CONSEQUENTIALISM AND CHARACTER

At first glance, it might appear just obvious what consequentialists should say about virtue. Virtues are simply character traits that are instrumental in promoting the good (for example, happiness). Indeed, this has been the view of many leading consequentialists: Bentham and Mill amongst the classical utilitarians, and Julia Driver and Roger Crisp amongst recent influential philosophers.Footnote ¹ Discussing dispositions of character, Bentham says: ‘It is with disposition as with everything else: it will be good or bad according to its effects: according to the effects it has in augmenting or diminishing the happiness of the community.’Footnote ² This instrumentalist view seems to emerge naturally out of the core consequentialist thought that what matters is that the good be maximized, and so character traits, just like anything else, should be assessed according to a general formula: the best x is the x that results in the best consequences. For example: the best available action is the one which results in the best consequences; the best set of political arrangements is the one that results in the best consequences; the best moral code is the one that results in the best consequences; the best motives are the ones that result in the best consequences.Footnote ³

There are two reasons why an instrumentalist view of virtue should not, in fact, simply be accepted as the default view of virtue for those who are consequentialists about the value of actions. The first is the straightforward point that consequentialists are not, and cannot be, instrumentalists about everything. They must regard some things as intrinsically good. Most consequentialists, for example, regard pleasure as one thing that is intrinsically good. The pleasure of playing football may help me relax and so be instrumental in promoting my long-term health, but the value of the pleasure is clearly not exhausted by this instrumental benefit. The pleasure also has intrinsic value; it is desirable for its own sake.

And virtue itself appears a serious candidate as something that has intrinsic value, and not just instrumental value. In Chapter 4 of Utilitarianism, Mill acknowledges that most people consider virtue to be intrinsically good, desirable for its own sake, and tries to accommodate this valuing of virtue.Footnote ⁴ In the Groundwork of the Metaphysics of Morals, Kant takes himself to be drawing out our ordinary, common-sense moral thought when he argues that we value virtue or ‘the good will’ more highly than anything else: ‘A good will is not good because of what it effects or accomplishes – because of its fitness for attaining some proposed end: it is good through its willing alone – that is, good in itself.’Footnote ⁵

Having taken into account the possibility that a character trait may itself be intrinsically good, we might think that consequentialists may eschew the claim that virtues are merely instrumentally good, but think that they are committed instead to the claim that the virtues are simply those character traits which bring about the most good, where ‘bring about’ is deliberately intended to be inclusive of goods realized in the trait itself, as well as in those extrinsic goods to which the trait is instrumental.

Treating the virtues in this way would also be a mistake. There is a second, perhaps less obvious, reason why consequentialists should not read off judgements about which traits are virtues from judgements about the amount of good the traits produce. This reason is that the concept of a virtue seems more closely linked to the question of what features of a trait give us reason to admire its possessor than to the question of what features of a trait give us reason to cultivate that trait. Even if we think that we should cultivate, in ourselves and in others, those traits which bring about the most good, it does not follow that we can tell which traits are virtues just by looking at how much good they bring about. It is important to leave open the possibility that we may have reason to cultivate traits which do not display any virtue and are perhaps even vicious, and that sometimes we may have strong reasons not to cultivate traits which are virtues. This will become apparent when we look at the major objection to the instrumentalist view in the next section.

The characterization of virtue in terms of what there is reason to admire can instructively be compared with Mill's characterization of moral obligation in terms of what there is reason to blame agents for doing. Many consequentialists claim that we have a moral obligation to do what brings about the best consequences. This seems wrong, however. It is an entrenched part of our moral thinking that there are actions which are supererogatory, which go beyond the call of duty, which are good but not morally obligatory. Mill makes clear his acceptance of this in August Comte and Positivism:

And in Utilitarianism:

Mill proposes that what is morally obligatory be characterized with reference to the moral sentiments − of blame, and guilt or conscience. To say that something is morally obligatory is not to say that it is the most ‘expedient’ action to perform. Rather it is to say that someone would merit feelings of blame from others, and the reproaches of one's conscience (which Mill conceives in a sentimentalist mannerFootnote ⁸), if he failed to perform it.Footnote ⁹

