Current philosophical debates revolving around the notion of moral character are dominated by a conception of character that originates in the Aristotelian tradition. This is true both for debates that engage the notion of moral character in an affirmative sense and for those that are driven by a largely critical impetus. An example of the former is the virtue ethical branch of contemporary moral philosophy;Footnote ¹ an example of the latter are current discussions of the so-called situationist challenge for the notion of character both in philosophy and social psychology.Footnote ² From the point of view of those who participate in these debates, one strong reason for this dominance is that the Aristotelian picture presumably reflects our pre-theoretical, common-sense conception of moral character. However, arguably this holds the other way around as well: given how theoretically influential it has been – not only in philosophy but in adjacent disciplines such as psychology and social psychology as well – it seems plausible to assume that our common-sense conception of moral character by now is strongly shaped by Aristotelian intuitions.

What are central tenets of the Aristotelian conception of character? We must be careful here to keep apart the simplified version of this conception as it has reached and continues to influence the theoretical mainstream from Aristotle’s proper view. The latter is formulated in a number of philosophical texts of great complexity and intricacy, most prominently the Nicomachean Ethics, and has been the subject of controversial discussion among commentators for centuries. In contrast to the latter, the former can be more easily summarized in a number of well-defined central theses. According to the folk-psychological/Aristotelian conception, a person’s moral character consists of a set of moral character traits, the virtues and vices. Familiar examples of virtues are temperance, courage, and generosity. Corresponding vices include intemperance, cowardice, and avarice. A virtue or vice is a disposition – put in more traditional terms, a habitus or hexis – to act, choose, or feel in a certain way. Thus, a virtue or vice is a causal power that determines a person’s actions, decisions, and emotions to a certain degree, in conjunction with relevant stimulating events or circumstances. While character traits thus have a certain deterministic power on the present view, this must not be understood as implying that they necessitate actions, decisions, or emotions, given the appropriate stimulating conditions. Rather, the view must leave room for the possibility that people sometimes act out of character, for instance, when they are under the sway of a strong passion, and hence that the causal power of the character disposition may be overruled by a stronger causal power. Nevertheless, on the present view, character functions as a default cause of action, such that all else equal we can explain and predict actions, choices, and patterns of behaviour by reference to character. This gives the Aristotelian conception of character a powerful theoretical function: moral character traits have both an explanatory and predictive function. We may say, for instance, “Anna gives 5 per cent of her gross income to charity because she is benevolent.” Here the ascription of a character trait is meant to explain why a certain action was performed. Furthermore, we can refer to an individual’s character traits in order to predict her actions. We may say, “Next year again, Anna will spend 5 per cent of her gross income to charity because she is benevolent.”

A second crucial aspect of the Aristotelian conception concerns the way moral character traits are acquired. From an Aristotelian point of view, it is clear that habituation plays an important role in the acquisition of both virtue and vice. One important implication of this is that whether or not an adult person ends up being virtuous or vicious is to a significant degree due to factors beyond her control: it depends heavily on her education, on whether or not the society and socio-cultural micro climate that surrounded her during her most pliable years was conducive to the acquisition of virtue or its opposite.

Third, since on the Aristotelian conception a character trait is a habitus, there is a certain diachronic requirement of repetition, of stability, and persistence implied by this. A person who performs one generous act does not thereby qualify as generous; nor would we label a person as avaricious merely based on one singular act of stinginess. Character traits, both morally praiseworthy and deplorable ones, bring with them repetition and persistence of performance. A generous person will repeatedly and reliably perform generous acts; an avaricious person will regularly plague her fellow human beings with acts of stinginess.

