It is uncontroversial that virtues and reasons are connected. But moral theorists differ widely regarding just what the connections are, and so far there has not been a fully adequate response to the question whether, in some important way, the category of reasons is more basic than that of virtues. This paper pursues that question. It begins with developmental considerations concerning what constitutes role modeling of the kind that best contributes to virtue. In this context, it explores what is required for a proper “aretaic” responsiveness to such role modeling. Reasons for action are shown to play an essential role in such responsiveness. The paper also shows that we cannot adequately understand virtues apart from appeal to the kinds of normatively significant reasons that must figure in their characteristic exercises. The paper connects this conclusion with epistemological points, such as those indicating how one can know that a person has a virtue, with ontological points concerning what psychological properties ground virtues, and with developmental points indicating how virtue must apparently be role modeled if it is to be (demonstratively) well taught. None of the emerging conclusions implies that virtues are of secondary importance in ethics, nor even that, apart from virtue, human beings can be expected to act morally with reliable steadfastness. Whether or not virtues are fundamental in one or another philosophically interesting way, they appear to be psychologically real traits of persons and both morally and intellectually indispensable.
Role modeling by someone virtuous is widely considered crucial for the development of virtue in others, especially children. But modeling the overt actions that go with virtuous conduct is not in general enough to elicit that development – or, at least, is not enough unless it results in successful learners not only doing or being suitably disposed to do the right deeds, but also in their doing or being disposed to do them for appropriate reasons. Moreover, doing the right thing for the right reason just once need not manifest virtue. It is even possible that some appropriate motivating reasons ground deeds of a given kind – say, helping others – repeatedly yet without the elements in character needed for doing the deeds in a way that bespeaks genuine virtue. Reasons of self-interest may masquerade as reasons of beneficence when doing good deeds for the powerful is a way to self-advancement. This is a pretense of virtue, and possibly habitual in some people, not the thing itself.
Genuine virtues, by contrast, are traits of character. What constitutes character and how it may be developed can be clarified in a distinctive way by considering the nature of role modeling and how it affects character. The dimensions of the role modeling important for developing virtue are complex and a challenge for both philosophical theory and developmental psychology. The philosophical challenge is to provide an adequate conception of role modeling and to formulate normative standards for doing it well; the psychological challenge is in providing descriptions, causal explanations, and empirical connections with (among other things) motivation and cognition. This paper chiefly concerns the philosophical challenge, though it is written with a sense of psychological realism and so makes minimal presuppositions about any empirical data available only through systematic psychological inquiry.
Our discussion will illustrate the centrality of reasons in the development, exercise, and understanding of virtue and, in that light, will pursue the question whether reasons for action are in some way more basic (in a sense to be clarified) than virtues, conceived as traits of character. One way to approach this priority question, as it might be called, and indeed to approach the philosophical understanding of virtue, is to explore how virtue is normally learned from role modeling, from instruction by precept, from verbal guidance of other kinds, and from those combined.
I begin with role modeling and will contend that neither this nor the responses to it characteristic of the acquisition of virtue can be understood apart from seeing how reasons figure in explicating both the role modeling of virtue and the kind of learning it properly aims at. This is a priority view: understanding role modeling (asymmetrically) depends on understanding reasons. The paper next proceeds to develop a three-dimensional conception of role modeling and the appropriate response to it by its potential beneficiaries. These notions, and through them virtue itself, are then further clarified by indicating some ways in which emotion is related to virtue. The concluding section shows how and why reasons – motivational as well as normative – are basic elements that must be taken fully into account by any adequate theory of virtue.
1 Aretaic Role Modeling
Role modeling, as understood here, is roughly doing something appropriate to a role, in circumstances to which it is appropriate, and perceptibly to someone who is a candidate for acquisition or reinforcement of the role in question. Teachers may role model passage analysis; farmers may role model the care of livestock; and – consciously or not – athletes may role model responses to winning and losing. When role modeling is of the kind important for the development of moral virtue, it requires action, but not just any kind of overt behavior of the sort characteristic of an agent exercising the virtue in question suffices for role modeling. The wrong kind of motivation may be readily detectable.
To be sure, not all virtues can be conceived in terms of characteristic roles, but there may still be a sense in which anyone acting virtuously is describable in terms of playing some role. Thus, if you have the virtue of justice and, from a sense of fairness, distribute responsibilities among disagreeing colleagues, you are acting in the role of just person (or judge, or perhaps both); and if, with practical wisdom, you give judicious advice to a struggling student or a distraught coworker, you are acting in the role of advisor – or benefactor, or senior colleague, or friend – even if none of these represents a role you usually play. Furthermore, a single segment of conduct can model two or more virtues, and indeed more than just virtue. Some of the theatrical associations of acting in a role in a play can be misleading here, particularly the suggestion that one is not being oneself. Where virtue is role modeled, moreover, the relevant actions may be performed without rehearsal or self-consciousness, even when they are exemplary. They may also be done with difficulty or with labored attention to the governing standards. In either case, they may admit of imitation and may teach something to an observer.
Role modeling of a moral virtue must meet certain objective standards: to a properly sensitive observer it must give the impression that the relevant deeds, those partly constitutive of that virtue, are motivated in a certain way. It need not evoke a belief attributing specific motivation; but under appropriate conditions it should create a sense of the agent’s being motivated in a certain way, say for reasons of beneficence. That sense, in turn, normally disposes the observer to form one or more beliefs appropriate to describing the motivation (though it does not require formation of such beliefs). Consider a case in which we have only the behavioral bare bones of aretaic role modeling: binding someone’s wound after a fall from a skateboard, done with no expression of sympathy or caring about the victim. Done in that way, this deed is not role modeling of the virtues of kindness or beneficence. A duly sensitive observer who cannot perceive signs of caring or some positive attitude may learn technique from it, but, as this example shows, virtues are not acquired just by learning the instrumental skills needed for their behavioral expression or even for accomplishing an aim appropriate to them.