I propose that the concept of virtue is likewise well understood by reference to our sentimental reactions. In this case, the relevant sentiment is admiration, or more narrowly, what we might call moral admiration.Footnote ¹⁰ On this proposal, a virtue is a character trait that we have reason to admire morally, or which gives us reason to admire its possessor. Further, we should note that it will not be plausible to say that just because some trait produces good consequences, then we have reason to admire it, or to admire a person who possesses that trait. Some despicable trait may, in unusual circumstances, turn out to have good consequences. Nor will it be plausible to say that there is reason to feel admiration for some trait just because my feeling admiration itself will have good consequences. It may have positive consequences if I feel admiration for my vain and bullying line manager, but this does nothing to show that he is admirable, that his character merits admiration.Footnote ¹¹ Just as with the question of what there is reason to believe, there seems in general to be no plausible straightforwardly consequentialist story to be given about what there is reason to feel. And if virtue, like moral obligation, is a matter of what there is reason to feel, then a straightforwardly instrumentalist account of virtue will not strike us as plausible.

THE APPROPRIATE RESPONSE VIEW

The leading competing candidate view for how consequentialists should regard virtue is what I will call the appropriate response account. This view is non-instrumentalist, allowing that virtue be intrinsically good, yet is fully compatible with a consequentialist theory of right action. A version of this view is defended by Thomas Hurka in his book Virtue, Vice and Value.Footnote ¹² In this section, I will argue that the appropriate response view shares the purported strengths of the instrumentalist view, and not only avoids the major objection facing instrumentalism, but gives a convincing diagnosis of where it goes wrong.

The appropriate response account is compatible with different axiologies and distributive principles. For simplicity, consider utilitarianism, which states that what is good is happiness or well-being. As already noted, it has often been objected to utilitarianism that it ignores the intrinsic goodness of virtue. In light of this, one might have expected more consequentialists to adopt an axiology that allows that there are two distinct intrinsic goods, happiness and virtue, which we have reason to promote. This would be the smallest modification we could make to utilitarianism, if we find ourselves attracted to the view, but feel that it ignores the intrinsic goodness of virtue. As Hurka suggests, one possible explanation of why consequentialist views which regard virtue as an intrinsic good distinct from happiness have not been more popular is that it is not obvious that consequentialists can consistently acclaim virtue as being intrinsically good. After all, virtue is already a moral notion. And it may be thought constitutive of consequentialism that it characterizes all moral notions in terms of goodness or badness:

However, as Hurka goes on to argue, there is another relation that virtue might have to the good, besides an instrumental one. Instead, it may be an intentional relation: virtue is having an appropriate response to the good, or being correctly oriented towards the good. Hurka calls this loving the good; Robert Adams calls it being for the good. We can characterize virtue as appropriate response, in our desires, feelings and actions, to what is good and bad (loving the good and hating the bad) and vice as inappropriate response to what is good and bad (hating the good, loving the bad, or being indifferent to either).Footnote ¹⁴ Appropriate response to value includes desiring that the good come about, taking pleasure in the fact that good has come about, and striving to bring about the good where one can. Appropriate and inappropriate responses to value are themselves very plausibly intrinsically good and intrinsically bad respectively. And so if virtue and vice consist in such responses, then virtue and vice are indeed respectively intrinsically good and intrinsically bad, just as common moral thought would have it.

A welfare-based form of consequentialism therefore might say:

1. (I) Happiness is intrinsically good; suffering is intrinsically bad.

2. (II) Appropriate response to happiness and to suffering is intrinsically good; inappropriate response to happiness and to suffering is intrinsically bad.

3. (III) Virtue consists in appropriate response to intrinsic goods and bads; vice consists in inappropriate response to intrinsic goods and bads.

4. (IV) So virtue is another thing besides happiness that is intrinsically good, and which we ought to promote; vice is another thing besides suffering which is intrinsically bad, and which we ought to eliminate.

Once we adopt an appropriate response account of virtue along these lines, the form of consequentialism just outlined, which recognizes two distinct goods that are to be promoted – happiness and virtue – begins to look very coherent and attractive, fitting nicely with the original spirit of utilitarianism. After all, just in calling happiness good, utilitarians already seem committed to the idea that a certain range of responses towards it is appropriate, and another range of responses inappropriate.

