusion can be challenged by pointing to a passage from the Mrongovius lecture notes that suggests that autocracy involves the self-determination of our reason, rather than simply presupposing it (L-Eth 29: 626). Since the legislative power is meant to involve the notion of self-determination, this causes problems for the reading according to which we should draw a sharp distinction between autocracy and autonomy. Compelling ourselves ‘to produce certain effects’, then, cannot be reduced to merely disciplining ourselves to follow certain rules.

Furthermore, in the Metaphysics of Morals, Kant confirms the intimate relationship between self-determination and moral feeling. Moral feeling is a distinctive kind of pleasure or displeasure that we feel whenever we become aware of our possible morally good or bad actions (MM 6: 399). It is ‘respect [Achtung]’ for the moral law ‘in its subjective aspect’ (MM 6: 464).Footnote ²⁷ Now, Kant’s point is that moral feeling is required for self-determination, which is widely held to belong to the legislative power. According to the lecture notes, however, moral feeling is equivalent to autocracy, or the executive power (L-Eth 27: 361–2, 29: 626), and the point that moral feeling is required for self-determination may then be taken to suggest that the executive power’s role is not limited to performing physical acts in accordance with the norms issued by the legislative power. In other words, it becomes difficult to defend the claim that autocracy and autonomy are two completely separate powers, and that autocracy only comes into play once the task of self-legislation has been properly fulfilled.

Kant argues that every determination of choice proceeds from the representation of a possible action through moral feeling (MM 6: 399). This distinctive feeling enables an actual determination of our choice by the moral law, conceived as the state in which we ‘take an interest in the action’ (MM 6: 399). In line with this, Kant explains that moral laws command morally necessary actions for which

This quotation confirms the thesis that moral feeling is necessary for the determination of our choice by practical laws. It also tells us that moral feeling is not the objective condition of morality that could be the cognitive basis of practical laws. Kant holds that the objective principles cannot be based on any kind of feeling.

However, this does not exclude the possibility that moral feeling, as one of the ‘subjective conditions for receptiveness to the concept of duty’ (MM 6: 399), is a necessary element in the process of adopting the particular moral maxims on which we actually act. It may still be needed when it comes to the adoption of subjective principles of volition through which we can determine our choice independently of external influences. Together with self-control, which is needed if we are to free ourselves from the influence of sensible impressions on our minds, moral feeling can still be required for the adoption of the rules that we make for ourselves on subjective grounds.

In fact, Kant argues that an incentive is necessarily involved in the maxim adoption characteristic of virtue. In the Metaphysics of Morals, he seems to retain the old distinction between the principle of appraisal of obligation and the principle of its execution: now as the difference between a law and an incentive. Whereas the former ‘makes an action a duty’ by representing it as ‘objectively necessary’, the latter ‘connects a ground for determining choice to this action subjectively with the representation of the law’ (MM 6: 218). But Kant now also suggests that every lawgiving must have these two elements and that the second element must be present if we really are to determine our choice (MM 6: 218). An incentive must be included in the process of self-legislation if we are to become morally motivated to perform an action; without a subjective ground for determining our choice, moral laws would never become subjectively practical for us. We would not strongly will what we ought to do, and this is precisely what virtue as moral strength involves.

This analysis is of particular importance in explaining the internal lawgiving that is constitutive of virtue. Unlike juridical or external lawgiving, internal lawgiving does not allow for inclinations or aversions as determining grounds of our choice. Virtuous lawgiving involves the proper incorporation of the incentive of the duty into the rules we prescribe to ourselves. Neither our own inclinations nor someone else can properly motivate us to act morally, or constrain us to make a free choice in the Kantian sense of determining our choice by the pure incentive. For these purposes, we need to acquire inner freedom. This is why Kant treats inner freedom as the condition of becoming virtuous. In line with this, Kant insists that the way of thinking characteristic of virtue can never become habitual, for virtue would then result from natural necessity. It would be a kind of unfree mechanism. He argues that virtue must always emerge entirely new and original from one’s way of thinking (Anth 7: 147) and that the maxims of virtue, which must be freely adopted, are in an unending progression (Anth 7: 147; MM 6: 409; CPrR 5: 32–33).

In order to properly accommodate Kant’s claim that virtue must be based on inner freedom and always proceed from freedom, we must also account for self-control in the process of thinking and end-setting. This is how we can explain the specific character of the duties of virtue, or spell out how we, in ever-new situations, determine our choice through the adoption of maxims of virtue. From my point of view, it is very unlikely that Kant’s claim is that autocracy, as consciousness of our purely instrumental capacity for self-control, is essential to his doctrine of virtue. Our analysis of self-control through the prism of ‘autocracy’ shows that it is better not to downgrade Kant’s conception of moral self-control to a mere tool for following established rules and sticking to our already made moral choices.

If we read Kantian self-control as abstraction, then we can easily accommodate Kant’s essential claims about virtue: because of our temptation to base our maxims on our inclinations and our tendency to self-deception, we must continuously exercise our capacity for abstraction in order to acquire inner freedom, that is, to put aside the influence of all sensible impressions on our minds. In what follows, I will therefore approach self-control as abstraction and explain how this ability can be used to acquire inner freedom.

3.1 Abstraction in Prudential and Moral Self-Control

For Kant, abstraction is essential to prudential self-control. Many people ‘are unhappy because they cannot abstract’; the one who plans to get married would have a good marriage if he could ‘overlook a wart on his beloved’s face’ (Anth 7: 131–2). We naturally pay attention to deficiencies, and it takes effort to look away from them or to abstract from that sensible representation of a wart. When properly developed, the faculty of abstraction becomes the strength of mind that enables us to set aside the obstacles that stand in the way of our happiness (Anth 7: 132).

The Mrongovius lecture notes (1784–1785) suggest a more nuanced version of this view: not all abstracting is proper (L-Anth 25: 1240). One abstracts too little if one decides against marrying an otherwise perfect woman simply because she has pockmarks, whereas one abstracts too much if one ignores a candidate spouse’s beauty and allows fear of infidelity to determine one’s choice; both end up unhappy, because they do not abstract at will (L-Anth 25: 1240). They involuntarily follow the natural flow of their attention. If they were to abstract voluntarily, they would exhibit self-control and manage to make themselves happy (L-Anth 25: 1239).

