II. CONSEQUENTIALIST THEORIES

Many consequentialists about right action are also inclined to characterize virtue and vice in terms of the consequential, or instrumental, properties of these character traits. For instance, Jeremy Bentham writes that ‘It is with dispositions [i.e. character traits] as with everything else: it will be good or bad according to its effects: according to the effects it has on augmenting or diminishing the happiness of the community. . .’Footnote ⁹ Similarly for John Stuart Mill, ‘The multiplication of happiness is, according to the utilitarian ethics, the object of virtue. . . actions and dispositions are only virtuous because. . . [of the] desirable consequences which [they] tend to produce, and on account of which [they are] held to be virtue[s].’Footnote ¹⁰ Similarly for Driver, as noted above, ‘a virtue is a character trait that produces more good. . . than not systematically. . . Vices on the other hand produce bad states of affairs.’Footnote ¹¹ Thus, benevolence is a virtue for these moral theorists because it facilitates cooperation between individuals and provides help for the needy, which, by and large, leads to more overall good. Malice is a vice because it breaks down social cooperation and, if acted on, usually leads to suffering, thereby promoting bad states of affairs.

However, not only theorists writing within the consequentialist tradition have consequentialist theories of virtue and vice. In her 1978 book Virtues and Vices Philippa Foot characterizes the virtues as states of character, involving choice, that are beneficial to human beings, and vices as states of character, involving choice, that are detrimental.Footnote ¹² Foot does not explain in Virtues and Vices what she means by beneficial and detrimental, except to say that beneficial characteristics are those that humans need for their own sake. In her 2001 book Natural Goodness she goes on to explain that the virtues are needed in the sense that they are required for human goodness.Footnote ¹³ According to Foot, human goodness requires more than survival and reproduction, including the ability to ‘understand stories, join in songs and dances. . . form family ties, friendships and special relationships with neighbors, [etc.]’.Footnote ¹⁴ The virtues are those characteristics or dispositions that make human goodness possible while the vices detract from human goodness. For example compassion, honesty and benevolence are virtues because they strengthen the bonds of community and promote social cooperation which is conducive to human survival and flourishing. Malice, envy and deceitfulness are vices since they weaken the bonds of community and break down the motivation for positive social interaction.Footnote ¹⁵ Foot's theory of virtue and vice is a consequentialist theory since it characterizes the virtues and vices in terms of the consequential properties of these character traits, i.e. according to their ability to promote or detract from human goodness.

III. THE IMPLAUSIBLE INSTABILITY ARGUMENT

My first argument against consequentialist theories of virtue and vice is that they make the virtues and vices implausibly unstable by implying that paradigmatic virtues, such as generosity, compassion and benevolence, could, in the right circumstances, be vices and that paradigmatic vices, such as malice, dishonesty and cruelty could, in the right circumstances, be virtues. But could malice really be a virtue and benevolence a vice? Hasn't something gone horribly wrong if our theory implies that such paradigmatic virtues and vices have such unstable moral statuses? It seems that theories that imply such instability must either mean something very different by virtue and vice then what is normally meant, or else they fail to capture these concepts correctly.

Driver anticipates this sort of objection and offers a few responses. First, she points out that on her account virtues are character traits that promote the good systematically and not necessarily in every instance.Footnote ¹⁶ So if honesty fails to promote the good in a particular situation it would still be a virtue to display this character trait so long as doing so generally promotes the good in that particular population or society. Similarly, if malice generally leads to bad consequences, it is still a vice to be malicious even if malice sometimes leads to good consequences, as it would in cases where a malicious klutz ends up benefiting her victims rather than causing them harm.

However, this response is not completely adequate since it is still possible (given the right society or population) that character traits that we typically think of as virtues might systematically promote the bad rather than the good, and that character traits we typically think of as vices might systematically promote the good rather than the bad. Imagine a world where maliciousness makes human beings on the whole better-off, while benevolence leads to suffering. That is, a world where, for some reason (perhaps because of a Cartesian-like evil demon), whenever we try to harm another person from a desire for their pain for pleasure we make everyone else better-off and thus increase overall happiness. We succeed at causing our victim to suffer but doing so increases overall happiness. We can imagine that a malicious person living in such a world might continue to cause other people suffering even though he is aware that he thereby makes human beings on the whole better-off. But he does not cause his victims suffering to make others better-off, his only aim is to take pleasure in his victim's suffering; he does not care one way or the other about human beings in general. The problem for consequentialist theories of virtue and vice is that, although maliciousness in such a world systematically promotes the good, it does not make the characters of malicious people, such as the one described above, any better, and thus, maliciousness is still a vice. Yet on a consequentialist account of virtue and vice, maliciousness would be a virtue in these worlds because it promotes the good.

Driver recognizes this counterintuitive implication of her theory but she does not think it gives us reason to reject her view. Instead, she believes that an error theory can be given to explain why our intuitions, and not her theory, go astray when evaluating these cases.Footnote ¹⁷ But it is hard to see what this error theory would be like. One possibility would be to suggest, as Driver does in her book, that any plausible theory of virtue and vice must be context-sensitive. For instance, Driver suggests that aggressiveness would be a virtue in a state of nature, e.g. in prehistoric times, although it is a vice in a contemporary society where social cooperation is important.Footnote ¹⁸ According to consequentialists such as Driver, aggressiveness would be a virtue in a state of nature because it would help aggressive individuals get what they want and need in the absence of social cooperation. Aggressiveness is a vice in social settings because it hampers social cooperation, thereby reducing the overall good. Similarly, while honesty is normally considered a virtue, we might rightly regard it as a vice if it generally led to more suffering than good as it might have done in Nazi Germany when being honest meant informing on neighbours and friends who were on the Nazi hit-list.