It is, of course, possible to characterize virtue as appropriate response to value (and thus in terms of the non-instrumental features of traits) while holding that such appropriate response is not itself intrinsically good. Not everything we take to be an appropriate response seems itself to be intrinsically valuable. For example, one may respond appropriately or inappropriately to the evidence one has. In this case, it would ordinarily be odd to say that appropriate response is intrinsically good, and inappropriate response intrinsically bad. But when considering appropriate and inappropriate response to value itself, it seems quite natural to regard the appropriate responses (such as taking pleasure in someone's happiness) themselves as intrinsically good, and inappropriate ones (taking pleasure in her suffering) as intrinsically bad. A world in which people have such appropriate responses seems to be better than a world that contains the same overall level of happiness, but in which people have the opposite responses, taking pleasure in the suffering of others and being pained when others are happy. Although utilitarians have generally not acknowledged appropriate response to happiness and suffering as further goods, it is not clear why they have not. If happiness really is so good, one should not be apathetic towards it. One should love it, in just the ways mentioned. These responses are called for directly by the nature of happiness. And it is quite intuitive to say that responding in such a way is itself good, something we have reason to promote for its own sake, and responding otherwise is bad.

Other variations are of course possible which have different axiologies or distributive principles, but which likewise make room for virtue, understood as appropriate response to value, as a distinct intrinsic good.Footnote ¹⁵ One interesting variation (akin to the prioritarian view that an increase in the welfare of a badly-off person is more valuable than a similar increase in the welfare of a better-off person) is the view that an increase in the welfare of a virtuous person is more valuable than a similar increase in the welfare of a vicious person. Such a variation promises to accommodate intuitive views about desert.Footnote ¹⁶ My aim here, however, is to establish the appeal of the appropriate response account, not to consider which version of it is the most attractive.Footnote ¹⁷

The appropriate response account appears, then, to be a strong competitor to the instrumentalist view, fitting well with our intuitions about virtue. However, the appropriate response account is not only independently appealing, but also gives a convincing diagnosis of where the instrumentalist view goes wrong. The central objection to the instrumentalist view is that, in unusual circumstances, it acclaims paradigmatic vices as virtues, and paradigmatic virtues as vices.Footnote ¹⁸ Julia Driver, the leading modern defender of the instrumentalist view, recognizes the objection, and offers her own example as the basis of discussion:

The reason why this sort of example amounts to a convincing objection against the instrumentalist view is that it vividly shows that the view commits us to accepting any trait as a virtue, no matter what kind of response to value it involves, so long as it is, in that world, systematically a good means to producing good consequences.Footnote ²⁰ The claim that the beaters’ trait is a virtue or a character excellence rings hollow precisely because it involves a wholly inappropriate response to the evil of suffering. Taking pleasure in the suffering of children is surely a vice, and what seems to make it so is that it involves an intrinsically bad orientation towards the badness of suffering.

Our natural response to this sort of case makes clear, I think, that our judgements of virtue and vice are governed not by considerations of whether there is reason to cultivate the trait in question, but rather by what we have reason to feel about someone who possesses the trait. We have no reason to admire the beaters simply in virtue of them having this systematically beneficial trait (since they do not care about the benefits their behaviour secures). In fact, we have reason to feel a moral disgust towards their sensibility.Footnote ²¹ It seems pretty clear that it is our response to the affective question rather than the practical one which lies behind our intuitive judgements about whether a trait is a virtue or a vice. It sits very uneasily with us to acclaim the Mutor beaters’ trait as a virtue, notwithstanding the plausibility of the claim that the Mutors have strong reasons to cultivate the trait in some of their number, given the very significant value of the consequences at stake. Such consequences fail to establish the trait as a virtue. Instead, it seems a clear case where we have instrumental reason to cultivate a vice in certain agents, because having this peculiarly delineated vice will have such good overall consequences.

The appropriate response account therefore not only avoids the very counter-intuitive conclusions to which the instrumentalist view is committed, but also implies the correct account of why the instrumentalist view gets the wrong answer in such cases. This is important in deciding between the two accounts, since the theories will, with regard to more orthodox character traits and more orthodox circumstances, generally acclaim very similar character traits as virtues, since appropriate response to value generally has systematically good consequences, and very many traits which have systematically good consequences are ones which involve appropriate response to value.

