1. The Dilemma for Virtue Cultivation

If you are aiming at cultivating a virtuous character, how should you conceive of this aim? Should you consider your own virtue to be your ultimate goal? Or should you consider it merely a means to your ultimate goal of good behaviour? This seems to be a dilemma. The first option seems unattractively narcissistic, since it sets oneself at the centre of all of one's ethical concern. The second option seems to embody a contradictory view of one's own action. It presupposes that action is under direct deliberative control, for if this were not the case there would seem to be no point in deciding to improve one's character in some particular way. But it also seems to involve the idea that behaviour flows directly from character dispositions, since that is the reason for paying ethical attention to these dispositions.

Thus it seems that if one aims at improving one's own character, one is engaged either in narcissism or in doublethink. These charges will be explored in more detail in the next two sections. But first, it is important to see that this dilemma does not track the metaethical distinction between ethical theories that treat virtue as normatively prior to good action and those that treat good action as normatively prior to virtue. We have seen that an ethical theory can be self-effacing, meaning that the value that it postulates as good is not one that the agent ought consciously to pursue. The claim that virtue is normatively fundamental, therefore, does not entail that one ought to treat one's own virtue as an end in itself. It does not entail that one ought to consciously pursue one's own virtue at all, and it does not rule out that one ought to consider one's own virtue only as a means to good action. Likewise, the view that virtue is good only because it leads to good action entails neither that one ought to cultivate virtue nor that one ought not aim at virtue as an end in itself.

To put this point another way, each side in the metaethical debate is faced with both of the options that the dilemma presents, but each side also has the further option of arguing that one should not treat one's own virtue as a goal at all. Philosophers have traditionally thought of virtuous character traits as dispositions that develop through habituation. One develops the disposition towards a certain kind of action by doing that action. ‘For example, people become builders by building, and cithara-players by playing the cithara; so too’, argues Aristotle, ‘we become just by doing just things, moderate by doing moderate things, and courageous by doing courageous things’.Footnote ⁴ Moreover, current work in experimental psychology provides plenty of evidence in favour of the view that character traits in general, not just virtues, develop in this way.Footnote ⁵ Any version of the claim that virtue is ethically important, therefore, could be combined with this view of character development to produce the practical ethical injunction to forget about one's character and aim only at acting in the right way.

The dilemma is intended to constitute the basic argument for the claim that virtue ethics needs to be self-effacing. This is because it is intended to incorporate the various different ways in which philosophers have objected to the idea that one ought to aim at cultivating one's own virtue. As we will see, each of these objections is ultimately either a form of the charge of narcissism or a form of the charge of doublethink. Some of these objections have been voiced as objections to metaethical claims about the normative status of virtue. But the fact that ethical theories can be self-effacing precludes conclusions on this matter being drawn so directly from considerations that essentially concern only the conscious pursuit of virtue.

This limitation on the impact of these objections ought to be evident from the fact that they are all specific to first-personal virtue cultivation. They are objections only to pursuing one's own virtue as an ethical goal, as we will see. The idea of virtue is not generally restricted in this way. We can, and perhaps should, aim at shaping the characters of others through our educational and penal systems, and less formally through the ways in which we raise our children. Objections specific to the conscious cultivation of one's own character could hardly be understood to undermine the pursuit of these goals, so could not undermine virtue ethics in general. Grouping these objections together to form a dilemma for first-personal virtue cultivation, moreover, not only maintains the proper focus of these objections. In so doing, it also helps to make clear their common weakness. For it will enable us to see that they rest on a mistaken idea of the first-personal perspective on character.

2. The Charge of Narcissism

Virtuous action should not be understood as action that explicitly aims at virtue. Although it might sometimes be appropriate to think of one's possibilities in terms of which would be the most virtuous, action can manifest virtue without the agent thinking in terms of virtue. Indeed, in some cases it seems essential to virtuous behaviour that one does not think of one's action as virtuous. The genuinely honest person behaves honestly because they see that as the right thing to do, not because they want to possess the quality of honesty. If modesty is a virtue, it is one that would seem inconsistent with the agent explicitly aiming to behave modestly and then recognising their achievement of modesty. Virtues, like character traits in general, are dispositions to perceive the world in certain ways, to think and feel in certain ways about it, and to act as a result of these perceptions, thoughts, and feelings. The dispositions that thus structure one's practical outlook need not, and in some cases should not, feature in that outlook.