Other instructive examples are role modeling the virtues of considerateness and generosity. Imagine a guest arriving unexpectedly at a family’s dinner hour. There may not be enough food prepared, but suppose the parents quickly set another place and, in view of the children, aged nine and eleven, serve themselves and the children less than usual. If they do it with alacrity and take pleasure in the guest’s appreciation of the good flavor and happy inclusion in the meal, they need not say anything to the children to have gotten across that on occasions of a sudden need that we can fulfill for people around us, doing so is a good thing. They show consideration for the feelings of the guests and show generosity in talking less in order to serve others. Granted, saying that we should share with our guests – or friends or neighbors – might be desirable here, but saying such things does not suffice for role modeling and is likely less effective than doing, in the right way, the things described.
We often think of moral education in relation to guiding children by both example and precept.1 Does good role modeling require precepts as elements, or is it more a matter of appropriate action of a non-didactic kind? On my view, role modeling is not merely behavioral; but, even when verbal, as with teaching the oral reading of poetry, it does not always require offering any commentary on what one is doing – valuable though that may be. Still, it would be at best abnormal for someone role modeling kindness not to be disposed, even if only afterwards, to attribute the relevant modeling deeds to some appropriate motive and, more generally, to criticize certain failures. One kind of failure is in not performing such deeds in circumstances that call for them; another kind is doing so for a reason of the wrong kind. This disposition seems especially likely to be manifested where the role model is teaching a child how to do things or is even properly conscious of being, if not an exemplar, at least an object of possible imitation. I take that consciousness to be characteristically manifested in acting on appropriate motivation. Recall the case of self-interest masquerading as beneficence. Even a child may, in time, detect insincerity in such cases. Not all motivated right action is for a reason appropriate to virtue, but all aretaically grounded action, the kind role modeling should achieve, is appropriately motivated.
If one reads Aristotle’s Nicomachean Ethics without keeping these points in mind, one might think that when he says (without qualification in the context) that virtue arises by habituation (e.g. in 1103a15ff and 1104a21-25), he is taking overt behavior to suffice for role modeling aimed at moral education and thereby for the development of virtue in learners, such as children, who may of course show their virtue by the right kinds of habitual acts. But this understanding would not take account of much else he says, particularly concerning what it is for an action to express virtue.2 In many passages, it seems clear that he has in mind the necessity of appropriate reasons as essential for action from virtue. If so, it would be strange indeed if he did not think that moral education, in role modeling and in other ways, should take account of the need for action from virtue to be motivated by appropriate reasons. This implies that moral education by role modeling should both communicate the appropriate underlying reason(s) and criticize doing the relevant deeds for reasons of the wrong kind.
I have described role modeling as partly constituted by overt behavior.3 This implies that role modeling is not possible simply by stating the relevant standards of action – say, by presenting them in a peptalk. But it does not follow that good role modeling is in general possible without such expressions of standards. For certain virtues, it normally is not. Consider courtesy, for instance. Even if there is a generic notion of courtesy that the virtue of courteousness must reflect, the appropriate behavior differs enough from one culture or subculture to another so that some descriptive explanation is likely needed, at least where there is not time and occasion for the full range of required behaviors to be demonstrated. On the assumption that intellectual virtues are in some ways parallel to moral virtues and in any case a good foil for them, note that for the intellectual virtues of clarity, rigor in reasoning, and judgment, the same point appears to hold. Indeed, intellectual virtues are so closely bound up with conceptualization of a kind that normally requires instruction that role modeling them usually requires commentary and sometimes calls for critique.
Role modeling, then, requires acting appropriately to the role in question. Good role modeling requires doing such things for the right kind of reason. Role modeling of virtues requires acting for a reason that is appropriate to some role characteristic of the virtue in the relevant circumstances, as where beneficence is modeled by a nurse who treats a wound not from a desire to gain favor but from concern for the patient’s health. This is not to imply that for every virtue there is a single specific role. The points made here apply to the multiple roles that go with exercising certain wide-scope virtues, such as beneficence, justice, fidelity and courage.
2 Aretaic Uptake
There are various ways in which philosophical inquiry can clarify virtue. One is direct analysis, but the approach here is different. It proceeds by clarifying what it is to role model virtue, how role modeling is connected with reasons for action, and what kind of response to role modeling is of the right sort to express virtue and apparently favorable to the development or reinforcement of virtue. As some of our examples indicate, role modeling does not imply being didactic or even intending to do anything approaching role modeling. The point is that the exercise of the virtue in question is done in an imitable way that one might hope is possible in child-rearing and moral education. But aretaic role modeling with no educative, reinforcing, or other distinctive effect is rather like a fine painting that no one appreciatively views. A teacher who does everything well but has students who do not open-mindedly pay attention will not succeed in educating them. Our understanding of virtue, especially from a developmental point of view, is incomplete without some account of what I am calling aretaic uptake in response to role modeling. Such uptake may be viewed as significantly analogous to appreciative viewing in art, particularly for those “learning” one or another of the virtuous traits in question. Merely seeing color and shape is not sufficient for appreciating paintings; one needs a kind of receptivity to perception of structure, design, and other elements that yield worthwhile results. Similarly, aretaic uptake requires a sense of the circumstances that call for manifestation of the virtue, a grasp of the appropriateness of certain responses to those circumstances, and a disposition to make them in those circumstances for an appropriate reason. This sense need not be articulable, but its scope and depth is intrinsically connected with six dimensions of aretaic analysis. When uptake is optimal, the agent’s instantiation of the virtue will reflect each of them. It will help us in reflecting on the nature and normal learning of virtue to consider them briefly here.