This kind of self-control is in accordance with the rules of prudence (L-Eth 27: 362). It is in the service of the human general desire for happiness. Of course, claiming that prudential self-control involves abstracting from our general inclination toward happiness or from all particular inclinations the satisfaction of which promotes our happiness, would make prudential self-control pointless. Kant’s claim cannot be that one acquires prudential self-control by abstracting from one’s desire to enjoy deserved leisure or a good meal and companionship; such desires are even in line with ‘the law of virtue’ and make virtue appealing.Footnote ²⁹

By Kant’s lights, only feelings and desires that make us lose self-control are obstacles to promoting our happiness, and we can use our ability to abstract from different sensible impressions to avoid such feelings and desires. Kant thought that having affects and passions is imprudent (Anth 7: 253, 267, 273). His example of a rich person ‘whose servant clumsily breaks a beautiful and rare crystal goblet’ (Anth 7: 254) illustrates the use of abstraction in preventing affects. The underlying thought is that it is possible for the rich man to avoid becoming overpowered by his anger. If he were to make a quick ‘calculation in thought’ and compare the pain he feels as a reaction to the accident to all the pleasures he enjoys as a rich man, he would not fear the loss of his entire happiness and would not descend into the affective state of anger. Gaining control over the condition of the representations in his mind by abstracting from the representation of his broken crystal goblet might prevent him from entering into an affective state in which he is no longer capable of comparing one feeling against the sum of other feelings. Kant analogously presents the imprudence of passions: When in a passionate state, we are no longer capable of comparing one desire with the sum of all other desires (Anth 7: 265). When making a choice, a person with a passion for avarice becomes blind to all other desires, such as her desire to be loved by others. Together with the insatiability of the passions (Anth 7: 266), this makes this person unhappy.Footnote ³⁰

And yet, Kant’s view is not that we merely use our capacity for abstraction in order to realize the ends we necessarily have, such as our happiness. This capacity can also be used in more complex cases where we ought to set ourselves certain ends and realize them. Many passages are suggestive of this, but the following seems most telling:

Since none of our inclinations should determine our maxim adoption, we must abstract from all inclinations when adopting virtuous maxims. This does not involve banishing inclinations from our minds. We cannot and should not get rid of our inclinations.Footnote ³² Abstracting from all inclinations, or redirecting our attention from them, is required if we are to take an interest in compliance with the moral law and set ourselves moral ends. When we exercise our capacity for abstraction in accordance with the categorical imperative – with the aim of properly incorporating the incentive of the moral law into our maxims – our use of this capacity is called moral.

Thus far, we know that abstraction can be voluntary and involuntary, and that voluntary abstraction, which is crucial for self-control, is opposed to our natural way of paying attention. However, Kant does not deny the intimate relationship between abstraction and attention. As his students’ notes suggest: ‘the same attention is present in abstraction, only the objects are different’ (L-Anth 25: 1239); attention does ‘not stop with abstraction’, and abstraction ‘is the actualization of attention’ (L-Met 29: 878). Abstraction is a more complex attentive activity. In the Attempt to Introduce the Concept of Negative Magnitudes into Philosophy (2: 190), Kant claims that abstraction can be called ‘negative attention’ because it is an effort to cancel ‘certain clear representations’ for the purposes of ensuring that what remains becomes ‘much more clearly represented’. Abstraction involves an effort to disregard certain representations and redirect our attention to others so that they become more clear. It can be said to involve attention, but not the attention that we naturally pay to sensible objects.

In his Anthropology (7: 131), Kant argues that abstraction consists neither in merely paying attention nor in a lack of attentiveness. Attention and abstraction are different ways of becoming conscious of certain representations. When abstracting, we are not simply distracted by something – we voluntarily pay attention to some of our representations by turning attention away from others. For Kant, abstraction is ‘a real act of the cognitive faculty of stopping a representation of which I am conscious from being in connection with other representations’ (Anth 7: 131). It is a cognitive act through which we intentionally sever the relation between certain representations in our minds with the aim of focusing our attention on something else. Kant further reveals what he means by abstraction:

The faculty of abstraction enables us to oppose natural necessity by ignoring representations of the sensible objects we encounter. We actually abstract from ‘a determination of the object’, as it were, incorporated in our representation (Anth 7: 131). When abstracting, we disregard the determinations that the sensible objects impose on us, and we do so by modifying the status of the representations of these objects in our consciousness. We cannot really banish such representations from our minds, but what we can do is gain control over their state in our minds by disregarding the influence of various sensible impressions. Our capacity for abstraction makes us capable of reasoning as if the sensible impressions that we receive from objects did not exist in our minds.Footnote ³³

In sum, abstraction is the cognitive activity of preventing the influence of various sensible objects on our consciousness by redirecting our attention from them in order to pay better attention to the representations that remain. As such, it need not only have the negative aim of disengaging from certain representations. In what follows, I will explain this by showing how Kant’s notion of abstraction helps us to get a better grip on his view of self-control, inner freedom and self-determination. Before doing so, however, I will sketch the rudimentary level of self-control needed for mental health.

3.2 The Elementary Level of Self-Control

Kant argues that abstraction should not be reduced to distraction (Anth 7: 131), but he sometimes also treats distraction as a kind of abstraction. He claims that distraction (Zerstreuung) is ‘the state of diverting attention (abstractio) away from certain ruling representations [Vorstellungen] by dispersing it among other, dissimilar ones’ (Anth 7: 206; translation modified). Distraction can be either involuntary or voluntary (Anth 7: 206). Involuntary distraction is absent-mindedness. Voluntary distraction or dissipation involves intentionally taking one’s mind off things, thereby creating a diversion from one’s ‘involuntary reproductive power of imagination’ (Anth 7: 206–7). With voluntary distraction we intentionally redirect our attention, but we do so without having the aim of paying better attention to the representations that remain. For example, when trying to get ‘rid of the object’ that makes us feel sad (L-Anth 25: 1240) we divert our attention from the representations that our recalcitrant power of imagination continuously reproduces, and we do so by dispersing attention to some other objects – for instance, by occupying ourselves ‘fleetingly with diverse objects in society’ (L-Anth 25: 1240).