We can also see how the psychological minimalism of the instrumentalist view fits ill with our common understanding of virtue if we consider another case addressed by Driver, that of Huckleberry Finn. In this example, Huck does what he believes to be wrong in helping Jim, a slave, escape from his owner. Driver explains:

Driver uses this example to argue against views of virtue that insist that knowledge of what is good or right is a necessary psychological condition for virtue. The instrumentalist view gives the more plausible verdict in this case, that Huck is indeed virtuous, since he acts from a trait that has systematically good consequences. However, the appropriate response account offers a more plausible necessary psychological condition for virtue than a correct moral belief condition, namely that one need be correctly responsive to value in order to count as virtuous. It is the fact that Huck responds appropriately to the goods at stake, and is motivated by these, that explains our verdict that he shows virtue in helping Jim escape. As John Skorupski writes, ‘[Huck] is responding to reasons which are morally salient, even though he does not recognize that that is what they are.’Footnote ²³ If Huck had some trait which happened to produce good consequences, but which did not involve motivation by the relevant goods at stake, we would not be so ready to acclaim his behaviour as virtuous. The appropriate response account's diagnosis of where other non-minimalist accounts go wrong seems clearly preferable then to the instrumentalist's diagnosis.

THE VALUE OF THE VIRTUES

The appropriate response account seems then to compare well against the instrumental view, correctly diagnosing where the instrumentalist view goes wrong. However, it might seem that the appropriate response account itself faces a strong objection – that when we examine the character traits that we standardly regard as paradigm virtues, we see that we value many of these traits predominantly for their instrumental features.

Consider honesty. If we were to ask why people should be honest, some might say that it is intrinsically a good way to be, or that it involves correct response to, or motivation by, things that we value. More likely, however, would be an appeal to the good consequences that honest behaviour is instrumental in securing. A society with widespread dishonesty will not flourish. Honest behaviour instils trust, and allows us to rely on each other, facilitating cooperation, enabling us to succeed in projects that we would not be able to succeed in on our own. The instrumental benefits of being able to rely on others to tell the truth are manifold. If people have a strong disposition to tell the truth, to refrain from deception, and to respect the property of others, and engage in other manifestations of honesty, then we can overcome ‘prisoner's dilemmas’, and secure results that are mutually beneficial, that would be unavailable if people were always prepared to lie or to steal when it was in their immediate interest.

Or consider patience, another virtue that seems to be valued largely for its instrumental benefits in a large variety of contexts. Think of the disposition of a teacher to persevere in explaining to a child who struggles to understand. Without patient behaviour, without a willingness painstakingly to go over material again with the struggling child, without a disposition to keep one's temper, the child will fail to learn. Or consider the patience of an innovator or enquirer, persevering to find the truth. There are manifest benefits of the spirit encapsulated by the famous quote (falsely) attributed to Edison, ‘I have not failed. I have just found 10,000 ways that do not work’. Such patience seems to be worth cultivating in ourselves and in others largely because it is a disposition of character that allows us to achieve good results that we would not achieve if we were instead disposed to become frustrated and dispirited, and to give up after early failures.

If we value many of the virtues predominantly for their instrumental benefits, does this not speak against the appropriate response view, and in favour of the instrumental view?

In fact, the defender of the appropriate response view can give a convincing two-part reply to this challenge. First, he can respond by insisting again on a sharp distinction between (i) the practical question of which traits we have reason to cultivate, and (ii) the separate normative question of what we have reason to feel towards the possessor of the trait, and more specifically whether we have reason morally to admire him. It is our judgements about what there is reason to feel that appear to underlie our judgements about what makes a trait a virtue. The appropriate response account can agree that many of our strongest practical reasons to cultivate the virtues are given by the instrumental benefits of the trait. For the type of consequentialism outlined above, for example, which recognizes two goods to be promoted, happiness and virtue, it will be natural to say that the happiness produced, and the suffering averted, by cultivating the traits we regard as virtues provide strong reasons to do so. But the question of how to feel about possessors of a given trait is clearly distinct.