To understand virtue – at least the practical kind that includes above all moral virtue – we must take account of at least six variables: (1) the field of a virtue, roughly the kind of human situation in which it characteristically operates, such as, for justice, determining sentences for crimes and, for fidelity, the range of behaviors to which the agent is committed; (2) the characteristic target(s) the virtue leads the agent to aim at, such as, for beneficence, the well-being of others; (3) the beneficiaries of the virtue, above all (and perhaps solely) the person(s) who properly benefit from realizing it, such as those needing help or counting on promise-keeping (this does not rule out agents themselves as major beneficiaries of virtue, as with self-discipline); (4) the agent’s understanding of the field, for instance of criteria for equal treatment of persons; (5) the agent’s motivation to act appropriately in that field in a certain way – the way that befits the virtue, say (in the case of fidelity) motivational reasons of moral commitment to promises and to relationships; and (6) the psychological grounding of the action tendencies characteristic of the virtue, where this requires that the acts manifesting it be (motivationally) explained by – and in that sense grounded on – an appropriate reason and not just rationalizable by appeal to it.4 These dimensions of analysis explain why habituation by itself – conceived as “automatically” doing a given kind of thing in certain circumstances – is not enough for development of virtue. They also indicate the requirements for the psychological profile of an agent who has developed a virtue, as where good parental role modeling achieves sound moral education.
In the light of these six considerations, we can approach the question whether the kind of behavior that can become habitual and can superficially “model” virtue is even necessary for role modeling virtuous action (we have seen why it is not sufficient). That it is necessary may be mistakenly inferred from the idea (easily attributed to Aristotle) that virtues arise by habituation. Granted, without certain habits one perhaps cannot have the traits of courtesy, veracity, or kindness, but even in these cases not every behavioral expression of the virtue need be an instance of habitual action. And what of the virtue of good judgment, which may double as a moral and intellectual virtue? That commonly requires reflection on alternatives, but a person of good judgment need not be in the habit of reflecting on alternatives whenever they are encountered and, in any case, the final judgment (whether intellectual or moral) that emerges as an expression of the virtue is not an action that, like slowing a car before a curve, is habitual.
Consider, too, generosity to favored causes. Here we might expect the donational motivation to arise spontaneously in certain situations; but no automatic giving can be expected in the common situations that present an agent with multiple deserving recipients. The agent may simply lack the resources to give at all or may agonize about which cause to support and then delay indefinitely. In any of these cases, moreover, principled exceptions, for both the rich and the poor, may be numerous.5
Exemplifications of virtue in action need not, then, be behavioral manifestations of habit. They also need not be perfect representations of the virtue in question. A courteous response to an unwanted question may have a rough edge; a kindness may fail to benefit the person needing it. In a virtuous person, however, one expects the aretaically constitutive dispositions to be self-corrective, a point stressed by Julia Annas, among others.6 Imperfection is one thing; unresponsiveness to perceptible deficiency is another. The former is compatible with virtue; the latter is generally not. As these points suggest, virtues are possessed to varying degrees, and there is no precise quantitative measure of a threshold for possession, either in strength of underlying motivational or cognitive elements or in frequency or strength of behavioral manifestations.
3 Virtuous Action and Motivating Reasons
It should now be clear that virtue is multi-dimensional in constitution, that, correspondingly, habituation, even with the right behavioral results, is not sufficient for its development, and that manifestations of virtue in action need not be performed from habit. But more should be said about how virtuous action may be conceived in the light of the six dimensions of aretaic analysis described in Section 2.
That actions expressing virtue, and so not merely characteristic of it – or “in conformity with it,” in Kantian language – must be performed for an appropriate reason is widely recognized. But this leaves open how the agent must understand the required reasons. They need not be conceived as reasons, though this may be common for sophisticated agents. Still, even a sophisticated virtuous agent may simply see certain deeds as, in the context, to be done. This may require seeing some act or desirable goal as indicating a need, as fitting one’s commitments, as benefiting another person, and in many other ways. But neither the notion of a reason nor any normative concept need figure in consciousness.7
The point that actions from virtue must be performed for an appropriate reason will tend to mislead if we do not distinguish between actions from virtue and actions that sometimes manifest it in readily recognizable ways but need not manifest it and may indeed have a basis other than virtue. Take sensitivity to the feelings of others. This will manifest itself in a manner of treatment that may be exemplified by many actions that are non-intentional or at least not aimed at some goal that, in the context, figures in the sensitivity. One might furrow the brow on hearing something, draw closer to someone to allow a lower tone in communicating something personal, or simply avoid asking certain questions natural in the situation. Some of these actions or manners of action need not be intentional, even though none is unintentional; and not all such actions derive from virtue, as opposed to variables (including momentary desires) that do not imply a trait of character. They may derive from virtue, but that will not always be apparent in the context of action. One important point here is that the manner of action – its adverbial dimension as opposed to its act-type – should not be overlooked in understanding virtue. The way in which we do things, e.g. deftly, vigorously, or cheerfully – even things not at the time required by any virtue – may indicate a particular virtue or at least constitute modes of behavior appropriate to some virtue. Possession of the virtue, for instance, may best explain the manner of action in question.
To summarize some of what we have now seen, both role modeling and aretaic uptake have three major dimensions. In full-blooded cases, they have what might be called a three-m structure: one dimension is act-type, the “matter” of action; the second is motivation; and the third is the manner in which the relevant acts are performed. This structure is to be expected given the dimensions of aretaic analysis sketched in Section 2 and will be evident in the relation of virtues to reasons to be described shortly. Virtuous action is (relative to its circumstances) of the right kind, motivated by the right kind of reason(s), and performed in an appropriate manner. The right kind of act is determined by the particular virtue in question and what it calls for in the circumstances; the right kind of reason is determined by the target(s) of the virtue and what, in the circumstances, will or may be reasonably thought to hit it; and the appropriate manner of performance is a holistic matter in which the agent must respond to at least the target(s), the circumstances, and the needs or capacities of anyone the action concerns.