Since these dispositions are developed through habituation, moreover, one can work at becoming virtuous by thinking only about the demands of one's situations, the possibilities that are open, and how best to respond to these. At least, one can do all of this without the idea of virtue featuring in unreflective engagement with the world. The charge of narcissism arises from the role of reflective deliberation in virtue cultivation. In order to try to become virtuous, one needs to reflect on oneself to identify one's shortcomings and to decide on appropriate strategies for self-improvement. It is here that the direction of one's ethical attention looks questionable.

Charles Larmore casts the problem as concerning the relation between reflective self-criticism and ethical engagement in the world. Since virtuous behaviour, such as honest or courageous behaviour, is not action explicitly aimed at manifesting virtue, Larmore suggests that ‘we can only cultivate our virtue on condition that we end up no longer thinking about it at the very moment when it is time to act’.Footnote ⁶ While that seems descriptively true, it does not allow the project of virtue cultivation to escape the charge of narcissism. For if one's purpose in reflective deliberation concerning one's own ethical performance is self-improvement as an end in itself, then one's ethical attention is ultimately directed towards oneself. Since ethics is essentially concerned with other people, that does seem a misdirection of ethical attention.

This worry about narcissism lies at the core of two recent objections to virtue cultivation. One is the objection that virtue ethics is a form of egoism, since it places one's own good at the centre of ethical concern. Thomas Hurka offers one form of this objection against theories that ground the normative priority of virtue in the claim that virtue is essential to flourishing. Ultimately, he argues, this entails that right action is right only because of its contribution to the agent's own flourishing.Footnote ⁷ Julia Annas has responded that this objection assumes that virtue is aimed at as a means to some further end of flourishing, which is itself a benefit to the agent. Those theories that recommend virtue for its own sake and those that identify flourishing with virtue, however, do not see virtue as instrumental to the agent's own good.Footnote ⁸

However, even if we accept this response, there remains something ethically troubling about treating one's own virtue as one's ultimate goal. The essence of the problem does not seem to lie in metaethical considerations of the relations between flourishing, virtue, and right action. This is because these considerations are distinct from the motivational question of how the agent conceives of the aim of virtue. The core of the egoism objection, then, is not that ethics is not about benefitting oneself, but that the ultimate terminus of one's ethical gaze should not be oneself.

The same seems to be true of the objection Williams raises, that virtue cultivation is ‘priggish’. Williams develops this by arguing that concern with virtue is ultimately concern with one's image. ‘Thinking about your possible states in terms of the virtues is not so much to think about your actions’, he claims, but ‘is rather to think about the way in which others might describe or comment on the way in which you think about your actions’.Footnote ⁹ There are two forms such a concern might take. One would be a concern with the opinions that actual other people hold about oneself. It would be the height of vanity to make this one's primary ethical concern. But vanity is an impure form of narcissism, one that involves a submission to other people's standards of appreciation. Narcissus was interested in his own beauty itself, not in anyone else's appreciation of it.

The other form of concern with image is a purer narcissism. It is a concern with the details of one's behaviour visible to others, irrespective of the evaluative perspective anyone might have on these details. This is the heart of the objection Williams raises, which is why he describes it as ‘priggishness’ or a self-righteous attention to detail. For his argument is that character trait terminology only describes a person as they appear from ‘the external point of view’. Just as Narcissus could only appreciate his physical beauty by looking at his reflection, Williams argues that you can only become aware of your own character traits by reflecting on yourself from an external perspective. It is a concern with this public image itself, rather than with anyone's evaluation of it, that is central to ethical narcissism.Footnote ¹⁰

Narcissism, then, is setting oneself as the ultimate goal of one's ethical endeavours. This does not require thinking of oneself in all that one does. Neither does it require aiming for some identifiable benefit to oneself. But it does mean that when one reflects on one's behaviour, one's ultimate purpose is a concern with oneself. This can take the forms of egoism or vanity, since it could be concern for one's own happiness or other people's appreciation of oneself, but it need not take either form. At its purest, it is just making oneself the terminus of one's ethical attention.