Against the background of this sketch of the (folk psychological version of the) Aristotelian conception, we can bring into view a further reason for its prevalence in our philosophical and pre-philosophical discourse: the conception is continuous in central respects with the predominant scientific conception of character. The current mainstream view in psychology and social psychology is dominated by a conception of character as a functional construct. On this view, a person’s character consists in a set of traits or dispositions concerning personal preferences and beliefs, which manifest themselves reliably and consistently in observable actions and behavioural patterns.Footnote ³ Methodologically, this means that character can be described in behavioural-functional terms and empirically observed in psychological studies.Footnote ⁴ A fundamental assumption in this view is that a person’s character must be sufficiently persistent and robust in order to allow for cross-temporal stability and cross-situational consistency, and hence for its ascription to a person from a third-person standpoint. This, in turn, makes explanations of observed behaviour and predictions of future actions possible.Footnote ⁵

To emphasize again, Aristotle’s position is more refined than and diverges in important respects from the view just sketched. As we shall see in more detail in Chapter 1, for Aristotle, the virtues and vices are not just behavioural dispositions but dispositions to undergo certain emotions, feelings, and desires – as well as, crucially, dispositions to choose certain actions in reaction to such feelings.Footnote ⁶ The latter aspect is made explicit in Aristotle’s thesis that a virtue or vice is a hexis prohairetike, a disposition involving choice.Footnote ⁷ Nevertheless, the Aristotelian understanding of virtue and vice converges with both the folk psychological and the scientific conception of (moral) character on the idea that an individual’s character is a disposition that determines her actions in a broadly causal way in conjunction with certain stimulating conditions or events.Footnote ⁸ This is precisely what makes it possible for the notion of character to fulfil both an explanatory and predictive function according to all of these views. Because character determines action, we can explain (all else equal) why a certain action occurred under certain circumstances by referring to the agent’s character. Furthermore, given that we know that a person has a certain character, we expect her to act in certain ways and may therefore predict (all else equal) her future actions under certain circumstances.

Although the Aristotelian view holds a central position in contemporary discussions of character, both in philosophy and beyond, it is worth noting that the past decades have seen a growing interest in Kant’s conception of moral character, especially in his conception of virtue.Footnote ⁹ At the same time, this increasing interest in Kant’s conception of virtue was to a significant degree shaped by the intention to demonstrate that when it comes to their conception of virtue and character, there is an unexpected theoretical proximity between Aristotle and Kant. Thus, on the view that has emerged from this field of research, Aristotle’s and Kant’s approach to moral philosophy share much more in common than the traditional textbook contrast between Kantian deontological ethics, on the one hand, and Aristotelian virtue ethics, on the other, suggests – and this is true, in particular, concerning their respective conception of virtue and moral character.Footnote ¹⁰ This movement has contributed greatly to shedding novel light on Kant’s moral philosophy and dispelling widespread prejudices about it. Moreover, this approach has helped to draw our attention to Kantian textual sources that used to live in the shadows of his major critical moral philosophical works, the Groundwork and the Critique of Practical Reason: notably Kant’s late moral philosophical works, such as the Doctrine of Virtue in his Metaphysics of Morals and the hauntingly obscure Religion within the Boundaries of mere Reason, as well as his writings and lectures on anthropology.

In this vein, Barbara Herman and Nancy Sherman, for instance, have argued that a Kantian moral agent, similarly to the Aristotelian virtuous agent, needs emotions in order to make moral judgments that can guide her actions in a competent way.Footnote ¹¹ Pointing out parallels between the Kantian and the Aristotelian conception of virtue acquisition, Robert Louden has drawn attention to the ‘impure’ aspect of Kant’s moral philosophy contained in his anthropology, emphasizing that morality for Kant has historical, educational, and generally anthropological preconditions.Footnote ¹² Other commentators have worked on developing a distinctively Kantian conception of moral character traits, specifically of Kantian virtue.Footnote ¹³ On the whole, this trend in Kant scholarship has helped to make generally acceptable the view that Kant, like Aristotle, is concerned with developing a rich, comprehensive conception of virtue that draws on considerations regarding human nature and holds that fully developed human virtue comprises dispositions to feel and act, moral interests, cognitive capacities, and acquired skills of moral judgment.