4 Virtue and Emotion: A Nexus of Intimate Relations
Emotion is a pervasive element in human life. It would be impossible for virtues of character to play the major part in the good life that they are supposed to play if they did not have an appropriate relation to emotion. This idea is prominent in Aristotle, who (in 1104b and elsewhere) maintained that “virtues are concerned with actions and feelings” and devoted considerable attention to how virtue is connected with emotions, such as fear in relation to courage. Virtue is still often associated with the control of emotions, especially where such common emotions as anger, fear, jealousy, and resentment lead to wrongdoing. But emotions can also be appropriate to virtue and positively influence it, as where empathy guides beneficence or indignation energizes justice. Moreover, some virtues, including compassionateness, courage, and lovingness, require emotion in their very constitution, though even in these cases the actions that manifest the virtues must be performed for appropriate reasons. This requirement, as we shall see, is consistent with the point that some virtues also apparently require, both for their acquisition and their possession, not only appropriate motivation but also emotions.8
To see the aretaic importance of emotions, consider compassionateness first. Normally, unless one has feelings significantly responsive to the plight of others, one cannot develop compassionateness. Such feelings, then, would be an element in genuinely aretaic uptake to role modeling of compassionateness. These, like emotions such as love and delight on the positive side and, on the negative side, resentment and anger, are developmentally important in relation to virtue – which demands the right aversions as well as the right attractions and a good balance between the two. Similarly, someone in whom no affectionate feelings toward anyone else occur cannot become affectionate; and normally, if one has no fear to be overcome, one cannot develop courage. In any of these cases, if brain manipulation is capable of inducing virtue, it must produce not only the right behavioral patterns and its rootedness in appropriate reasons, but also emotional tendencies.9
Here courage, which is a paradigm of a virtue whose target includes control of emotions, differs from the others just mentioned in a way that merits comment: once courage is achieved, the courageous agent need not feel fear but can face danger without it – the crucial thing is a steadfast willingness to risk something one values for the sake of something else. If fear is normally required for development of courage – since mastering it is typically crucial for that – too much of it can make acquisition more difficult. A person too easily frightened may not be able to develop courage. There are, however, some emotions that may facilitate developing virtues of any kind. Such, for Linda Zagzebski, is (occurrent) admiration, provided it is directed toward the right elements in someone else’s behavior and character.10 This idea deserves discussion (more indeed than is possible here). Admiration usually implies desire to be like the person admired in respect of the element(s) for which the person is admired. It typically implies a kind of approbative pleasure in witnessing manifestations of the admired trait or behavior; and such admiration, in turn, tends to make more likely the development of both emulative motivation and sensitivity to the reasons for action that go with acting in expression of that trait.
It is one thing, however, to take admiration to tend to produce desire for similarity; it is quite another thing to give it a major epistemic role, as Zagzebski often does, or to consider it a necessary condition for aretaic uptake.11 In any case, even if admiration is an appropriate and common response to a person’s perception of virtue in one or more other people, it seems possible to recognize a virtue in someone without admiring it. This may indeed be illustrated by Iago in Shakespeare’s Othello. Iago scorns certain virtues in others or perhaps is jealous because of them, but his discernment neither comes from nor constantly reveals admiration. If Othello role modeled honesty for his troops, Iago is a paradigm of understanding without aretaic uptake.
It also appears that admiration can be misplaced in such a way as to yield something perhaps liked but not virtuous. Related to this, is there any evidence that virtues cannot be understood apart from determining what is admired (as opposed to meriting a certain kind of approval) or (though this is not implied by the view in question) that virtues cannot be acquired without the help of admiration? To be sure, even if the view that admiration is central for understanding and development of virtues is mistaken, it remains true that admiration and other emotions can play a major developmental role regarding moral virtue and can help in distinguishing virtues from merely socially approved traits.
What has emerged in this section provides good reason to integrate a theory of emotion into our understanding both of what constitutes virtue and of how it is role modeled and acquired. Emotions can be constituents in virtue, guides in its expression, sources of motivation to act virtuously, and, like admiration, elements favorable to aretaic uptake. But we must not allow the merits of an “emotionalist” view of aretaic uptake (which is suggested by some of what Zagzebski has said) to prevent seeing the merits of a wider approach. Outlining a wider approach is the task of the next section.
5 The Primacy of Reasons in the Understanding and Development of Virtues
Reasons for action are important in all the major ethical traditions, and a major concern of this paper is to determine whether, from the perspective of ethical theory, they are properly viewed as in some ways more basic than virtues. One kind of priority here (equivalent to a sense in which reasons are more basic) is conceptual: it seems clear that there could be reasons for action even if there were no virtues – though these reasons might include reasons to develop virtues. Another kind of priority is ontic: actions from virtue must be performed for an appropriate reason, one appropriate to the virtue in question, but actions performed for an appropriate reason need not express a virtue of character. Similarly, actions not performed for such a reason – even if they are both obligatory and have good effects – cannot be virtuous.12 These points are connected with still another kind of priority. Virtues have an explanatory dependence on reasons: actions can be explained by appeal to reasons – say by citing beliefs and intentions – even if there should be no virtues, but not conversely. If there were no reasons for action, there would be no virtues.
Kant and other leading deontologists, such as W. D. Ross, have maintained a point parallel to the ontic priority view, with ‘moral worth’ substituted for ‘being virtuous’: actions cannot have moral worth or be virtuous unless performed for at least one reason of the right kind, but an action might have moral worth without being virtuous – or at least without manifesting a virtue. One reason, as we have seen, is that an agent might not have achieved that positive element in character even if able, in one instance, to do a deed virtuously. The matter is more complicated for utilitarians, but many versions of utilitarianism may provide plausible grounds for a similar point (as indicated by J. S. Mill).13
Given these points, how should virtue ethics be viewed from the point of view of ethical theory? One might think that from the perspective of a plausible virtue ethics, fully moral action is not possible without virtue. But this claim would surely go too far and is not entailed by all forms of virtue ethics. Consider a virtue ethics for which virtue is not the basic normative concept (a metaethical possibility for some virtue theories) but can nonetheless provide both a normative framework extensionally equivalent to the best deontological competitor(s) and insights in ethics and moral psychology. Let me briefly explain.
The possibility of extensional equivalence is not difficult to see. There are at least two points here. First, nothing prevents a virtue ethics from so conceiving virtuous action that it occurs when and only when the act in question is obligatory on some other theory (say a version of Kant’s).14 This is equivalence concerning what acts are obligatory. Second, the two kinds of theory can agree in viewing aretaically adequate deeds – roughly those done from virtue – as coinciding with those performed for a reason that, viewed from a deontological perspective and considered as their basis, gives them moral worth.