The goods to which a virtue is instrumental will frequently be greater than the intrinsic good of the virtue itself (the good of appropriate response). Sometimes, these former goods will be best secured, and evils best averted, by inculcating habits and dispositions which do not involve appropriate response to value, and so which we have no special reason to admire and should not be regarded as virtues. The most striking examples of this will be traits like that of the Mutor beaters, which involves positively inappropriate response to value. But there are also more this-worldly cases of dispositions or habits of character which we have strong reason to cultivate because of their instrumental benefits, but which do not involve any particular appropriate response to the good and so are in no way morally admirable or virtuous. Consider people's motivations to obey the law. It may be ideal if people are characteristically motivated to obey laws by the thought of the goods secured by conformity with the laws in question. But if a disposition to obey the law from fear of punishment is more reliable in motivating people to obey, then a trait which involves this motive is one we have reason to cultivate, even though it bespeaks no particular virtue.

The second part of the reply that the appropriate response account should make is to emphasize that the view will in fact end up acclaiming traits such as honesty and patience, which we value primarily for their effects, as paradigmatic virtues. One only properly has the virtue of honesty or of patience if one is motivated by the goods that are typically secured by the behaviour in question. Granted, someone who has the mere habit of perseverance, or of truth-telling, without being at all motivated by the goods that their habit secures, might loosely be described as patient, or as honest. But unless he is appropriately motivated, we will be less inclined to say that he is a truly patient or a truly honest person; we will not be inclined to say that he has the virtue of patience, or of honesty. If one has the behavioural disposition as a pure habit of action, or perhaps out of fear, with no sensitivity to the goods that are liable to be secured by this type of behaviour, one shows no virtue in having the disposition.

The truly patient teacher is not just one who has a disposition to persevere, without interrupting or losing her temper, but one who behaves in this way, and cultivates this disposition, out of a concern for the goods that can be secured by this behaviour – out of concern that the pupil comes to learn what she otherwise would not learn. If one simply has a habit of not interrupting as the struggling child tries to articulate herself, without any care for the benefits to the child of one's forbearance, then one does not count as properly having the virtue of patience. True virtue then is not something possessed as a mere habit of action, but instead involves a disposition to perform certain actions that is in some way motivated by the goods typically instantiated or produced by such actions.

This account nicely accommodates the appealing idea that an agent may exhibit the virtue of honesty even in cases where telling the truth will not best promote the good. Having a well-entrenched desire to tell the truth, having a deep-seated dislike of lying, is likely to be a better bet in promoting the good systematically than possessing a mindset which involves no emotional commitment whatsoever to truth-telling as such. But in order to be fully virtuous, this emotional commitment must have a motivational link to the goods typically secured, or evils typically averted, by truth-telling. If instead an agent had a general emotional commitment to truth-telling which entirely floated free from any responsiveness to those goods which truth-telling typically secures, that would seem to bespeak a form of mere fetishism rather than virtue.

It should be emphasized here that the appropriate response view is not committed to the idea that, for each virtue, there must be some particular distinctive good at stake to which the agent correctly responds. One could think, along with Hurka, that the virtue of honesty specifically involves appropriate response to the good of knowledge, but an appropriate response theorist is not committed to this. The goods which honest behaviour typically produces might be very disparate. (On the welfare-based view outlined above, all the goods would, of course, have to come under the general heading of contributions to well-being, but again this view is only one version of consequentialism which can make room for an appropriate response account of virtue.) On this model, honest behaviour is picked out not by reference to some single good distinctively promoted by such behaviour, but instead by reference to specific types of behaviour (those involving truth-telling, promise-keeping and adherence to property rules, for example), which prove to be instrumental to various goods.

To conclude, the appropriate response account can acknowledge the practical reasons generated by the instrumental value of paradigm virtues like honesty and patience, while also doing justice to their motivational component.

THE AGENT-RELATIVITY CHALLENGE

The view that virtue consists in appropriate response to value may be thought to face a second objection. Value is always agent-neutral: the goodness of a person's well-being, for example, does not vary according to the relation in which he or she stands to any given agent. But much which is commonly regarded as virtuous seems to be agent-relative: it involves patterns of emotional response, and of behaviour, which exhibit partiality to some over others, dependent on the relation in which they stand to the agent.