3. The Charge of Doublethink

One can avoid the narcissism objection by cultivating virtue not as an end in itself, but as a means to the goal of better behaviour. But this seems to involve a kind of doublethink, which George Orwell defined as the ability ‘to hold simultaneously two opinions which cancelled out, knowing them to be contradictory and believing in both of them’ and ‘to forget whatever it was necessary to forget, then to draw it back into memory again at the moment when it was needed, and then promptly to forget it again’.Footnote ¹¹ In this case, one needs to see oneself from an external point of view as a bundle of dispositions that cause one's actions, since the purpose of working on those dispositions is to alter one's patterns of behaviour, but from the internal point of view of agency one must regard one's behaviour as responsive to the reasons one considers in practical deliberation, otherwise this reflective deliberation about one's character could not result in actions that would lead to a change in character.

The problem is not that there is some outright contradiction between the view of one's agency available from the internal point of view and that available from the external point of view. It is not, for example, that practical reasoning involves an indeterminacy to be bridged by the agent's own commitment or rational endorsement, which is then invisible from the external point of view. For if this were the case then the outcome of practical reasoning could not be predicted on the basis of external knowledge of an agent's dispositions, yet we are able to make predictions of other people's behaviour in this way. These predictions are not always accurate, but this can be explained by the imperfection of our knowledge of the agent's dispositions. The fact that one's predictions of one's own behaviour do not displace the need to decide what to do, moreover, does not show that there is some act of decision or endorsement needed to step from the outcome of one's dispositions to action. It is true that when the prediction concerns oneself one cannot just sit back and watch it come true, but this is only because the prediction was that one would deliberate and decide, rather than merely sit back and watch the action unfold.

Rational deliberation may seem, from the internal point of view, to require a kind of endorsement or commitment that goes beyond the products of one's dispositions, but if so this can only be an illusion generated by an imperfect knowledge of one's own dispositions together with the sense that one is indeed committing oneself.Footnote ¹² This is not to deny that agency involves rational commitment, but it is to deny that this fact engenders any deep asymmetry between predictions of one's own behaviour and predictions of the behaviour of someone else. In both cases, the requirement of rational commitment is presupposed by the prediction. The prediction, that is to say, is a prediction of the rational commitment that the agent will make. The asymmetry is merely due to the superficial fact that some predictions concern a rational commitment that oneself will make, and that is therefore under one's own control, whereas other predictions concern commitments that can only be made by other people, and are therefore beyond one's own control.Footnote ¹³

Although the external point of view on one's agency presupposes that one's actions reflect one's rational commitments, this presupposition is suppressed in the project of cultivating one's own character in order to bring about good actions. The problem is not simply that this project requires one to see oneself from the outside, but rather that it requires a distortion of this view of oneself. This distortion is, as Alan Thomas has put it, thinking of oneself ‘merely as an object’. Thomas characterises this as the failure to recognise that the ‘capacity for executive decision is never determined in advance’ by one's character.Footnote ¹⁴ We have seen that the role of rational commitment should not be understood as bridging some gap of indeterminacy between the outcome of dispositions and action. But there remains an important sense in which the project of cultivating virtue for the sake of good behaviour involves falsely viewing oneself ‘merely as an object’.

It is a sense that is well captured by Thomas in another context. The attempt to treat a prior resolution as something with the power to determine one's action, he argues, is bound to fail precisely because maintaining a resolution requires focusing attention on the reasons for that resolution, the facts that motivated the resolution, which are no longer the focus of attention when one instead thinks about the resolution itself. ‘It is as though one had put in place of one's ongoing rational commitment a mechanism to whose operation one is now indifferent’, and then finds that the mechanism fails to motivate in the way that a rational commitment would.Footnote ¹⁵

Cultivating virtue as a means to good action faces the objection that it involves this same kind of mischaracterisation of one's own dispositions as constituting a mere mechanism productive of behaviour. For the project requires that one recognise certain character traits as virtues. Correctly understood, these virtues are dispositions to recognise certain kinds of situational features as reasons to behave in particular ways. Honesty, for example, is the disposition to endorse certain considerations, such as the importance of telling the truth, as significant reasons for action. If one genuinely wants to cultivate a given virtue, therefore, and understands that virtue correctly, then one already recognises the relevant considerations as important reasons for action. In which case, one should just commit to acting on those reasons. To understand the virtue correctly, that is to say, is to make the aim of cultivating that virtue redundant; one ought instead to aim directly at good behaviour.