Regarding the bearing of an adequate virtue ethics on moral psychology, recall our discussion of how role models may succeed in helping children to become virtuous. Mere imitation of the relevant act-types is not enough. Doing the right thing for an appropriate reason is needed, and it also seems sufficient – as we would expect given the three-m structure of virtuous action – for the kind of conduct that, when spontaneous in an appropriate range of situations and performed in an appropriate manner, constitutes acting virtuously. Virtuous action, then, can occur on the way to developing virtue. Those with virtue tend to act in fully virtuous ways; but acting in those ways may occur en route to developing virtue or otherwise without the agent’s having achieved that status. Apart from a view at least close to this, it is at best difficult to see how Aristotle could have given habit such a large role in the genesis of virtue. One sound idea in this vicinity (as he may have presupposed) is that habit commonly fixes in the agent’s character a tendency to perform the right kind of action – virtuous action – not that the right kinds of action cannot even occur until the relevant trait of character is developed.15
In the light of the conceptual and developmental points made in this section and earlier, we are in a good position to pursue the question whether reasons are normatively as basic as virtues. Suppose that (1) role modeling normally requires an understanding of an appropriate range of reasons (even if not necessarily in terms of the concept of a reason) for the behavior that is partly constitutive of the modeling.16 This indicates a conceptual as well as normative dependence of role modeling on reasons for action. Suppose also that (2) a kind of imitation of virtuous conduct, provided the imitation occurs for the right kind of reason, suffices for virtuous action and that such imitation may be a step toward developing the virtue(s) in question. This second point indicates (as our examples have) a dependence on a kind of response to reasons as a condition for aretaic uptake. Given all these points, it is plausible to hold that (3) if having a virtue is itself partly constituted by a proper reasons-responsiveness, we should understand fully moral action – the kind that is right, creditworthy for the agent, and includes virtuous action as a special case – as a matter of doing the right thing for an appropriate reason, and in the right way.17 One might wonder about the partial constitution idea in (3), but if we take a proper reasons-responsiveness to be an element in character and to imply an appropriately strong pattern of responses, it becomes plausible to consider that element largely constitutive of at least moral virtue. The just person, for instance, is sensitive to facts indicating a need for a just distribution, takes such facts and related ones concerning the relevant situation as reasons to act, is appropriately strongly motivated to act for the reasons in question, and competently does so if not prevented.
Suppose these three points are correct. That does not require denying something a strong virtue ethics implies – that what is obligatory for us in a given situation is what would be done in that situation by a perfectly virtuous agent. This equivalence is not implausible, but it does not by itself imply the position a virtue theorist tends to take – that an understanding of perfect virtue is a basic route, even if not the only route, to determining one’s obligations. This view seems too strong. Given how virtuous action must be understood and role modeled, it is at best unclear how we can determine what virtuous agents would intend or would do apart from an independent knowledge both of what should be done and of appropriate reasons for doing it. These considerations seem required in the first place. There is apparently an epistemic dependence of knowledge of what a virtuous agent would do on knowledge of these other considerations.
This epistemic dependence apparently explains in part why, in the normative order, reasons for action and hence, indirectly, right and wrong acts and good and bad acts are conceptually more basic than virtues. I am not suggesting that highly specific rules of action must be presupposed for the development of virtue, something one might think is implied by Kant’s emphasis on duty or from his apparently top-down conception of ethics. Indeed, we have seen how a broad spectrum of motivating reasons can be appropriate even for a single virtue, such as justice. The same spectrum is apparent in the range of types of aretaic modeling. Moreover, role models can exercise influence quite beyond eliciting virtuous action. When they are admired or even simply identified with, their being the (or a) kind of person they are can represent an aspiration and can generate its own psychologically autonomous reasons in those who admire them. If I admire you for important traits you have, I might adopt some of your likes and dislikes, which would influence desire, action, and emotions.
If, in the normative order, and especially in determining what we ought to do, reasons are more basic than virtues, they need not also be more basic in the psychological order – that of the explanation of action and other psychological elements. Suppose, however, that they are psychologically more basic, on the ground that virtues themselves are at least largely constituted by long-standing and stable patterns of intentions and desires (as motivational elements) and beliefs (as cognitive elements),18 as well as dispositions to form such intentional elements in new situations. The priority here would be ontic as well as psychological: ontic because the constituents of virtues can exist without them, but not conversely, and psychological because in the order of psychological concepts and explanations the relevant motivational and cognitive notions are prior. Neither kind of priority implies that appeals to virtue have no explanatory power; even derivative power is still power.
Moreover, even if the grounds of virtue include propositional attitudes whose contents, like those of beliefs, express reasons, it does not follow that aretaic explanations of action are reducible to explanations of them citing the relevant beliefs and motivating desires. Consider explaining why a shape is visible on white paper because it is colored. This is not reducible to explaining the same phenomenon by appeal to redness even if the color is grounded in the redness, which (as a specific color) is more basic than simply being colored. To take a moral case, suppose an agent acted from (e.g.) beneficence. The deed may depend on beliefs about what will benefit someone and on wanting to achieve that benefit, and citing these might explain the action. Still, the full content and the scope of aretaic explanations – such as one citing beneficence as explaining the deed in question – as compared with cognitive-motivational ones, is different. An aretaic explanation of an action, say explaining it as due to beneficence, locates it in the context of the character of the agent, which may normally be considered stable. This kind of explanation does not indicate just what desire or belief underlies the action; but although explaining action by appeal to a particular desire-belief pair indicates the underlying grounds, it may suggest little or nothing about overall character. At the descriptive level, moreover, attributions of moral virtues are very different from, and generally more comprehensive and temporally inclusive than, attributions of the individual motivational and cognitive elements that (at least largely) ground the virtues in question.