However, in contrast to this trend, other commentators have pointed out that there is a potential tension between this line of interpretation and those aspects of Kant’s moral philosophy that are essentially connected to his transcendental methodology, in particular his emphasis on the reciprocity of morality and freedom. Paradigmatic for this approach is Patrick Frierson’s book Freedom and Anthropology in Kant’s Moral Philosophy.Footnote ¹⁴ Frierson tackles in particular the notion, stressed by some of the Kantian virtue theorists just mentioned, that broadly anthropological sources and considerations ought to be considered as forming a constitutive part of Kant’s moral philosophy. As Frierson emphasizes, one of the uncontested tenets of Kant’s moral philosophy is that the moral law’s demand is unconditionally valid, that is, valid for every rational agent at any time and under all circumstances. On Kant’s account, this view is inseparably linked to two further central claims of his moral philosophy, which together constitute his theory of freedom. One is the thesis that moral agents are transcendentally free, that is, free from causal predetermination through natural laws. That we are transcendentally free and subject to the moral law, for Kant, is something that can be determined a priori, on the basis of theoretical reason (which assures its possibility) and practical reason (which assures its reality). The second is the thesis that can be summarized in the slogan “You can because you ought to”: Kant believes that every rational agent can at any time heed the call of the moral law by performing morally worthy actions. This means that whether or not an agent has a good will and performs actions of proper moral worth is solely a matter of whether she makes the right choices from the right motives. In contrast, if one holds, as some Kantian virtue theorists do, that in order to perform actions of proper moral worth an agent needs virtue understood as a trait grounded in anthropological, educational, historical, or social conditions, this would seem to conflict with Kant’s emphasis on the freedom of moral action. For, in this case, it follows that only agents who fulfil the specified conditions are subject to the moral law and in a position to properly heed its call. Accordingly, Frierson argues, commentators who hold that virtue for Kant is grounded in anthropological conditions face a dilemma: either they must deny a central tenet of Kant’s moral philosophy or they must concede that for Kant anthropological, educational, historical, or social conditions can only render indirect, optional support to morally worthy actions, rather than being necessary conditions for or constitutive elements of it.Footnote ¹⁵

Frierson focuses primarily on the notion that virtue is grounded in anthropological conditions, but it should be noted that his line of criticism can be expanded to apply to other aspects of the interpretation of Kant’s conception of virtue just outlined. In each case, we end up facing a similar dilemma. Consider the notion, for instance, that virtue is acquired through a process of self-cultivation and habituation, an idea that is emphasized in Aristotelian virtue theory. If it is held that one needs acquired virtue in order to perform actions of proper moral worth, again, this would seem to conflict with the idea that any rational agent can perform a morally worthy action by acting from duty at any time. Alternatively, one may concede that acquired virtue is merely ancillary to the performance of actions of proper moral worth. Most strikingly perhaps, the notion that a virtue is a habitus that determines our choices and actions in a causal sense is strikingly at odds with the Kantian creed that in our moral choices we are fundamentally free in the sense of not being predetermined by any causal powers. Again, one may alternatively concede that a habitus for Kant can never be constitutive of a moral character trait as such; it can merely render support to the speedy and efficient execution of a moral choice which is itself undetermined and free.

What we should take away from this brief and rough sketch of a series of interpretive tensions is that when seen from a certain point of view, the differences between Kant’s and Aristotle’s conception of virtue are more striking than the similarities. Furthermore, this is the point of view that puts centre stage the most pivotal aspects of Kant’s transcendental moral philosophy, in particular his emphasis on the reciprocity of morality and freedom. At the same time, the commentators mentioned earlier are undoubtedly right to emphasize that the notion of moral character and related concepts, such as virtue and vice, are being given increasing attention in Kant’s later moral philosophical works. One cannot deny, then, that Kant was increasingly concerned, in these later works, with developing a theory of moral character and virtue.Footnote ¹⁶