The charge of doublethink, then, is the objection that the aim of cultivating virtue as a means to good behaviour makes sense only if one fails to recognise the role of practical reasoning in the manifestation of character. Once one is clearly aware that actions reflect rational commitments, the aim of better action seems better served by making such commitments. This is doublethink rather than mere error because the very role of practical reasoning that one fails to recognise in thinking of one's own character as productive of action is required for the project of virtue cultivation. For unless my actions were determined by my rational commitments, there would be no point in making the commitment to bring my dispositions into line with some theory of good action. For this reason, it seems that if one's ultimate ethical aim is better behaviour, one should just aim to behave better; virtue cultivation would be a misdirection of ethical attention.

4. Reflection and the Practical Role of Character

This dilemma for virtue cultivation rests on a particular account of epistemic access to one's own character. Common to the objections encompassed by the dilemma, that is to say, is the view that one can know one's own character only from the external point of view. The charge of narcissism is essentially that cultivating virtue for its own sake makes oneself, as seen from this external point of view, the final end of one's ethical attention. The charge of doublethink is that the aim of virtue cultivation for the sake of better behaviour requires a distorted vision of one's character from this external point of view, as a mere object causing behaviour, while also holding the contrary supposition that one's actions manifest rational decisions. The claim that one has access to one's own character only from the external point of view has a significant philosophical history, appearing at least as far back as Adam Smith's Theory of Moral Sentiments.Footnote ¹⁶

The basic motivation for it is rooted in recognition of the role that character traits play in decision and action. Traits are not to be understood simply as dispositions towards particular kinds of action, but rather as manifested in the way the agent perceives their situations, thinks about them, and feels about them. Character, that is to say, structures the agent's environment into a field of reasons. In unreflective experience, we are engaged in a world of invitations, demands, proscriptions, and opportunities, where this practical structure mirrors our own character traits. Although character structures experience in this way, it does not explicitly figure in that experience itself. This is why the honest person should not be understood as someone who sees the world as a range of opportunities for being an honest person, but rather as someone who sees the world through the lens of honesty. Thus, we do not have direct epistemic access to our character traits in unreflective experience. All that we find before us is the world of reasons.

When we reflect on this experience, the argument runs, we are again simply confronted with the world of reasons. For the experience itself is ‘transparent’. That is to say, it is because unreflective experience is nothing but a presentation of the world that reflection on that experience can deliver nothing more than the world presented. But we can take up a different kind of reflective stance. Rather than reflect on our subjective experiences, we can reflect on the patterns in our behaviour. In this kind of reflective experience, we draw inferences about our character under the supposition that our behaviour is rationally guided. We come to understand, that is to say, the patterns in the reasons for our behaviour, patterns that manifest our dispositions. This kind of reflection rests on publicly available information. We do not have any special kind of epistemic access to this information, although we do have the advantage of witnessing more of our own behaviour than any other person witnesses. This is, therefore, an ‘external’ point of view on oneself, since it is based on information available to anyone. From the ‘internal’ point of view, from the point of view available only to oneself, all that one can see is the world.Footnote ¹⁷

This role of character as imposing a practical structure on the world of experience explains why the prediction of someone's behaviour on the basis of their character is a prediction of the decision that they will make. For the understanding of their character on which the prediction is based is an understanding of how the situation will seem to them. It is an understanding of the reasons that they will find in the situation. On the assumption of practical rationality, this understanding of their character licenses a prediction of their response to these reasons. This is also why it is a distortion of character to see it just as a set of dispositions towards particular behavioural responses to particular situational stimuli. Although character does dispose towards action, the operation of this disposition cannot be mechanistic. Because character is manifested in the presentation of reasons, any behavioural outcome requires practical reasoning. This point is at the heart of the charge of doublethink.