The case of consequentialism is difficult to deal with in connection with the issues we have been considering. But suppose we consider not a hedonistic version but a valuationally more pluralistic position such as we find in Sidgwick and Moore.19 Reasons for action may be considered conceptually autonomous even if the overall moral directive is to maximize intrinsic value. To illustrate, producing an equal distribution might be a basic reason for action of the kind important both for acting justly and for developing the virtue of justice. It need not be considered a reason for action only insofar as acting on it tends to maximize intrinsic value. Its doing that may be viewed as a favorable causal consequence; but, for certain versions of consequentialism, it need not be considered a conceptual requirement for normative status, as opposed to, say, empirically equivalent to right action. If, however, a consequentialism is (like Mill’s, one might argue) top down and conceptual in its grounding of reasons in considerations of pleasure and pain, then virtuous action and virtue itself must be understood differently – as conceptually subordinate to the master reason for action: contributing as much as one can, given one’s alternatives, to enhancing the balance of the intrinsically good to the intrinsically bad. This view does not provide a good basis for understanding either virtue or common-sense morality. In my view, it tends to assimilate what is praiseworthy, in the intuitive sense in which virtuous actions merit that description, to what is worth praising, in the sense that praising it would have optimal consequences.20
6 Virtue Ethics and the Theory of Virtue
Once it is appreciated that intellectual virtue is parallel to moral virtue, it will be clear that we should not take virtue ethics to exhaust the realm of virtue theories. There is also virtue epistemology and, applicable to both, the theory of virtue as the philosophical study of the entire aretaic domain. This paper has so far addressed virtue ethics, but virtue epistemology has been viewed by some writers as a foil,21 and considering certain aspects of it is illuminating in understanding the theory of moral virtue. I have stressed the primacy of reasons (and grounds) in understanding virtue. This final section will bear on both virtue ethics and virtue theory in general. Its point of departure is two plausible contemporary conceptions of virtue.
The first broad conception of virtue I have in mind is that of moral virtue as “excellence in being for the good,” a view developed by Robert M. Adams.22 The second is the idea that “Virtues are ways of being well oriented to morally relevant reasons,” a view developed by Garrett Cullity in “Moral Virtues and Responsiveness to Reasons.”23 Detailed discussion of either author would require more space than I have, and in any case my concern is to view these formulations (even if intended to hold mainly for moral virtue) in relation to the primacy of reasons thesis concerning virtue: the view that conceptually, ontically, developmentally, and normatively, reasons (broadly understood) are more basic than virtues.
Consider first the view that virtues are ways of being persistently for the good – ways that exhibit a kind of excellence – and assume (as most writers on virtue do) that virtues are traits of character. Can we understand what it is to be for (pro) the good simply by identifying virtues and ascertaining their targets? Perhaps we could in principle, but how are we to identify the virtues and their targets in the first place if we do not presuppose the kinds of desires and reasons for which the agent must act to manifest the virtue and, accordingly, what aretaic targets of conduct go with successfully acting for one or more such reasons?
Regarding both the pro-goodness conception of moral virtue and the reasons-responsiveness conception of it, we have seen much of the basis of these views. As argued in sections 1 and 2, both role modeling virtues and acquiring them in the ways normal human development require, on the modeling side, representing aretaically appropriate reasons and, on the uptake side, sensitivity to such reasons and, for successful development of virtue, internalizing a sufficiently strong responsiveness to them. Such internalization is central in being persistently for the good, and it requires appropriate motivational and cognitive elements – reason-expressing psychological properties that, as constitutive elements in virtues, are ontically prior to them. As to the normative side, granting that virtues are not only instrumentally good but good in themselves, are they not good at least mainly on the basis of how they manifest appropriate reasons and how, motivated by those reasons, the agent tends to hit appropriate targets connected with the reasons? The beneficent, for instance, must both tend (sufficiently strongly) to do things that conduce to the good of others and do them for non-self-interested reasons whose content reflects a kind of caring about their good for its own sake.24
The developmental points made in parts i and ii also accommodate plausible elements in the reasons-responsiveness and pro-goodness conceptions of moral virtue. Recall the case made here for the importance of reasons for action in understanding the development and constitution of virtue. Can we give an adequate account of the development of virtue – and, as part of such an account, of the distinction between acting virtuously and action from virtue – without giving reasons an essential role? Reflection on the six elements of aretaic analysis sketched in Section 2 will confirm the suggestions these rhetorical questions indicate. So, I believe, will an account of what constitutes the pluralistic good that virtues are plausibly taken to be for. Speaking more generally, in a way that holds for any virtue ethics, virtuous agents are motivationally oriented toward the goods that represent targets of the virtues; they are cognitively guided by beliefs that indicate (among other things) what action will (or might) realize those goods; and, especially in interpersonal situations, they strongly tend to act for the reasons that are based on what is required to realize or advance those goods.25
Four points will further clarify how primacy thesis accommodates the main elements in the reasons-responsiveness conception of virtue. First, on that conception, reasons have conceptual priority in that virtue is taken by its very nature to be a characteristic on the basis of which the agent responds appropriately to reasons, which are explicated independently of appeals to virtue. Second, since reasons for action are efficacious in action through the motivational and cognitive elements that represent them in the agent’s psychology, these elements are essential for actual reasons-responsiveness – they are its main causal basis – and, since their existence is prior to that of virtues, they are ontically more basic than the virtues of which they are appropriate psychological manifestations. Third, given the reasons-responsiveness conception of virtue, for those role modeling or otherwise teaching or engendering virtue, the developmental role of reasons should be expected to center on modeling not just of behavior appropriate to virtue, but also of the ground of that behavior, which is an appropriate response to reasons and, often, part of a pattern to be internalized. Fourth, the reasons-responsiveness conception indicates that it is in terms of manifesting reasons and hitting the targets they lead the agent to aim at that, in the main, we understand the goodness of virtue.