Where should we go from here? It seems that we have reached an interpretive impasse. The notions of virtue and moral character do seem to play an important role in Kant’s moral philosophy, especially in his later works. At the same time, at first sight it is hard to make sense of quite what theoretical role it is supposed to play without coming into conflict with central tenets of Kant’s overall philosophical enterprise. Note that what has led us into this impasse was the more or less tacit assumption that we must look at Kant’s theory of moral character through a broadly Aristotelian lens, thus arriving at a picture according to which these two philosophers’ conceptions of moral character and virtue can be seen as being much closer to each other than previously assumed. I suggest that we should consider putting this assumption aside and instead try to explore what Kant’s unique contribution to our understanding of moral character and virtue might look like, thus delineating a uniquely Kantian theory of moral character and virtue. By that, I mean two things: first, a theory that is not driven by the explicit or implicit assumption that the resulting theory can be expected to bear close resemblance to the Aristotelian conception; second, a theory that is grounded in and arises from the central tenets of Kant’s transcendental moral philosophy. Most importantly for the purposes of my project, it is to be grounded in Kant’s view of morality as based on an unconditional law of pure practical reason that moral agents are always at liberty to follow in order to thus bestow full moral worth on their actions.Footnote ¹⁷

There are two worries that might lead one to be hesitant about such a project. The first one is that, as we have seen, on the Aristotelian conception, the notion of moral character in general and of virtue in particular fulfils a well-defined theoretical function: its function is to explain and predict actions. If the Kantian theory turns out to diverge radically from the Aristotelian one, will it still be able to fulfil this function – and if not, which theoretical function, if any, is it going to fulfil? The second, closely related worry is that if Kant’s conception of moral character and virtue looks nothing like Aristotle’s, is it going to be even recognizable as a theory of moral character? Both of these worries bear witness to the extent to which our thinking about moral character is influenced by the Aristotelian conception. We have become so used to thinking about moral character along Aristotelian lines that it seems to us that unless Kant’s conception is close enough to the Aristotelian one, it will have no significant theoretical function and amount to a view that effectively reduces the notion of moral character to something else.

I want to show in the following that both of these worries are unfounded. There is theoretical room for a uniquely Kantian conception of moral character. To be sure, this conception will appear unfamiliar in some respects. Most importantly perhaps, in stark contrast to Aristotle’s theory, it conceives of the moral agent as a human subject whose principal moral task consists in heeding as conscientiously as possible the call of an unconditional practical law. At the same time, the human being is, for Kant, essentially a creature who is subject to inclinations that may diverge from the moral law and thus, depending on the situation, present strong incentives for moral transgression. In fact, virtue for Kant, most fundamentally, consists in a “moral disposition in conflict” (CprR, 5:84): it takes shape, essentially, as a conflict between the incentive to follow the moral law and diverging incentives of self-love. In Kant, we have to understand the notion of moral character as being situated against this background of continuous conflict and struggle. One of my central theses in the following is that this means that from a Kantian point of view, in contrast to Aristotle, the definite and complete acquisition of a good or virtuous character is essentially out of reach. Such a character, by contrast, is something that we must continuously and progressively strive for. In this striving, the idea of moral completion can serve only as a guiding light, an ideal that unifies our particular efforts and gives them a sense of direction. At the same time, while each step seems to bring us closer to this end, it also pushes it further away again, like a horizon towards which we are walking. This is not merely an epistemic point: the issue is not just that we never know for sure in Kant’s view whether or not we really are good. Rather, there is a kind of ontological indeterminacy here: for Kant, we never are good but permanently on our way towards being good. In Kant, then, the notion of moral character assumes a radically different function than on the Aristotelian conception. Not only is it a practical rather than a theoretical concept: that is to say, its function is to guide us in our moral agency, rather than to feature in theoretical knowledge claims about moral agents (ourselves or others). Furthermore, there is something specific about the way it is applied in our moral practice, which can be captured by the Kantian conception of a regulative idea: we look towards it for our practical orientation as an ideal towards which we strive, rather than as a category that we apply to ourselves or others in order to determine someone’s moral status in a definitive way.