Voluntary distraction is a precondition of mental health (Anth 7: 207). Kant’s discussion of hypochondria explains why voluntary distraction is an elementary level of self-control and how self-control is related to some other capacities. Picture someone who interprets every little sniff and cough as a sign of a serious disease and who suffers from obsessive fear and anxiety as a result. We can apply Kant’s insights to this case. This person is involuntarily distracted – he clings to certain representations so strongly that he cannot let them go, and he is in an unhealthy state in which he lacks self-control (L-Anth 25: 1240). True, his condition is not as severe as madness, where ‘fantasy plays completely with the human being and the unfortunate victim has no control at all over the course of his representations’ (Anth 7: 181). Still, he cannot freely use his capacities for imagination, reason and feeling. Hypochondriacs have a diseased imagination (Einbildungskrankheit) (Anth 7: 213). The imagination of the person in question turns into mere fantasy because he cannot restrain its play at will. As a mere quasi-mechanical activity, his imagination regularly misinterprets certain physical sensations as the symptoms of disease. He cannot intentionally turn his attention away from every little sniff and cough and refrain from seeing them as harbingers of doom. Moreover, he cannot help thinking that he has a serious disease, although he might at times realize that this is an irrational belief to hold. In the case of hypochondria, ‘the patient is aware that something is not going right with the course of his thoughts, insofar as his reason has insufficient control over itself, to direct, stop or impel the course of his thoughts’ (Anth 7: 202). Furthermore, his obsessive fear and anxiety are the signs of his hypochondria, which is the opposite of the mind’s power to master ill feelings (CF 7: 103). In sum, he cannot voluntarily distract himself from the chimerical representations that the reproductive power of the imagination unrestrainedly produces; his reason cannot control itself by disregarding the influence of certain sensible impressions, and he cannot free himself from the ill feelings elicited by the figments of his imagination. This lack of control over his own capacities destroys the balance of the soul necessary for his mental health.

The capacity for abstraction in this elementary form, conceived as a certain degree of control over the course of one’s sensible representations, is not only a prerequisite for mental health. It is also a precondition for exercising self-control on a higher level that is more directly required for the fulfilment of one’s moral obligations. Kant states that being subjected to affects and passions is ‘probably always an illness of the mind’ (Anth 7: 251), and he describes moral strength of soul as a state of health in moral life (MM 6: 409; 6: 384). As in the case of mental health, when Kant speaks of virtue as moral strength he seems to have in mind a balance of the soul that involves having control over all its powers. I will examine how we acquire the inner freedom of virtue by exercising our capacity for self-control: first as preventing affects and acquired passions, then as abstracting from other inclinations and feelings on which they are based.

3.3 Acquiring Inner Freedom

Defining virtue merely as self-constraint does not capture its essence: Virtue would then be a battle of our inclinations in which the stronger inclination wins. It must be free self-constraint (MM 6: 383). Kant therefore discusses virtue as ‘self-constraint in accordance with a principle of inner freedom’ and as ‘a moral constraint’, which is possible ‘in accordance with the laws of inner freedom’ (MM 6: 394; 6: 405). Inner freedom is the constitutive basis of virtue, and it is inseparable from self-control:

Inner freedom is ‘the condition of all duties of virtue’ (MM 6: 406). In the first instance, it is this condition merely in a negative sense – as a way of dealing with our inner obstacles to morality.Footnote ³⁴ Acquiring inner freedom is necessary for becoming virtuous, because removing outer obstacles would only suffice for acting in accordance with duty and the fulfilment of duties of right. With regard to duties of right, it does not matter whether the end that one intends is moral, or whether one’s maxim is genuinely moral. By contrast, if the moral agent is to act for the sake of duty, she must set herself moral ends, which requires that she fulfils the duty of apathy.

The above quotation also suggests that inner freedom surpasses the fulfilment of the duty of apathy. Inner freedom is also the condition of virtue in a positive sense. It entails a positive command that human beings rule or control themselves. By holding ‘the reins of government in its own hands’, or by governing our feelings and desires, our reason prevents them from mastering us. As I will clarify, this self-government or self-rule is self-determination. The virtuous way of thinking is one by which we freely determine ourselves ‘to act through the thought of the law’ (MM 6: 407). What matters here is the way in which agents determine their choices: Kantian virtuous agents must freely adopt their maxims, and not because of the feelings and desires they happen to have. Such agents must set themselves moral ends. They must redirect their attention to these ‘pure’ ends.

In the Metaphysics of Morals, Kant discusses two requirements of inner freedom (MM 6: 407): freedom with regard to affects and freedom with regard to passions. These requirements represent different aspects of the capacity for self-control.Footnote ³⁵ Even more, I take Kant to be suggesting that there is also a third requirement of inner freedom, which involves the crucial, but neglected aspect of moral self-control. Here, I have in mind Kant’s claim that inner freedom is the capacity to release ourselves from all inclinations and corresponding feelings (CPrR 5: 161).

In what follows, I analyse these three requirements of inner freedom as aspects of self-control, conceived as abstraction.Footnote ³⁶ My analysis shows that the use of the capacity for self-control to prevent affects and passions, which would suffice for following maxims of virtue, serves as a kind of preparatory ground for their adoption, whereas the actual adoption of these maxims also requires that we set aside all inclinations and the feelings on which they are based. On this basis, I conclude that Kant’s conception of moral self-control necessarily involves two intimately related levels, and that it can be central to virtue only if understood in this way.

4.1 Different Interpretations of Moral Strength

Virtue signifies ‘a moral strength of the human will [eine moralische Stärke des Willens]’ (MM 6: 405).Footnote ⁴⁶ It is ‘the moral strength of a human being’s will in fulfilling his duty’ (MM 6: 405), ‘the strength of a human being’s maxims in fulfilling his duty’ (MM 6: 394), and ‘the strength of intention [die Stärke des Vorsatzes]’ (MM 6: 390).Footnote ⁴⁷ Other definitions of virtue, such as ‘moral disposition in struggle’’ (CPrR 5: 84) and moral self-constraint, presuppose the idea that we ought to acquire moral strength.