The first point to note in reply to this worry is that the appropriate response account is not committed to the idea that the most virtuous available pattern of emotional response to disparate goods is one where love or concern is perfectly proportioned to the degree of goodness involved. The most fitting degree of love towards any valuable object, considered individually, is that degree which does justice to its object. But there are simply far too many valuable objects in the world for us, given our severely limited emotional capacities, to have the most fitting degree of love for each. Every human person is precious, and the most fitting attitude towards each is one of deep love, concern and devotion. But while it may be possible for God to have such an attitude towards every person, we clearly cannot. The question which confronts us then is: what is the most fitting available pattern of emotional responses to the many disparate goods with which we meet? How may we best love the good, given our extremely limited capacities for love?

A standard agent-neutral conception of value does not commit us to the notion that the best set of emotional responses to equally valuable objects is to feel exactly alike towards each, when we can instead fully do justice to the value of some, while merely ‘acknowledging’ the value of others. It is open to the appropriate response theorist to say that one better loves the good by fully fittingly loving some good things, while loving others less, than by perfectly proportioning one's limited love across all goods, and falling short of fully fittingly loving any good. As Robert Adams writes:

Suppose that a close relative of mine has a successful operation to cure a long-standing illness that has been for her the source of intense anxiety and physical pain. My natural response to this is the most intense joy and relief; I rejoice for her sake. In one clear sense, these emotions are quite simply the fitting responses to her being cured. Suppose now that an alert system is set up whereby I receive a message from the National Health Service every time someone somewhere in the country has a successful operation that affects his or her life in a similar way. Clearly, I cannot have the fully fitting response to every such happy event. Nor should I strive to: attempting such full-hearted emotional engagement with every person, aiming to proportion my love perfectly to the size of every good, will end in having an insufficient response to any good, rather than to having a sufficient love for every good. The same points are even clearer when we consider the appropriate response of being pained by the suffering of others. A fully fitting response is most typically felt specifically in response to the suffering of loved ones, and only to the suffering of strangers when it is made especially vivid to us. But striving to perfectly proportion our emotional responses to the degree of badness of each person's suffering could result only in one of two undesirable outcomes: an inability to respond fully fittingly to any person's suffering, or else a devastatingly incapacitating emotional exhaustion.

Our limited emotional capacities also have important implications in determining what is the most appropriate response in action to disparate goods. It has been well noted by consequentialists that aiming to promote the good in a persistently direct manner is liable to be much less effective than doing so in indirect ways which involve behaviour that looks agent-relative.Footnote ²⁵ Our best bet in promoting the good will not be constantly to have in direct view the goal of promoting the overall good, but instead to entrench ourselves in deep personal relationships, involving strongly partial emotions, in strong commitment to specific personal projects, and in general patterns of thought, attention and emotional engagement which home in on particular goods.

Close personal relationships that involve partial behaviour will be an integral part of any lifestyle likely to promote the overall good, for the following familiar reasons: (i) Deep human relationships, which involve the sorts of emotional responses discussed above, which we cannot have towards everyone, are a central (perhaps the central) constituent of human welfare. (ii) Each of us is especially well positioned to promote and protect the well-being of those close to us. (iii) Intimate relationships that involve strong partial commitments provide necessary psychological foundations for a lifestyle that involves doing good for those beyond the immediate circle of our personal affections.Footnote ²⁶ (iv) Given the natural affections we have, we naturally come to form close relationships which generate expectations of attention, love and support, and the disappointment of such expectations is something which is strikingly painful (hence the virtue of loyalty, construed as a strong disposition to avoid violating the natural expectations of those to whom we are committed). (v) Only when our emotions are engaged – when we feel a love, passion or commitment to some person or project – will we be able persistently to maintain the sort of attention, focus and ability to persevere that are necessary prerequisites for achieving many of the most valuable things in life; in order to promote the good in a consistent way, we need an ability to commit to a particular project which, in Adams's words, ‘engages enough of our love to motivate us to follow it’.Footnote ²⁷