Kant holds that the notion that any moral law worthy of the name must be one that commands our unconditional respect perfectly captures our modern common-sense understanding of morality. Accordingly, if the Kantian theory of moral character is just the theory of moral character we need in order to accommodate this understanding, this may be a strong reason in its favor. In any case, we may begin to consider the possibility at this point that our modern common-sense pre-theoretical, intuitive conception of moral character is not perfectly aligned with the Aristotelian one.

In sum, it may well turn out that a uniquely Kantian theory of moral character not only is theoretically viable but also has a lot more intuitive appeal than may appear at first sight. I hope that this study will contribute to showing that this is the case.

Overview of Individual Chapters

Chapter 1, “The Significance of Character”, serves as a reference point for the subsequent discussion of Kant in two ways. First, it systematically examines a reciprocal relationship between the idea of committing oneself to a normative principle of action and the concept of moral character. On one hand, by committing to a normative principle, a person not only decides how to act in the present moment but also views this decision as a precedent for future actions, thus projecting themselves as a consistent character into future situations of choice. On the other hand, to genuinely commit to a principle, a person must act in accordance with it consistently and non-accidentally. As will be explored in later chapters, Kant can be interpreted as defending similar claims, thereby aligning with our common-sense understanding of character. Second, the chapter offers a more detailed outline of the Aristotelian conception of moral character and its underlying motivation, also linking character to commitment to a normative principle. This outline serves as a contrast for the discussion of Kant in the chapters that follow.

Chapter 2, “Setting the Stage: Kant and the Puzzle of Moral Character”, sets out by bringing into view a problem revolving around Kant’s conception of moral character: at first sight at least, the very idea that the notion of moral character plays a significant theoretical role in Kant’s moral philosophy is at odds with central Kantian theses concerning the nature of morality. Most importantly, Kant is unwaveringly committed to the principle “You can because you ought to” – every rational agent can in any situation freely decide to perform a morally worthy action, an action from duty. This idea seems to sit uneasily with the notion that one needs an enduring character acquired through practice and training in order to perform a morally worthy action. This tension is made explicit in an interpretive dilemma similar to the one sketched above: as a commentator of Kant, one appears to be faced with the choice of either relegating the notion of moral character to a secondary place in Kant or accepting that Kant has a highly revisionary understanding of what moral character consists in. The chapter also gives a suggestion for how to move forward from here, which points the way for the following chapters: the solution lies in understanding that and how in Kant the very idea of moral choice and the moral worth of individual actions is inherently tied to a conception of moral character.

Chapter 3, “Intelligible Character, Empirical Character, and Moral Character”, starts out with a discussion of the metaphysical foundations of Kant’s conception of moral character as set out in his Critique of Pure Reason. According to Kant, we must ascribe to ourselves and others an intelligible character over and above our empirically observable and causally determined empirical character in order to be able to think of ourselves as transcendentally free or spontaneous. Our spontaneity manifests itself in our capacity to make choices, the capacity of Willkür, and is complemented by our ability to determine ourselves in accord with normative reasons, instead of being determined by causal natural laws. The chapter outlines the development of Kant’s conception of intelligible character, starting from a broad understanding of intelligible causality as presented in the Critique of Pure Reason, and progressing, in his later moral philosophical works, towards a morally charged concept that aligns more closely with the notion of character as an individual’s moral identity exhibiting a certain diachronic stability. Kant’s term for this morally charged concept of intelligible character is ‘Gesinnung’.Footnote ¹⁸ Therefore, the chapter examines the etymology, semantics, and usage patterns of this term in Kant’s writings and lecture courses from the 1770s until the 1790s. The chapter concludes by posing the question of Kant’s philosophical motivations for not only imparting a moral dimension to the concept of intelligible character but also for envisioning it as a persistent moral identity; this question will be addressed and answered in the following chapters.