Some Kant scholars seem to hold that moral strength is not essential to virtue. Reference Johnson and CuretonRobert Johnson and Adam Cureton (2017) argue that one can act from duty while lacking moral strength when one is not tempted to act otherwise. Along these lines, Reference PapishLaura Papish (2007: 141–42) suggests that the moral worth of an action does not depend on moral strength. Although it sounds plausible to say that moral strength is not needed if an agent is already inclined to help, Kant’s view seems to be that even this agent must control himself not to perform a morally good action simply from his natural inclination to help. Fulfilling the duty of beneficence requires a proper maxim, and the inclination to help cannot produce such a maxim (CPrR 5: 118). If the agent is to adopt a maxim of virtue, he must acquire inner freedom by setting aside all inclinations and their ends. Otherwise, he will not be in a position to set himself the moral end of increasing others’ happiness by adopting a maxim of helping them (MM 6: 452). Hence, I believe that moral strength is needed whenever we are to perform a morally worthy action.Footnote ⁴⁸

In line with this, Kant points out that moral strength is one aspect of our capacity for self-control: ‘For while the capacity (facultas) to overcome all sensible impulses can and must be simply presupposed in man on account of his freedom, yet this capacity as strength (robur) is something he must acquire …’ (MM 6: 397) [italic in the original]. The assumption here is that all of us have an innate capacity for self-control, because ‘ought’ implies ‘can’. We have a duty to acquire moral strength by constantly developing this capacity through its free exercise in accordance with the moral law. Furthermore, ‘true strength’ of virtue is not only ‘a tranquil mind’ but also ‘a considered and firm resolution to put the law of virtue into practice’ (MM 6: 409). Virtue as moral strength is also a kind of self-rule that surpasses the fulfilment of the duty of apathy. This specific aspect of self-control has been read as self-determination. Accordingly, Kant speaks of the moral strength of the human will and maxims.

Kant’s conception of moral strength is sometimes not explained in terms of self-control. Reference McCartyRichard McCarty (2009: 196, 230) seems to interpret moral strength as the psychological force of the moral incentive that all of us happen to have. Reference GuyerPaul Guyer (2000: 307) argues that virtue in the sense of moral strength is ‘caused’ by virtue in the sense of a virtuous disposition: moral strength results from an act of inner freedom, which is to be understood as ‘an agent’s adoption of respect for the moral law as his fundamental maxim’. If I understand it correctly, moral strength is here presented as a causally produced mental state that is not required for acquiring inner freedom.Footnote ⁴⁹

Anne Margaret Baxley does explain moral strength as the strength of the power of self-control. She states that moral strength is required ‘to enforce the morally good choices we legislate to ourselves as autonomous rational agents’ (Reference BaxleyBaxley, 2010: 57). Autocracy is a form of self-control that ‘arms us with moral strength to execute self-legislated principles’ (Reference BaxleyBaxley, 2010: 83). This executive power enables compliance with maxims, whereas autonomy is tasked with issuing them.

Reference Grenberg and DenisJeanine Grenberg (2010) interestingly interprets moral strength as the realization of inner freedom. Vice amounts to ‘weak’ realizations of inner freedom, which is opposed to ‘truer, more complete realizations of inner freedom’ (Reference Grenberg and DenisGrenberg, 2010: 163). Whereas the former involves a moment of rationalization, the latter involves keen ‘attentiveness’ to the moral law by which we fully accept our moral obligations (Reference Grenberg and DenisGrenberg, 2010: 165). Such ‘attentiveness’ occurs via our moral feeling, and it is by engaging in contemplation of the moral law that we strengthen this feeling. Grenberg seems to hold that this happens before we adopt maxims.

I believe that we need to extend the scope of our reading of Kantian moral strength by further explaining the close tie between moral strength and maxim adoption – in conjunction with its connection to Kant’s rejection of the model of virtue as a mere habit. In the Metaphysics of Morals, Kant objects to defining virtue as ‘a long-standing habit [Gewohnheit] of morally good actions acquired through practice’ (6: 383). In the Anthropology, he argues against reducing virtue to a ‘skill in free lawful actions’ (7: 147). But the core of his argument is the same: Were virtue a mere habit, we would have to embrace the unacceptable claim that virtue is a kind of natural mechanism.

In the Metaphysics of Morals, Kant therefore emphasizes that virtue can only be a kind of ‘free skill (habitus libertatis)’ (6: 407) [italic in the original]. It cannot be reduced to consistently acting in accordance with previously established rules but must also involve the free determination of choice in ever-new situations. As Kant explains:

This freedom in adopting maxims is the inner freedom of virtue, and it is why practicing virtue can never become a mere habit.

In line with this, Kant insists that virtue is ‘always in progress’ (MM 6: 409) and ‘can never be completed’ (CPrR 5: 33). He suggests that the best we can do is to ensure the ‘unending progress’ of our maxims and ‘their constancy in continual progress’ (CPrR 5: 32–33). Kant’s emphasis on this unending progress may concern the motivating ground of our maxims – which should be constantly renewed – whereas the constancy in this continual progress may concern their purely cognitive, theoretical basis.

In the Anthropology, Kant explicitly contrasts virtue as moral strength with skill in performing free lawful actions. He does so by arguing that ‘virtue is moral strength in fulfilling one’s duty, which never should become habit [Gewohnheit] but should always emerge entirely new and original from one’s way of thinking [immer ganz neu und ursprünglich aus der Denkungsart hervorgehen soll]’ (Anth 7: 147) [italic in the original]. His point is that virtue as moral strength must always go beyond a mere mechanism of applying certain rules, because the way of thinking characteristic of virtue, or the way of thinking according to moral laws (CPrR 5: 160), can never become habitual or unfree. Moral strength is needed, in ever-new situations, for the process of becoming morally motivated by a proper way of thinking.