Moreover, it is the general acceptance of this practical role of character among virtue ethicists that explains why there are only two basic metaethical theories of the normative status of virtue. For if character is manifested in perception, thought, and feeling, as well as in action, then one might wonder why there are only theories that assign normative priority to virtue and theories that assign it to action. Why, one might wonder, is there not a third position that assigns normative priority to experiencing the world in the right way? Could one not hold good action to be action resulting from the world being perceived as the right set of reasons, with the right emotional texture and the right deliberative considerations, and likewise hold virtues to be those traits that dispose towards this experience? Such a position would need to assign a value to the practical structure of experience without thereby assigning it to character. But this cannot be done, for to experience the world as having a particular practical structure just is to possess a certain character trait. To assign normative priority to the right way of experiencing the world is to assign it to virtue, since that is what virtue is.Footnote ¹⁸

Although this account of the rational role of character motivates the claim that one has epistemic access to one's own character only from the external point of view, it does not entail it. Indeed, as we will see, this widely accepted view of the nature of character in fact provides the basis for a different account of epistemic access to one's own character. Once that account is in place, we will see that the dilemma for virtue cultivation is mistaken, as are the objections it encompasses. One can cultivate virtue as an end in itself without being narcissistic and one can cultivate virtue as a means to good action without engaging in doublethink. But before we see why that is the case, we will consider further our epistemic access to our own character.

5. From a Phenomenological Point of View

If character is manifested in the practical structure of experience, then reflection on that practical structure affords epistemic access to one's character. This does not require the denial of the transparency of unreflective experience. We should accept that in unreflective experience the agent is faced with a practically ordered world and nothing more. We should also accept that reflection on transparent experience cannot deliver anything other than that practically ordered world. But it does not follow that this reflection cannot deliver any information that was not delivered in unreflective experience. For if it can deliver the object of the unreflective experience in a different perspective, or in a different light, then it is possible that we can learn something about that object that cannot be learned from the initial unreflective experience. What is more, this new information would not be available from the external point of view. It would be uniquely first-personal, because one has no direct epistemic access to the practical structure of someone else's experience of the world.

In developing this view of reflection on the practical structure of one's experience, we will draw on Jean-Paul Sartre's phenomenology of consciousness. There is some irony in this, since some of the arguments subsumed into the dilemma for virtue cultivation also draw on Sartre. Whether this is due ultimately to an inconsistency in Sartre's philosophy or merely to some infelicity in its expression is a question we will leave open here.Footnote ¹⁹ Sartre certainly does sometimes sound as though he holds that epistemic access to one's own character is restricted to reflection from the external point of view. ‘Consciousness does not know its own character – unless in determining itself reflectively from the standpoint of another's point of view’, he writes.Footnote ²⁰ But his account of the nature of character and his theory of reflection, which grounds his method of phenomenology, together suggest epistemic access to one's own character that does not rely on the external point of view.

Sartre holds that character is responsible for the structure of the objects of experience as a world of invitations, demands, proscriptions, and opportunities, even though one's character itself does not appear in that experience. Or, as he puts it at one point, consciousness ‘exists its character in pure indistinction non-thematically and non-thetically […] in the nihilation by which it recognizes and surpasses its facticity’.Footnote ²¹ In reflection, this unreflective experience of the world becomes the object of a further mental state. This reflective mental state is not directed towards the objects of the unreflective experience, but towards the experiencing of those objects. It presents the way in which the world is structured in that unreflective experience as itself an object for consideration. Unreflective experience is absorbed in the practical structure of the world, in responding to the possibilities it presents. It is because reflection on that experience is not likewise absorbed in the practical structure of the world, but rather ‘the reflecting consciousness posits the consciousness reflected-on as its object’, that reflection allows me to ‘pass judgment on the consciousness reflected-on’.Footnote ²²

Sartre does not give a clear exposition of his method of phenomenology in Being and Nothingness. But in earlier philosophical works, he is very clear that the aim is to describe the way objects appear in various types of experience by reflecting on examples of those types of experience. He begins The Imaginary, for example, by distinguishing between imagining something and reflecting on that imagining. He argues that the reflection always reveals the original imagining as an imagining, that it is impossible to mistake it in reflection for a perception or some other experience. This is because imaginings ‘present themselves to reflection with certain marks, certain characteristics’. It is this that grounds his method: ‘produce images in ourselves, reflect on these images, describe them, which is to say, try to determine and classify their distinctive characteristics’.Footnote ²³ He goes on to argue that this method reveals that imagination is distinguished from perception by characteristics of the way the object of experience is presented in the experience.