Persistent direct consequentialist deliberation is counter-productive. A disposition constantly to look askance to see if there are alternative goods that we could be better promoting is liable to undermine the sort of focused attention and full emotional engagement with particular goods which is a necessary condition of many of the most valuable achievements of which we are capable. One who thinks that virtue is appropriate response to the good will emphasize that we ought from time to time to reflect on the value in the wider scheme of things of the particular goods to which we commit ourselves. We should not run the risk of devoting our lives, loves and energies to something trivial, or worse still to something evil. The most virtuous agent then will be one who has fully fitting emotional engagement with particular goods, and who has the sorts of deep relationships, and devotion to particular projects, which are likely to be necessary for him to promote the good, but who occasionally steps back from such commitments and assesses whether his way of life embodies a sufficiently good overall pattern of response to the good, given human limitations. This might appear an overly intellectualized conception of what is required in order to be virtuous. But striking the right balance between commitment to loved ones and personal projects, and sensitivity to other goods, is a central part of our ordinary views of virtue and vice. Someone who is so entrenched in particular relationships and projects that she never takes an impartial perspective on them – who is perhaps so devoted to her own family, or her country, that she loses all sensitivity to the good of others – clearly has a serious fault of character.

The foregoing observations can lay the foundations for more detailed treatments of specific virtues, such as loyalty, which seem centrally to involve agent-relative responses. Giving certain types of disproportionate attention and concern to those to whom one stands in a special relationship may be at the core of loyalty, but we can see now that this is not precluded by the appropriate response account. Interestingly, attributions of disloyalty very rarely occur in response to a tendency to give strictly equal weight to the concerns of strangers to those of people to whom we stand in a special relationship. Instead common instances involve, like so many other vices, giving a disproportionately large weight to one's own concerns. The fair-weather friend who abandons a supposed loved one who falls on hard times or becomes ill, in favour of new, easier friendships; the captain who deserts a sinking ship, leaving his crew in the lurch; the traitor who defects in exchange for a bribe; the teenager who gives in to the pleasures of gossip and reveals her friend's secrets – all these paradigm instances of disloyalty involve ignoring the claims and interests of those close to us, not for the sake of promoting the impartial good, but for the sake of one's own interests. Someone who is genuinely, non-self-deceptively motivated by considerations of the overall good in giving preference to strangers over loved ones is far less clearly guilty of a vice of disloyalty.

Some other virtues seem agent-relative in that they require the agent to give special attention to aspects of his own behaviour which he is not required to give to the same features of others’ behaviour. We generally think of the person who has the virtue of honesty, for example, as one who has a special concern with his own truth-telling, fidelity to promises or respect for property, rather than having an equal concern for anyone’s truth-telling, fidelity to promises or respect for property. But again this specialized concern can be endorsed by the appropriate response account. The best means available to most of us for promoting the goods which distinctively result from truth-telling, promise-keeping, respect for property and so on is to develop in oneself a disposition to act in these sorts of ways. (Another is the inculcation of such habits in one's children.) A disposition to make it one's business to ensure that others – in particular, rational adults – keep their promises or do not tell lies is far more likely to have mixed results, given the limited control I have over the dispositions of others, and the clear undesirability of my attempting to acquire such control. Fostering one's own honest dispositions is simply a much stronger bet in promoting the relevant goods than policing the same dispositions in others.Footnote ²⁸

The fit between the sorts of justified partiality recommended by the account proposed here, and the sorts of partiality regarded as justified by pre-theoretical moral thought, is likely to be imperfect. Some revisionism may be required, and the plausibility of such revisions will play into the overall assessment of the account.Footnote ²⁹

THE DISJUNCTIVE VIEW AND THE MIXED VIEW

Let me conclude by outlining how the foregoing discussion not only puts us in a position to see why an ‘intrinsicalist’ account like the appropriate response account should be preferred to a merely instrumentalist account, but also why it should be preferred to two other accounts of virtue discussed in the recent literature: the ‘disjunctive minimalism’ defended by Luke Russell,Footnote ³⁰ and the ‘mixed view’ discussed by Driver and by Todd Calder.

The disjunctive view of virtue says that either instrumental or intrinsic goodness can be sufficient in order for a trait to count as virtue. Russell's view attempts to accommodate the attractions of both instrumentalism and intrinsicalism. In particular, it endorses Driver's psychological minimalism, the idea that any character trait can be a virtue so long as it has good consequences, irrespective of its non-instrumental features, even where they involve false belief or incorrect orientation to value. But it also says that if a trait is intrinsically good, for example by virtue of involving correct orientation to value, then that trait is a virtue. A trait may then simultaneously be both a vice and a virtue. The theory purports to have the best of both worlds.