Without any further explanation of the close link between moral strength and maxim adoption, we risk presupposing a static account of maxims and a mechanistic account of rule application, both of which are ruled out by the above-discussed passages.Footnote ⁵⁰ By making it seem as though moral strength is needed only for following maxims once the activity of adopting them has been completed, we lose a useful tool for accommodating Kant’s claim that virtue, as moral strength of the human will, can never become a mere habit. In my view, we can properly capture the freedom in adopting maxims that characterizes maxims of virtue only through providing a proper account of self-control’s role in ethical end-setting. Without such an account, we risk underestimating the active and dynamic aspects of Kantian moral strength of will.

What is more, if we were to hold that the strength of intention characteristic of virtue merely concerns an intention to follow established maxims, we would have to say that the fulfilment of duties of virtue merely involves compelling ourselves to perform certain actions. In this way, we would sidestep the very essence of Kantian virtue, which is free self-constraint in end-setting.

4.2 The Two Faces of Moral Strength

In his notes to the Doctrine of Virtue (23: 394), Kant writes that moral strength is strength of intention (Vorsatz) and strength in action (That). By building on this claim, I address two aspects of moral strength. I contend that strength of intention, which is essential to Kantian virtue, cannot be a mere intention to follow our maxims, for it must also be an intention by which we set ourselves moral ends in ever-new situations. This intention is constitutive of maxims of ends, and an end is not simply an action but an aim we intend to realize through the performance (or avoidance) of a particular action. By strength in acting, I mean mere consistency in performing (or avoiding) actions by which we realize moral ends or adhere to our moral maxims.

5.1 The Puzzle of Weakness of Will

At first glance, weakness of will might seem easy to explain. It is acting against our better judgement. But how weakness of will is possible is a matter of debate.Footnote ⁵⁹ Consider the one who judges that he should stop smoking for the sake of his health but still lights up another cigarette. There is a principle that he holds dear, and there is his failure to act accordingly. The problem arises when it comes to explaining this failure such that weakness of will involves freely acting against one’s better judgement. If the agent’s smoking another cigarette is explained by his being overpowered by an irresistible desire to smoke – such that he could not do otherwise – this would seem to be a case of compulsion. And if the agent decides to smoke the cigarette, or forms an intention to do so, we might think that he has changed his mind, but then the conflict that is essential to weakness of will disappears. The further question is how this decision, or intention, is related to his better judgement. Perhaps he has decided to smoke just this last cigarette because he is nervous and believes that smoking will calm him. Deep down, he may still hold that he should stop smoking for health reasons, but he now has one more, conflicting reason. One may be tempted to say that the solution is simply that his strongest reason or desire wins, but this simple answer pushes us back to the problem of how to explain freely acting against one’s better judgement.

Situating Kant’s brief treatment of moral weakness within his own ethical framework complicates the issue even further.Footnote ⁶⁰ On the one hand, Kant suggests that moral weakness is a mere failure to follow our otherwise morally good maxims. He mentions the ‘weakness of the human heart in complying with the adopted maxims’ (Rel 6: 29) and argues that the weak are not ‘strong enough to comply with’ their ‘adopted principles’ (Rel 6: 37).Footnote ⁶¹ This easily leads to the conclusion that moral weakness is merely expressed at the level of following maxims. On the other hand, Kant addresses moral weakness as the first grade of our propensity to evil, which implies that moral weakness must also be expressed at the level of maxim adoption. This propensity ‘must reside in the subjective ground of the possibility of the deviation of the maxims from the moral law’ (Rel 6: 29), and it is ‘the subjective universal ground of the adoption of a transgression into our maxim’ (Rel 6: 41). It is still ambiguous whether, and if so how, these two aspects of moral weakness can be combined.

Furthermore, it has been argued that what we usually call weakness of will is incompatible with Henry Allison’s incorporation thesis, which is his interpretation of Kant’s passage on what characterizes freedom of the power of choice (Rel 6: 24). Reference AllisonAllison (1990: 40) claims that ‘an inclination or desire does not of itself constitute a reason for acting’; it becomes such a reason only when we incorporate it into one of our maxims.Footnote ⁶² The general concern is that Allison’s incorporation thesis makes acting against one’s better judgement impossible: moral weakness, as a failure to act in accordance with an adopted maxim, would then be based on yet another maxim, which again presupposes the incorporation of incentives.Footnote ⁶³

On the other hand, the claim that Kant’s notion of weakness is not expressed at the level of the incorporation of incentives falls prey to the difficulty of how to account for moral weakness as freely acting against one’s better judgement and as a manifestation of our propensity to evil. The morally weak agent does incorporate the law into his maxim, although not in a fully satisfying way (Rel 6: 29). If so, then we need an account of what goes wrong when it comes to the weak-willed person’s incorporation of incentives into her maxims.

There is an additional problem of how to distinguish between the weak, the impure, the vicious and the virtuous, especially with regard to how they incorporate incentives into their maxims. Impurity (Unlauterkeit) is the second stage of our propensity to evil, which leads to actions that are not done ‘purely from duty’ (Rel 6: 30). Viciousness (Bösartigkeit) is the third and worst grade of evil, which involves the subordination of the incentive of the moral law to the incentives of inclinations (Rel 6: 30). Unlike the vicious agent, the weak agent does not incorporate deviation from the moral law into her maxims by allowing the incentives of inclination to determine her choice, but it is less clear how she differs from the impure agent and the virtuous agent.

Kant scholars have come up with a variety of creative solutions to the outlined problems. Some employ a conceptual distinction between motivating and justifying reasons. Reference SussmanDavid Sussman (2001) uses Kant’s treatment of the passions to account for weakness, whereas Reference FriersonPatrick Frierson (2014) argues that weakness is a defect of volition that is opposed to passions. However, very few of the available solutions are based on Kant’s notion of the moral strength that is constitutive of virtue. This is surprising, especially because Kant explains moral weakness as a mere lack of moral strength (MM 6: 384, 6: 390).

5.2 A Look at Paradigmatic Solutions

The widespread assumption is that Kantian moral weakness can only be exhibited at the level of following morally good maxims. Reference EngstromStephen Engstrom (1988: 441) holds that ‘the frail agent’s weakness is not expressed in any maxim’, and Reference BorgesMaria Borges (2019: 24) claims that weakness ‘is an exception not reflected in the maxim’. Reference McCartyRichard McCarty (1993) argues that Kant’s treatment of weakness enables us to accommodate those cases in which we have a genuinely moral maxim but fail to live up to it. His point is that the weak-willed agent recognizes the moral law as providing a sufficient reason but fails to act morally because his moral feelings do not happen to be sufficiently strong.