It should be stressed, here as in relation to any other implication of the primacy thesis regarding reasons and virtues, that the thesis allows for important equivalences and that, on this count and others, it accommodates important roles for aretaic concepts in both ethics and epistemology. A good person, for instance, can still be informatively and most briefly described in aretaic language; right acts may still be usefully characterized as consonant with moral virtue; and obligatory acts may be plausibly said to be acts whose non-performance would be inconsistent with virtue. Similarly, justified beliefs might be seen as equivalent to those grounded in an “exercise” of or an element in one or more virtues of intellect (in “reason,” one might say); and knowledge may be viewed as constituted by true belief so grounded. Inferential power is one example of an intellectual virtue; perceptiveness regarding the physical world is another. Both commonly lead to new knowledge when exercised in the right way.26
The primacy thesis has a perhaps unexpected benefit for answering the “situationist” critique of virtue ethics. Central to this critique is the idea that traits of character and even those constituting virtues are neither significantly predictive of the kinds of deeds one would expect from the nature or “content” of the trait or virtue, nor genuinely explanatory of the actions one would expect.27 If, in the ways I have outlined, virtues are understood in terms of the primacy thesis, then one would neither predict nor try to explain behavior simply on the basis of a virtue-ascription. On any plausible virtue theory, including a properly qualified Aristotelian one, we should not assume that virtues yield behavior on their own, since they do so in a context that may activate multiple and possibly conflicting reasons. Nor should we assume that virtues cannot yield conflicting tendencies, as where fidelity to a promise calls for action that would harm an innocent person and thereby creates a tension with kindness; and we should not assume, in either explanation or prediction, that reasons characteristic of virtue always outweigh competing reasons. Human agents are creatures with a multitude of interests, emotions, traits of character – and reasons for action. If reasons are properly taken into account – especially motivational reasons (reasons for which one acts), as based on desires, and cognitive reasons, as expressed by beliefs that guide action toward an appropriate target – then virtue theories can often be as predictively and explanatorily adequate as the reasons-responsiveness theories that are sometimes opposed to them. The level of accuracy here is still below the level found in much of physics and chemistry; but that is no surprise and goes well with Aristotle’s caution that we should demand no greater precision than the subject matter admits of.
* * *
Virtues and reasons are interconnected, and this paper shows how, in several respects, reasons have a kind of priority relative to virtues. Virtues require a major role for reasons; and virtues do not, and apparently cannot, play the role in guiding thought and action that manifests virtue unless they are internally empowered by the cognitive and motivational elements that express reasons for action and by which virtues themselves are at least largely constituted. Those cognitive and motivational elements empower much thought and action and supply reasons for thinking and acting – propositions in the case of beliefs, and projected states of affairs in the case of the motivational elements. These reasons go with the telos, the “aim,” as it were, of the virtue. Whether or not reasons are normatively fundamental or can be accounted for in terms of, say, values as still more basic normative elements, it appears that reasons are conceptually, ontically, developmentally, and normatively more basic than virtues. We cannot explicate virtues apart from appeal to the kinds of reasons for which their genuine exercises must occur, but one need not – even if one can – explicate normative reasons by appeal to the virtues specially related to them. We cannot understand the normal development of virtues, or indeed the areataic uptake to role modeling of virtues, apart from some account of how they are connected to reasons-responsiveness in their possessors. The primacy of reasons for understanding virtues is also manifested in our inability, at least for moral virtues, to understand even what constitutes them apart from positing longstanding and stable sets of wants and beliefs. These conceptual, ontological, and developmental conclusions about virtue are connected with epistemological points, such as those indicating how one can know that a person has a virtue, with ontological points connected with what psychological properties ground virtues, and with developmental points indicating how virtue is to be role modeled if it is to be well taught. None of these points implies that virtues are of secondary importance in ethical or intellectual matters, nor even that, apart from having virtue, human beings can be expected to act morally with reliable steadfastness. Whether or not virtues are fundamental in one or another philosophically interesting way, they are morally and intellectually indispensable.
Part of this paper was presented at the International Conference on Reasons and Virtues at the Australian Catholic University in Melbourne, May 2015. A commentary by Garrett Cullity was of substantial help in revising it, and a later version of the paper benefited greatly from numerous detailed comments by Christian Miller. I also want to thank two anonymous referees for the Journal.
Linda Zagzebski goes so far as to say, “Moral learning, like most other forms of learning, is principally done by imitation.” See “Moral exemplars in theory and practice,” Theory and Research in Education 11, 2 (2013), 193–206 (p. 200). Even with the qualification ‘principally,’ imitation must be taken to encompass more than just performing the morally appropriate act-type – that allows the wrong motivating reason(s).
See Nicomachaen Ethics, Book ii, esp. 1105a. See, e.g., 20–30, in which he says, “actions are not enough” (for becoming just or even learning a craft) and explains why.
This is the normal case, but in principle someone could be given systematic hallucinatory experiences – or direct access to a role model’s silent reflections – that might well suffice to teach the same patterns.
These dimensions of analysis are introduced in my “Acting from Virtue,” Mind 104, 414 (1995), 449–471.
Nothing said here implies either that virtues enable highly reliable predictions of action as opposed to patterns of action in certain circumstances (which may be difficult to ascertain or even describe); nor is it implied that virtues are possessed to the same degree by all who have them or that any specific proportion of normal persons possess them at all (a variable that may differ among cultures or subcultures). For much corroborative data on these qualifications, see Christian B. Miller, Moral Character: An Empirical Theory (Oxford: Oxford University Press, 2013), esp. Part iii.
See, for instance, her Intelligent Virtue (Oxford: Oxford University Press, 2011), e.g. p. 14.
This is argued in detail in “Acting from Virtue.”
I refer to normal acquisition and characteristic expressions of virtue. On plausible assumptions about the relation between the mental and the physical and the (broadly causal) determination of the former by the latter, certain brain manipulations could produce at least a large number of virtues. This would be an example of abnormal and doubtless not creditworthy acquisition of virtue, but it is significant as a conceptual possibility.