If reflection can reveal the way the object of that experience was presented, then given that character is manifested in the way objects of experience are presented, reflection ought to reveal features of experience that manifest character. Such reflection is not from the internal point of view of the original experience, since it is not the same mental event as the experience reflected on. Where the object of the original experience is the world, the object of this reflection is the practical structure that the original experience presents the world as having. Neither is this reflection from an external point view available to any observer, since one can only reflect in this way on one's own experiences. Your access to the way the world seems to me is secondary to my own access to that information, since your access relies on my reports of, or other reactions to, the way the world seems to me. This kind of reflection is from neither the internal point of view nor the external point of view, therefore, as these have been understood in the debate over cultivating one's own virtue. We should keep this clear by giving this reflective perspective a third label. Since its object is the way the world appears, I suggest we call it the phenomenological point of view.

This phenomenological point of view is a form of privileged access, since one can take up this perspective only on one's own experiences. But this does not entail either of two further claims that have often been made for self-knowledge under the title ‘privileged access’. It does not follow that this epistemic access to one's own character is infallible. Neither does it follow that the full detail of one's character is available from the phenomenological point of view. Sartre is well aware of this. He distinguishes ‘the certain’ features of experience available to phenomenological reflection from ‘the probable’ conclusions that can be inferred about the underlying causes of those features.Footnote ²⁴ Moreover, he argues that a sufficiently deep motivation to see oneself in a particular light might distort one's reflection on the way the world appears to be.Footnote ²⁵ Nevertheless, the availability of this phenomenological point of view is overlooked by the dilemma for virtue cultivation, which presupposes that epistemic access to one's own character requires the external point of view. We will see that the objections encompassed by that dilemma cannot be raised against phenomenological virtue cultivation. But first we must give further substance to the idea that one can cultivate virtue in this way.

6. Practical Experience and Reflective Endorsement

Resistance to the idea of virtue cultivation from a phenomenological point of view might be rooted in either of two objections. One would be that there is no genuine distinction between unreflective experience of the world with a particular practical structure and this purported reflective experience of the practical structure of the experienced world. If this is no more than a difference in description that fails to be matched by a difference in experience, then there is no genuine phenomenological point of view. Rather, all we would find when reflectively considering an experience is the practically structured world exactly as unreflectively experienced. Certainly, we have not yet seen reason enough to insist that the phenomenological point of view is genuinely distinct from unreflective experience in the information it delivers. So to meet this objection, more needs to be said about the nature of this difference.

The second objection might concede the possibility of a difference in content between unreflective experience and phenomenological reflection, but deny that genuine self-criticism could be rooted in such reflection. The idea here would be that the reflective point of view remains one occupied by the same subject who is purportedly under criticism. If the critic shares precisely the same character with the criticised, and if character indeed bestows the practical structure of experience, then it would seem that the critic would lack the required critical distance to be able to critique their subject effectively. To put this point another way, if one is assessing one's practical commitments from the perspective of those very practical commitments, then it seems that one is destined to affirm the commitments that one is supposedly critiquing. Perhaps one needs to take up an external point of view, either the point of view of some particular real person, or that of some imagined other person, or the abstract point of view of one's society as a whole or of an ideal spectator, in order to introduce the intellectual distance required for self-criticism.Footnote ²⁶

Neither of these objections is sufficient to undermine the idea of self-criticism based on phenomenological reflection. Central to this idea is the distinction between those values that one consciously endorses and those that are sufficiently embedded in one's cognitive (and affective) system to contribute automatically to the constitution of the practical structure of experience. For not all of our consciously endorsed values are sufficiently embedded to operate in this way. Conversely, not all of those values that do structure our experience are ones that we are even consciously aware of holding, never mind ones that we would endorse were we to become aware of them. Then there is the third category, the set of values that we do consciously endorse and that are sufficiently embedded in our cognitive architecture to contribute to the practical structure of our unreflective experience. The aim of cultivating virtue is the aim of enlarging this third category, with the ideal goal of holding only values that are both consciously endorsed and automatically activated in unreflective experience. This is the practice that virtue ethicists from Aristotle onwards have described as rationally guided habituation.Footnote ²⁷