However, this need not exhaust the scope of Kantian moral weakness. The fact that Kant addresses weakness as the first stage of our propensity to evil speaks against the idea that there is nothing wrong with the maxims of the morally weak. Kant emphasizes that evil must not be sought in inclinations, but in one’s ‘perverted maxims’: ‘genuine evil consists in our will not to resist the inclinations’ (Rel 6: 58–9). Moreover, Kantian moral weakness cannot be explained in terms of the strength of the incentives that we happen to have. As ‘the impotence of the incentive of reason’ (Rel 6: 59), moral weakness must also be explained as the weakness of the human will in facing temptations.

Some insightful interpretations leave room for weakness at the level of maxim adoption. The inner conflict experienced by the morally weak agent is conceived as a conflict between his good underlying maxim and his morally incorrect particular maxim. Reference HillThomas Hill (2012: 146) argues that the morally weak agent ‘must be viewed as having two conflicting maxims: a basic maxim to conform to morality’s unconditional requirements and a shorter-term maxim reflecting an intention to indulge self-love on the particular occasion’. Robert Johnson similarly argues that the Kantian weak-willed agent’s disposition or underlying maxim is morally good, whereas his specific maxims are not: just like the vicious agent, he incorporates ‘wayward incentives’ into his ‘motives’ (Reference JohnsonJohnson, 1998: 362). Both have and follow morally incorrect particular maxims. When making a snide comment to a colleague, the weak and the vicious are therefore both ‘motivated by a maxim of doing so’ (Reference JohnsonJohnson, 1998: 361–62). The only difference is that the weak agent acts against his own deepest commitments, because he does not have an evil disposition.

First, this kind of philosophically appealing solution seems to work only on the assumption that the morally weak agent’s underlying maxim or disposition is genuinely good. But as I will explain, this presupposition can be challenged by textual evidence. Second, the above-described way of accounting for the inner conflict experienced by the morally weak agent seems to rest on a very sharp distinction between particular and underlying maxims. The latter are thought to be static, for they are meant to represent our deepest commitments, which are made once and for all, independently of our particular maxims. The implication of this view seems to be that one who often adopts and follows specific immoral maxims, might still be said to have a good underlying maxim. Hence, even someone who often intentionally reverses the ethical order of the incentives in his particular maxims, might still be said to have a good subjective, motivating ground of his particular maxims. Third, as I elaborate below, the claim that the weak person adopts the same particular maxims as the vicious person falls prey to the difficulty of explaining why Kant highlights that vice, unlike weakness, is an ‘intentional [vorsetzliche]’ transgression that ‘has become a principle’ (MM 6: 390). If understood as an intentional transgression of the moral law based on a maxim, making a snide comment to a colleague would seem to illustrate the Kantian vice of arrogance rather than weakness. Finally, even if we could distinguish between the vicious and the weak in terms of their different underlying maxims, this would leave no room for the impure. To the extent that the impure agent has a morally good disposition, his failure is conflated with weakness, and to the extent that his disposition is evil, it is conflated with vice.

Reference TimmonsMark Timmons (1994) interestingly applies the distinction between motivating and justifying reasons by arguing that the morally weak person still has an evil disposition. He leaves room for moral weakness at the level of adopting maxims, but his explanation depends on the idea of moral luck. The question is whether such an explanation can accommodate the freedom condition of weakness of will and other responsibility-related issues. Adopting maxims and acting accordingly should not be a matter of luck on a Kantian picture. As Kant suggests, this is better spelled out in terms of self-control, or the lack thereof. In what follows, I examine moral weakness as a failure to properly exercise our capacity for self-control when setting ourselves particular moral ends and realizing them.Footnote ⁶⁴ This examination will help us to understand what is going wrong at the level of maxim adoption when it comes to the morally weak.

5.3 A Self-Control-Based Solution

Weakness is a ‘mere lack of virtue [blos Untugend]’, or a mere ‘lack of moral strength (defectus moralis)’ (MM 6: 390; 6: 384). It is possible to lack virtue as moral strength in two ways. One can exhibit either a ‘negative lack of virtue’ or a lack of virtue that is also positive (MM 6: 384). The former is weakness and the latter is vice. If virtue is ‘+a’, then weakness is ‘0’ and vice is ‘−a’ (MM 6: 384 and Rel 6: 22 n).

Vice is an intentional transgression of the moral law based on a morally incorrect maxim, in which one consciously reverses the ethical order of the incentives. The vicious agent is aware of the moral law, but she does not properly incorporate it into her maxims. When adopting her maxims, she starts from the ends that she is anyway eager to adopt and does not constrain herself to adopt moral ends. Just like Kant’s moral egoist, the vicious person ‘puts the supreme determining ground of his will simply in utility and happiness, not in the thought of duty’ (Anth 7: 130).Footnote ⁶⁵

By claiming that the weak agent adopts the same maxims as the vicious agent, we are then conflating a merely negative lack of virtue with a lack of virtue that is also positive. The morally weak agent does not yet seem to locate the determining ground of her will in one of her self-seeking interests. She does not yet seem to adopt maxims on empirical grounds by intentionally reversing the ethical order of her incentives. The maxim that guided Sulla in his bloody revenge against his enemies cannot be attributed to the weak agent. The latter is someone who is willing to help others but fails to do so.

This easily leads to the conclusion that the weak agent has the same particular maxims as the virtuous agent but simply fails to perform certain actions. But if so, it becomes difficult to explain why Kant treats weakness as the first stage of evil. In an attempt to solve this problem, we can best analyse weakness as a lack of moral strength of the human will and maxims (MM 6: 447, 6: 394, 6: 405). For example, the duty of beneficence ‘consists in the subject’s being constrained by his reason to adopt this maxim as a universal law’ (MM 6: 452); it does not require that the agent performs the action of helping whenever he sees someone in need. On its own, the mere omission of an action can hardly count as a lack of Kantian virtue. I believe that Kantian moral weakness is best understood as a mere lack of moral strength that plays out on two levels. After briefly clarifying the weak-willed agent’s lack of self-control as a mere failure to adhere to his maxims, I try to describe his lack of self-control at the level of maxim adoption.