To be sure, if virtues are historical, at least in the way remembering is (entailing the person’s pre-existing the time in question), the idea of immediate producibility by brain manipulation is ruled out. I suggest, however, that duplication of a person is possible, at least by an omnipotent God, and that reflection will show that there is no good reason to deny to one’s duplicate all one’s virtues that – unlike any that entail a past, as does regularly remembering people one has met in the past – depend on one’s basic psychological constitution. This is arguable for all of the basic virtues.
See her “Moral exemplars in theory and practice,” cited above.
She grants that we can be “mistaken in our judgment that some person we identify as paradigmatically good is really good” but immediately adds, “I don’t think that we could be mistaken about most exemplars [the kinds of persons we admire] for the same reason that we cannot be mistaken that most of what we take to be water is water” (p. 200).
A possible exception would be what might be called collaterally virtuous actions: non-intentional but anticipated as done in or by doing what one intends, say avoiding language that would offend, though one’s aim (and all one relevant intends) is just speaking neutrally to the point at issue.
See esp. note 3 in ch 2 of Utilitarianism.
Strictly speaking, on neither kind of theory need there always be just one optimal act (unless it is disjunctive). Where two or more acts are equally adequate morally, each might be virtuous and only their disjunction is obligatory; given this, the theories may still be extensionally equivalent as regards obligation.
It might be argued that actions cannot be virtuous unless the agent has the virtue in question. This would require holding that, where the only difference between two agents who do the same thing, for the same reason, in the same way, and in exactly the same kind of circumstances, is that one has the trait of courage and the other does not but will develop it with more practice, only the former can act virtuously. I cannot argue the point here but suggest that such reasoning weights the fixity of virtues as traits too heavily relative to the main elements in them that make virtuous action admirable in the first place.
Cf. Annas’s point, regarding teaching situations and with virtue in mind as importantly similar to skill, that “With skills of any complexity, what is conveyed from the expert to the learner will require the giving of reasons. … The ability both to teach and to learn a skill depends on … giving and receiving reasons” (op. cit., p. 19).
It may seem that we may take descriptions of ways of doing (given typically by adverbs of manner) things to be reducible to actions in their own right, as speaking very loudly might be thought to reduce to yelling. But this would be a mistake, as I have argued in Means, Ends, and Persons: The Meaning and Psychological Dimensions of Kant’s Humanity Formula (New York and Oxford: Oxford University Press, 2016). On the nature and importance of this adverbial dimension of action Donald Davidson’s “The Logical Form of Action Sentences” is instructive in his Essays on Actions and Events (Oxford: Oxford University Press, 1980), 105–148.
This conception of virtue is developed in earlier work of mine, e.g. “Responsible Action and Virtuous Character,” Ethics 101, 2 (1991), 304–321, and “The Psychology of Character and the Theology of Virtue,” in Craig Steven Titus, ed., The Psychology of Character and Virtue (Arlington: Institute for the Psychological Sciences Press, 2009), 101–118. For a related conception of virtue viewed in the light of empirical inquiry, see Miller, op. cit.
Hedonism is itself pluralistic given how many kinds of things yield pleasure and how many yield pain, but knowledge and virtue need yield neither, which is one reason their views are more pluralistic than hedonism. See, e.g., G. E. Moore, Ethics (Oxford: Oxford University Press, 1912), Ch 8; and for an extensive pertinent discussion of both Sidgwick and Moore, see Thomas Hurka, British Ethical Theorists from Sidgwick to Ewing (Oxford: Oxford University Press, 2014).
This is argued in detail in my “Internalism and Externalism in Moral Epistemology,” Logos 10 (1989), 10–37.
See, e.g., Linda Zagzebski, Virtues of the Mind (Cambridge: Cambridge University Press, 1996) and also Richard Foley, Working Without a Net (Oxford: Oxford University Press, 1992).
See A Theory of Virtue: Excellence in Being for the Good (Oxford: Oxford University Press, 2006).
Presented at the international conference on Reasons and Values, Australian Catholic University, May 2015.
This point is argued in detail in my Means, Ends, and Persons.
Cf. Miller’s conception of a character trait (of which virtues are a species) as “A disposition to form beliefs and/or desires of a certain sort and (in many cases) to act in a certain way, when in conditions relevant to that disposition” (op. cit., p. 6).
This conception of knowledge fits well with some significant versions of virtue epistemology. See, e.g., Linda Zagzebski, especially her Virtues of the Mind (op. cit.); Ernest Sosa, A Virtue Epistemology (Oxford: Oxford University Press, 2007) and Knowing Full Well (Oxford: Oxford University Press, 2011); and John Greco, Achieving Knowledge (Cambridge: Cambridge University Press, 2010). Critical discussion of the kind of externalist view developed in this tradition is found in John Greco, ed., Ernest Sosa and His Critics (Oxford: Blackwell, 2004). The critical appraisals include my “Intellectual Power and Epistemic Virtue,” pp. 3–16. For wide-ranging discussion of virtue epistemology in the wide context of virtue theory in both the practical and theoretical domains, see Heather Battaly and Michael Slote, “Virtue Epistemology and Virtue Ethics,” in Lorraine Besser-Jones and Michael Slote, eds., The Routledge Companion to Virtue Ethics (New York: Routledge, 2015), 253–269. This paper also considers the virtue-responsibilism of James Montmarquet (255–56) in his Epistemic Virtue and Doxastic Responsibility (Lanham, 1993) and others.
The situationist critique is well developed in John M. Doris, Lack of Character (Cambridge: Cambridge University Press, 2002) and, in a different way, by Mark Alfano in Character as Moral Fiction (Cambridge: Cambridge University Press, 2013). Informative critical discussion of this critique is provided by Lorraine Besser-Jones, “The Situationist Critique,” in Besser-Jones and Slote, op. cit., 375–84. She takes “moral character to consist in one’s moral commitments, one’s dispositions to act, how it is that one’s moral commitments influence and interact with one’s dispositions,” which implies that there is a “gap between one’s moral commitments and one’s behavioral dispositions” (p. 381). This contrasts with the stereotypic trait-by-trait conception of character implicit in some discussions of character or virtue, and it indicates a more complicated basis for prediction and explanation.