It is important that virtue cultivation is reflective deliberation in this sense, rather than the philosophical consideration of ethics that Williams, in Ethics and the Limits of Philosophy, is concerned cannot be conducted independently of the commitments that already structure the character of the person reflecting, or the societal values that character embodies, without thereby undermining commitment to the importance of ethical value.Footnote ²⁸ The reflective deliberation involved in virtue cultivation, by contrast, is concerned with whether the agent's own dispositions genuinely embody the values that agent would consciously endorse. It is a self-critique that takes conscious commitments largely for granted, though it might lead to some consideration of their overall coherence, and aims to bring the practical structure of the agent's experience into line with those commitments. In other words, this reflective self-critique is involved in habituating in oneself the character traits required to live up to the values that one endorses.

In unreflective experience, the values that are deeply embedded in our cognitive architecture provide the practical structure of the world that we experience. The world is thus experienced as a set of demands, invitations, proscriptions, and opportunities. Because this practical structure is bestowed automatically, the world is experienced unreflectively as exerting directive pressure. Reflection on this experience takes the experience of this directive pressure as its object. This reflection itself has a practical structure, but this embodies one's deeply held values as they apply to the object of this reflective experience, not to the object of the unreflective experience. That is to say, in this reflection the practical structure is applied to the unreflective experiencing of the world as practically structured, not to that world itself. It is this that grounds the distinction between unreflective experience and phenomenological reflection on that experience. In phenomenological reflection, one's values are applied not to the object of the unreflective experience, but to the practical pressure exerted by the object of unreflective experience. One considers the way that the authority figure's instructions seemed to demand compliance, or that the stranger's suffering seemed like an obstacle to one's action, or that the driver in front seemed to be thwarting one's goals, rather than now feeling that demand, difficulty, or frustration.

Reflective experience is thus disengaged from the directive pressure exerted by the practical structure of the unreflective experience. From the phenomenological point of view, we observe rather than feel that directive pressure. This opens up the critical distance required for virtue cultivation. For in the cool light of phenomenological reflection, one can consider whether that directive pressure accords with one's consciously endorsed values. One can pass judgment on the practical structure of the unreflective experience. In so doing, one is passing judgment on the set of deeply held values that provide that practical structure. One might, for example, reflect that it really had seemed appropriate to obey the authority figure's instructions, or to ignore the suffering of the stranger, or to express annoyance at the driver of the car in front, and that this does not now, on reflection, seem appropriate at all. One might be dismayed, that is to say, not only by one's actions, but also by the ways in which one experiences the world and the influence this has over one's actions. The aim of cultivating virtue should be understood as the aim of getting that practical structure of experience right.

7. The Dilemma Dissolved

The dilemma for virtue cultivation is based on the assumption that epistemic access to one's own character is restricted to reflection from an external point of view. In this kind of reflection, one considers the patterns in one's behaviour and draws conclusions about one's underlying dispositions. This access to one's character is available to anyone who can witness one's behaviour, although it is true that one witnesses more of one's own behaviour than anyone else does. Given this epistemic access to character, the aim of cultivating virtue would be the aim of reflectively critiquing and aiming to improve the dispositions that underlie one's behaviour. To aim at virtue as the ultimate end of this activity is narcissistic, since it sets oneself as the ultimate terminus of one's ethical attention. To undertake this activity as a means to better behaviour seems committed to doublethink: it treats action as merely caused by character dispositions, since otherwise one should just commit to respecting the right reasons in action; but it also rests on the assumption that practical reasoning about one's character can lead to actions aimed at cultivating virtue.

Neither the charge of narcissism nor the charge of doublethink can be brought against virtue cultivation on the basis of phenomenological reflection on the practical structure of experience. One's aim in this activity is to come to experience the world as it ought to be experienced. What one is cultivating is the tendency to recognise in the world the reasons that one ought to recognise there. If one aims at this for its own sake, one can hardly be charged with narcissism. For the terminus of one's ethical gaze is not oneself, but the reasons that one finds in the world. One is simply aiming to understand the world as it should be understood. Moreover, if one is aiming at this in order to behave better, then one is not overlooking the role of practical reasoning in action. For one is not aiming to instil dispositions that cause behaviour, but rather to recognise the reasons which one ought to recognise in practical deliberation.