Kant describes weakness as follows:

This passage makes clear that Kantian moral weakness is a lack of moral strength in compelling oneself to realize moral ends. Kant also mentions ‘the general weakness [Schwäche] of the human heart in complying with the adopted maxims anyway [überhaupt]’ (Rel 6: 29; translation modified). The weak agent may lack the skill to compel himself to act a certain way. He might lack the self-discipline necessary to obey rules or fail to acquire a habit of acting a certain way.

The weak agent may also fail to follow his maxims because he has failed to cultivate his capacities in order to use them as means to realize all kinds of ends. For example, by exercising one’s capacity for judging in different situations one becomes more skilful in fulfilling the duties of virtue. The skill of postponing judgement, developed through practice, ‘indicates great strength of mind’ by which we can avoid performing bad actions out of anger (L-Eth 27: 365).Footnote ⁶⁶

The weak agent might also be under the sway of an affect and therefore momentarily incapable of acting in accordance with his maxims. Kant claims that ‘weakness in the use of one’s understanding coupled with the strength of one’s emotions’ is ‘only a lack of virtue [Untugend]’ (MM 6: 408). A lack of moral strength due to affects can be an aspect of Kant’s conception of moral weakness. The weak agent might fail to cultivate his capacity for self-control and therefore descend into affective states that make him momentarily incapable of controlling himself and adhering to his maxims. Furthermore, by allowing his feelings to become affects, he also creates the obstacles that stand in the way of self-determination and proper maxim adoption. And yet, if understood simply as lack of moral strength due to affects, lack of virtue cannot provide a full account of Kantian moral weakness. We also need to address lack of moral strength in setting aside those feelings on which inclinations are based. In this way, one cultivates or strengthens one’s susceptibility to moral feelings. As elaborated in the previous two sections, this cultivation enables the adoption of maxims of virtue.

A fuller application of what has been said about the connection between moral feeling and moral strength provides us with a plausible portrait of the morally weak agent. If the weak agent fails to cultivate moral feeling and if it is by strengthening our moral feeling that we make our object every particular end that is a duty (MM 6: 387), then the weak agent fails to intend a particular moral end. She takes no interest in particular moral ends because she has failed to cultivate her capacity for moral feeling, conceived as the capacity to take a pure interest (CPrR 5: 79–80). For this reason, she remains unaffected by the concepts of duty. Her general, abstract knowledge of what is morally right or wrong does not effectively motivate her. By failing to cultivate one of the natural predispositions of the mind to being affected by the concepts of duty (MM 6: 399), or the subjective ground on which she is morally to determine her choices, she fails to enter into a state in which the moral law actually determines her power of choice. Her maxims remain weak or impotent in practice.

Finally, Kant also implies that moral feeling is the purest virtuous disposition and that this disposition, conceived as the end of moral perfection, can be attained by cultivating the will or our ‘moral way of thinking’ (MM 6: 387).Footnote ⁶⁷ If so, then the weak agent may fail to achieve this disposition because she fails to cultivate her moral way of thinking by adopting maxims of virtue – the subjective principles of actions through which we actually determine our choices. Since she fails to cultivate her natural capacity for moral feeling, she can be said to set herself the end of moral perfection, but only in the sense of having an overly general, wishful intention to cultivate her will. Her commitment to the moral law remains a wishful moral intention, because she fails to strengthen or renew this general intention by continually exercising her capacity for self-control in order to abstract from sensible impressions. She fails to acquire the strength of the intention that is constitutive of maxims of virtue or maxims of ends.

This puts us in a better position to answer the question of whether the weak agent’s disposition, or her fundamental maxim, is a genuinely good one. I agree with Johnson and Hill that the morally weak agent has a kind of general, pure commitment to do what is right. Unlike the vicious agent, the weak agent can be said to have a general intention to do what the moral law demands. The weak agent wants to do what she ought to do (Rel 6: 29). She can be said to set herself moral ends in a purely intellectual and abstract way. The objective determining ground of her choice might be characterized as good. As Kant explains, the weak agent incorporates the moral law as the good that is ‘an irresistible incentive objectively’ (Rel 6: 29). However, I do not think that this explanation suffices to show that the disposition of the morally weak agent is genuinely good, because her abstractly good commitment to the moral law remains fragile in practice. The Kantian weak-willed agent might take mere wishes, which ‘always remain empty of deeds, for proof of a good heart’ (MM 6: 441). She might deceive herself into thinking that she cannot compel herself to act morally in some situations.Footnote ⁶⁸ Since there is no middle position between good and evil for Kant (Rel 6: 24; 6: 22 n), we must conclude that the weak agent’s disposition is evil.

We may also resist the conclusion that the disposition of the morally weak agent is genuinely good by appealing to Kant’s sophisticated version of the criterion for distinguishing moral good from evil – we incorporate both the incentive of the moral law and the incentives of inclination, so that the difference between good and evil must lie in the way they are incorporated (Rel 6: 36). The suggestion is that being morally good means that one has incorporated the law into one’s fundamental maxim (oberste Maxime) as a by-itself-sufficient (für sich allein hinreichend) determination of one’s choice (Rel 6: 36). The weak agent might then incorporate the incentive of the moral law but fail to incorporate it as a self-sufficient incentive.

The impure agent can be said to do the same, however. Kant points out that his maxims are ‘not purely moral’ (Rel 6: 30). The impure agent intends to comply with the law from morally unacceptable motives, and the weak agent wishes to comply with the law from morally acceptable motives. There is also a further difference with regard to the quality of their maxims. The maxims of the impure agent can be ‘powerful enough in practice’ (Rel 6: 30) in that they result in legally good actions, whereas the maxims of the weak agent are not effective in practice. So, although the weak and the impure can both be said to incorporate the incentive of the moral law improperly, they seem to do so for different reasons and in slightly different ways. The subjective motivating ground of the maxims of the weak agent is pure, but his pure intention to follow the law is ultimately impotent in practice because he seems to postpone the adoption of particular moral maxims that are efficient in practice. The weak agent does not seem to adopt impure maxims based on the incentives of inclination, but he fails to renew his general commitment to the moral law by reassessing his incentives in new situations. The impure agent seems to take a step further to the extent that he adopts impure maxims in which priority is more explicitly given to morally unacceptable incentives to follow the moral law.