Chapter 4, “Kant’s Gesinnung: From Groundwork to Religion”, shifts focus, transitioning from the metaphysical foundations of Kant’s conception of moral character and Gesinnung to its moral philosophical justification and implications. I argue that that the notion of the Gesinnung is introduced by Kant already in the Groundwork of the Metaphysics of Morals in order to fulfil the crucial theoretical function of ascribing moral worth to an agent’s individual actions: an action is morally worthy if and only if it expresses a good Gesinnung. I then outline how this conception of the Gesinnung remains consistent while also being gradually enriched and elaborated throughout Kant’s subsequent moral philosophical works, notably the Critique of Practical Reason and the Religion within the Boundaries of mere Reason. Insofar as the Gesinnung constitutes the moral dimension of a person’s intelligible character, it must be conceived of as an object of free choice. For Kant, the objects of choice, more specifically, are principles of action, or maxims. Thus, the chapter explores Kant’s intricate, two-tiered theory of moral choice, according to which a person’s Gesinnung is expressed in the most fundamental, inner maxim guiding their actions.

In Chapter 5, “Against Foundationalism about Moral Character”, I argue against what I call a foundationalist reading of Kant’s conception of moral character or Gesinnung. Central to this reading is the view that moral character functions as a ground that precedes and pre-determines particular moral choices and actions. This is one of the crucial claims of the Aristotelian conception of character as well: according to the Aristotelian conception, actions and decisions can be explained and predicted with reference to a person’s character, because character precedes and predetermines particular moral choices and actions. I discuss two types of foundationalist readings of Kant. According to dispositional foundationalism, moral character for Kant is a non-empirical, intelligible kind of disposition that determines an agent’s choices in a quasi-causal way. According to principle foundationalism, moral character for Kant consists in the choice to commit oneself to following a fundamental principle that precedes and pre-determines subsequent moral decisions. This is the interpretation put forward by most contemporary readers of Kant’s conception of moral character. I present arguments against both versions of the foundationalist reading of Kant’s notion of the Gesinnung.

Chapter 6, “Complete and Incomplete Virtue”, introduces an alternative, non-foundationalist reading of moral character in Kant. Three elements are central to this reading. First, a morally excellent character or Gesinnung, on this view, is not fixed prior to an agent’s particular moral choices and actions, such that it can precede and pre-determine them. Instead, it only takes shape gradually and progressively through a series of particular moral choices. Before the series of particular choices has been completed, the definite moral quality of a person’s moral character – good or evil – is not yet fully determined. Second, in this gradual progression of particular choices, the idea of a completed, fully determined excellent moral character, of fully accomplished virtue, serves as a regulative guiding thought for the moral agent. Third, the completed excellent moral character has a holistic structure, in the sense that in this structure the moral value of a particular action depends on the whole of which it is a part.

Chapter 7, “Kant on Hard Moral Choices”, discusses an essential feature of human moral life according to Kant. For Kant, human moral life is an ongoing struggle, not only in the sense that we must fight to uphold our moral resolutions once made but also in the sense that the question of what the right resolutions are can never be fully or definitively resolved at any point in life. In other words, human moral agents for Kant routinely face hard moral choices in the form of choices between being moral or pursuing one’s own interest. This amounts to an additional argument against principle foundationalism: according to the latter view, once we have made a fundamental moral choice, we only ever have to fight to remain strong enough to stay faithful to this choice; the choice itself is no longer contested afterwards. However, I argue, this fails to adequately capture the Kantian conception of moral trial as a contested and therefore hard choice, rather than as a struggle to maintain the strength to carry out the choices already made.

Chapter 8, “Kant and Aristotelian Neo-Kantianism”, raises the question of whether in Kant’s view there is a psychological mechanism or entity that renders an individual’s good Gesinnung abiding and stable. I consider a number of recent Aristotelian interpretations of Kant that aim to identify a parallel between Kant and Aristotle in this regard: they seek to identify psychological structures inherent in a person’s empirical character that render support to a person’ good Gesinnung. I argue against these readings, showing that such psychological structures cannot render support either in the sense of necessary and sufficient conditions or in the sense of enabling conditions for excellent moral choice. What results from this is a picture of human moral identity as fundamentally precarious and exposed to uncertainty, which is radically unlike the Aristotelian conception. I finish by showing how this Kantian picture naturally opens up a space for divine grace to intervene in human moral life.