We ought to reject this view nonetheless. The main problem is that it faces similar objections to the instrumental view. As we saw above, it is implausible to acclaim a trait as a virtue if it involves a wholly inappropriate response to value, as do maliciousness, cruelty or callousness. This is so even if the consequences of having the trait turn out to be systematically good ones. Russell's disjunctive view is committed to acclaiming the Mutor beaters’ trait as a virtue, because it has such good consequences, even though the essence of the trait is a disposition to malicious pleasure. Russell's account differs from Driver's in being able to add that the beaters’ trait is also a vice, since it is intrinsically bad. But intuitively, this is not enough. Calling the trait a virtue, though it involves taking active delight in the suffering of innocent children, simply rings false.

We might put the point by saying that we evaluate something as a virtue or as a vice according to a set of internal standards. Compare the case of acclaiming some portrait as a great work of art. In some unusual circumstances, my hastily sketched drawing of a stick man may somehow turn out to have significant instrumental benefits. Perhaps my drawing somehow leads to you meeting the love of your life. Or perhaps a benevolent demon sets up the world in such a way that whenever I make a hasty sketch, the lives of many innocent and happy people are extended by 50 per cent. These fortunate effects of the picture are very good, and if I know about them, give me strong reason to do the sketch. But they do nothing to affect the status of my drawing as a great work of art or otherwise. We would emphatically not say that it is a great work of art in virtue of its happy consequences, but also at the same time a poor work of art in virtue of its intrinsic features.

In just the same way, the happy consequences of the Mutor beaters’ trait simply do nothing to affect the status of the trait as a virtue or as a vice. Whether a trait is a virtue or a vice is to be decided according to the relevant internal criteria, namely whether the trait possesses features which make it morally admirable or morally condemnable; the incidental benefits of the trait are not relevant to that question, even if they are relevant to the practical question of whether to cultivate the trait and relevant to the question of whether we should be happy under the circumstances that some people have the trait.

Russell himself concedes that his theory of virtue is ‘messy’. It is messy because it gives the status of being a virtue and the status of being a vice to one and the same trait.Footnote ³¹ This messiness is easily avoided. In cases where the intrinsic features of a trait are bad but the instrumental features are good, it is tidier and simpler to limit ourselves simply to saying that there is something bad about the trait and something good about it. It may after all be true of some other trait that it has some instrumental benefits and also some instrumental disbenefits. Are we likewise to say of that trait that it is both a virtue and a vice, since there is something good about it and something bad about it? This surely would be messier still, but it follows the same logic as Russell's view. In spite of the messiness involved in his own account, Russell feels driven to it by the need to accommodate the attractions of the instrumentalist view. But this is unnecessary.

A second alternative is what Driver calls a ‘mixed view’ of virtue, which requires that a trait's status as a virtue is dependent on both its intrinsic, psychological features and its consequences. Todd Calder convincingly argues that mixed accounts fare less well than pure ‘intrinsicalist’ views like the appropriate response account. While the mixed view avoids having to acclaim the Mutor beaters’ trait as a virtue, it does result in similarly counter-intuitive results in failing to acclaim benevolence, for example, as a virtue in unusual worlds where it fails to have the systematically good consequences that it has in the actual world. We should note again that whatever attractions this view has are better satisfied by the appropriate response account. Driver locates the appeal of the mixed view as follows: ‘Most people find some kind of mixed view the most intuitively plausible because such a view accommodates the intuition that “good intentions” are somehow necessary to being a good person, but that “good intentions” are not enough.’Footnote ³² But what is needed to complement ‘mere’ good intentions in order to have authentic virtue is clearly not just that happy consequences result. Instead, what is needed is the accompaniment of good intentions with will or effort to secure the good when possible. The identification of active striving for the good as one way, perhaps the central way, of being for the good better accommodates the intuition in question than appeal to the actual consequences of the trait. One is deficient in virtue if one desires the good of others ‘as a mere wish’, in Kant's phrase, while being unwilling to make any active effort to secure that good when one is in a position to do so.