We can clarify what goes wrong at the level of maxim adoption when it comes to the morally weak by recalling Kant’s claim that all lawgiving consists of two elements: a law, which represents an action as objectively necessary, and an incentive, which makes that action also subjectively necessary (MM 6: 218). The first element suffices for a possible determination of choice, and the second is required for actual self-determination (MM 6: 218). Accordingly, even though the purely theoretical basis of the morally weak agent’s maxims is good, there may still be something wrong with the subjective, motivating ground of his maxims, which is what enables actual self-determination. Something may still not be totally right with the content or ends of his maxims. The maxims of the weak agent seem to lack a proper element that would make them, as Kant puts it, ‘subjectively practical’ (L-Met 28: 317) or ‘subjectively possible’ (L-Met 29: 900). Objectively, as regards the rule, his maxims are good, but subjectively, as regards the incentive, they are not (Rel 6: 58 n). The incentive of the moral law may be irresistible ideally – ‘in thesi’, but not also ‘in hypothesi’ (Rel 6: 29). In the human condition, the subjective ground of its irresistibility may be weaker than the inclinations. One can will the good but not strongly enough to move one to perform a morally right action. It is through the constant exercise of our capacity for moral self‐control that our maxims become strong enough to result in actions, for this is how we set ourselves particular moral ends.Footnote ⁶⁹

The maxims of the morally weak agent may thus be more like practical laws, which would serve him ‘subjectively as the practical principles’ of his action if his reason were to gain control over his faculty of desire (G 4: 401 n). Since he does not properly exercise his capacity for self-control, his self-imposed rules do not function as volitional principles that actually motivate him to act morally. His maxims are not strong enough to actually move him to act. This may also be why the weak agent is not ‘strong enough to comply with’ his ‘adopted principles [genommenen Grundsätze]’ (Rel 6: 37).

As I have shown, it is through the constant exercise of our capacity for moral self-control, or through the acquisition of moral strength, that our maxims become the principles that actually guide our actions in practice. It is only via a constant effort to set aside the influence of sensible impressions on our mind that we can set ourselves moral ends in new situations. My point has been that the intention the strength of which Kant calls virtue (MM 6: 390) is an intention by which we set ourselves particular moral ends and that our particular maxims must include this intention if we are to be motivated by the pure moral incentive to perform an action. Without such an intention, our maxims would not be subjectively practical principles of our own volition.

On the basis of these considerations, I now conclude that the morally weak agent lacks virtue, understood as acquired moral strength. He lacks the strength of intention needed to set himself moral ends in ever-new situations. For this reason, the weak agent’s maxims do not seem to be subjectively practical. The agent who does not acquire moral strength by properly exercising his capacity for self-control is weak. He fails to make a continuous effort to sustain the firmness of his general moral intention in the face of contrary inclinations. The subjective determining ground of his choice remains impotent in practice, because he fails to continuously use his capacity for self-control to put aside the incentives of inclinations. By failing to gradually acquire virtue, he fails to restore his original predisposition to the good. This reading is in agreement with Kant’s point that from our own perspective, the reformation of our propensity to evil as a perverted way of thinking (verkehrter Denkungsart) must be gradual because we can judge the strength of our maxims only on the basis of the control over the input of sensibility that we gain over time (Rel 6: 48).

The weak agent may not be strong enough to comply with the principles that he endorses on a purely theoretical level because he seems to postpone making these principles subjectively practical. For example, although he generally holds that he should help those in need and makes the happiness of others his end in abstracto, this does not sufficiently motivate him to help others. The rule ‘help others in need’ does not become a subjectively practical principle that moves him to perform the relevant actions.

Once he succumbs to the temptation to adopt impure maxims, he becomes not only weak but also impure. To use the same example, he adopts the maxims that often result in acts of helping others, but he does not do so ‘from duty’. But even if the weak agent does not become impure, his way of thinking cannot rightly be characterized as virtuous, because he fails to carry out his overly general commitment to the law in practice. He fails to continuously renew his commitment to the moral law by adopting virtuous maxims of ends. This is to say that he fails to cultivate the deep motivating subjective ground of his maxims by reassessing his incentives in different situations. For this reason, the subjective principle of his particular maxims, or his disposition, remains impotent in practice.

5.4 Favourable Implications

My reading has promising implications. First, it opens up the possibility of accounting for the inner conflict experienced by the morally weak agent without making problematic assumptions about her possession of a good underlying maxim and her morally incorrect specific maxims. The weak-willed agent has the will to do what the moral law requires in abstracto, whereas in concreto she fails to make an effort to strengthen her will. Second, the proposed reading does not commit us to the view that the morally weak agent simply changes her mind by dropping her adopted principle. Third, it accommodates Kant’s suggestion that moral weakness is expressed both at the level of adopting maxims and at the level of following them. The Kantian morally weak agent is also someone who lacks moral strength in constraining herself to adopt particular maxims that are powerful in practice. She wills the good but lacks a settled intention to determine her choice by diverting attention away from her inclinations and focusing on a particular moral end. She can be described not only as lacking an intention to perform an action in order to follow a maxim but also as lacking the intention that is essentially involved in adopting maxims of virtue. Fourth, by addressing the neglected connections between weakness, moral strength, moral feeling and the activity of setting ourselves moral ends, this reading highlights important aspects of moral weakness that have been overlooked thus far. Fifth, it enables us to distinguish the weak from the impure and the vicious. Finally, it does not compel us to abandon the incorporation thesis in order to save the phenomenon of weakness of will, and it can tell us why moral weakness counts as moral evil. In a certain way, the morally weak agent takes an active stance regarding her inclinations. However, she still fails to enter into a state of actual self-determination because her moral intention is not firm in the sense of continually engaging in the purifying (or self-controlling) activity that the free adoption of maxims of virtue requires.