The charge of doublethink can be cast as the objection that if one genuinely wants to behave better, one should not focus on behavioural dispositions but commit to respecting the right reasons in action. Phenomenological virtue cultivation just is commitment to respecting the right reasons. Or, to put it another way, rational commitment requires more than intellectual endorsement, since it also requires habituation to embed the endorsed values in one's cognitive architecture sufficiently to be manifested in the practical structure of unreflective experience. Such rational commitment requires the reflective critique of the practical structure of one's experience that is available from the phenomenological point of view.

Although phenomenological reflection is intrinsically first-personal, since one has direct epistemic access to the practical structure of one's own experience but not to the practical structure of anyone else's experience, it does not follow that phenomenological virtue cultivation is inherently solipsistic. The advice of other people, particularly when this is grounded in systematic external study of the origins of behaviour, will also be required. For what is available to phenomenological reflection is the practical structure of experience itself, not the underlying causes of that practical structure. One might be dissatisfied with this practical structure without being in a position to formulate successful strategies for altering it. This is for two reasons. First, it might be that one cannot discern the precise nature of the aspect of the practical structure of one's experience that one wishes to alter. Second, it would seem to be an empirical rather than a phenomenological question how best to alter the patterns in the practical structures of one's experience.

Empirical research into psychological priming supports the first of these points. For example, one recent experiment found that drivers of red cars are subject to more aggression from other drivers than are drivers of blue, green, black, or white cars. The experiment involved the experimenter's car waiting at traffic lights after the signal had turned green and recording whether the car behind responded aggressively and, if so, how much time elapsed between the signal change and the response. They found that significantly more drivers responded aggressively to the red car blocking them in this way than they did to blue, green, black, or white cars, and that aggressive responses to the red car were generally quicker than to the other cars.Footnote ²⁹ Some of the drivers who responded aggressively to the red car, therefore, would not have done so had the car not been red. It seems likely that this would come as a surprise to those drivers. They are aware of their aggressive response, that is to say, and aware that they saw the driver in front of them as behaving in a way that seemed to call for such a response. But they are likely to be unaware that the driver would not have seemed to be calling for such a response had the car not been red. They are very unlikely to mention the redness of the car as partly explaining their response. This may seem a trivial example, but other experiments have found behaviour to be similarly biased according to more important factors, such as the ethnicity of the person responded to.Footnote ³⁰

It is likewise an empirical question how one should best go about altering unwanted aspects of the practical structure of one's experience. One method would be to try to eliminate the disposition that it manifests. In the case of the red car, one might aim to eliminate one's aggression generally, or one might try to eliminate the disposition to experience redness as calling for aggression. Although it is a normative question which of these strategies is preferable, empirical information about the role and value of aggression should inform answers to that question. It might be, for example, that the association of redness with aggression is so deeply embedded in our cognitive architecture, it having been an important association throughout our evolutionary past, that one is less likely to be able to eliminate that association than to reduce one's aggressive disposition generally. Moreover, it is an empirical question whether one should even aim to eliminate one's undesirable dispositions at all, rather than to embed strong countervailing dispositions that will effectively cancel out the undesirable ones.

The project of virtue cultivation through critical reflection on the practical structure of one's experience, then, has a strong empirical dimension, one that should be the subject of further research in philosophical moral psychology. But it remains that virtue cultivation is best understood as a phenomenological enterprise. For this kind of virtue cultivation does not face the dilemma of narcissism and doublethink. That dilemma encapsulates a range of objections that have been raised against virtue cultivation on the assumption that epistemic access to one's own character is restricted to the external point of view. But the phenomnenological point of view offers a different kind of access to one's own character, which grounds a form of virtue cultivation that is not subject to the objections encapuslated in the dilemma. We should not accept, therefore, the conclusion of the dilemma, which is that one's own virtue can only be a self-effacing goal if it is to be a goal at all. Moreover, since character provides the practical structure of experience, it is difficult to see how the aim of acting on the right reasons could be better served than through the cultivation of the right character traits. We should aim to cultivate good character, therefore, not through critical reflection on our own behaviour from the external point of view, but rather through the privileged access of phenomenology.Footnote ³¹