Thus, an error theory along the lines suggested by Driver might contend that once we are aware of the context-sensitivity of the virtues and vices we will be less likely to think it strange that malice could ever be a virtue or benevolence a vice. The moral status of malice and benevolence depend on the circumstances of the population in which they occur, in particular, on whether these character traits generally lead to good or bad consequences, just as the moral status of aggressiveness and honesty depend on context.

Further, defenders of consequentialist theories might argue that we can see that paradigmatic virtues and vices, such as malice and benevolence, are context-sensitive if we consider how a virtuous person would act in the evil-demon-world discussed above. That is, a virtuous person who realized that she could make many people better-off by being malicious to a few may very well choose to be malicious. Similarly, if benevolence generally led to more suffering overall, a virtuous person may well abstain from benevolence. Thus, it might be argued that since the virtuous person in such a world would choose to be malicious and to abstain from benevolence, then, in that world, malice is a virtue and benevolence a vice.

But this conclusion does not follow from the premises offered. The virtuous person in this world does not choose maliciousness for its own sake, she chooses it as a means to promote the good. Character traits chosen by virtuous people as a means, and not for their own sakes, are not necessarily virtues.Footnote ¹⁹ The situation is best seen as a case of conflicting virtues rather than as a case of character traits turning from virtues into vices and vice versa. In a world where being malicious leads to a significant increase in well-being for others, a virtuous person may very well (but not necessarily) choose to be malicious to some individuals as a means of being benevolent to many others. But that does not mean that malice is no longer a vice. It just means that in certain circumstances, or populations, a virtuous person may choose to be vicious as a means of exhibiting some other virtue.

Similarly a virtuous person living in a world where being benevolent to one person causes suffering to many others may very well choose to abstain from benevolence. In this case she chooses compassion for many people over benevolence for a few.Footnote ²⁰ But that does not mean that benevolence loses its status as a virtue. It just means that when forced to choose between two competing virtues a virtuous person will set the less important virtue to the side to exhibit the other.

The same goes for honesty. If in Nazi Germany being honest meant subjecting friends and neighbours to torture and death, a virtuous person may very well have chosen to be less than truthful out of compassion for others. But that does not mean that honesty was not a virtue in Nazi Germany. It means that honesty was a virtue that was routinely passed over by virtuous people so that they could exhibit competing virtues.

The apparent variable moral status of aggressiveness is different. For one, aggressiveness is not really a character trait, it is a way of responding to the world which is akin to feelings such as fear or confidence when it comes to the virtues and vices. Aggressiveness, in itself, is neither a virtue nor a vice, just as fear and confidence, in and of themselves, are neither virtues nor vices.Footnote ²¹ It is a vice to be fearful or confident when one should not be, or to lack fear or confidence when one should be fearful or confident. Similarly, it is a vice to be aggressive when one should not be, or to lack aggressiveness when one should be aggressive. Just as the virtue with regard to fear and confidence is courage, the virtue with regard to aggressiveness is assertiveness.Footnote ²² While it is a vice to be excessively aggressive or not aggressive enough, it is a virtue to be assertive. To be assertive we may need to be more aggressive in some contexts than in others, but it is not proper to say that aggressiveness, itself, is a virtue in some contexts and a vice in others.

The amount of aggressiveness required for assertiveness is relative to the individual and to the circumstances. For instance, a small person in a callous or hostile environment would need to be more aggressive to be assertive than a big person in a caring or peaceful environment. Other virtues and vices are relative in this respect. For example temperance in food consumption is relative to the size and activity of the individual. However, other virtues and vices such as honesty and malice are not relative in this respect. For instance, to be untruthful to any degree in any context is dishonest. Similarly, desiring someone else's pain for pleasure to any degree and in any context is malicious.Footnote ²³

Hence, there are two reasons why we cannot alleviate worries raised by the implausible instability argument by pointing to the context-sensitivity of the virtues and vices. First, it is not clear that the virtues and vices are context-sensitive in the respect that in different contexts paradigmatic virtues become vices and vice versa. It seems rather that certain virtues such as assertiveness and temperance, while retaining their moral status as virtues and vices, are relative in the respect that what they require of a person is relative to the person and to her circumstances. Second, not all virtues and vices are relative, or context-sensitive, even in this respect; for instance some paradigmatic virtues and vices such as malice and honesty are not. Thus, an error theory that points to the context-sensitivity of the virtues and vices fails. In lieu of an adequate error theory, it is a serious problem for consequentialist theories of virtue and vice that they imply the instability of paradigmatic virtues and vices.

IV. THE INCOMPLETENESS ARGUMENT

According to the incompleteness argument, since consequentialist theories define virtue and vice in terms of the consequential properties of these character traits rather than in terms of their intrinsic properties, the accounts are less complete and satisfying than intrinsic theories. Consequentialist theories are less complete and satisfying because we want to know more than that virtues are states of character that lead to good consequences and that vices are states of character that lead to bad consequences. We want to know, in addition, what these states of character are like. Consequentialist theories, by their very nature, give us no insight into the intrinsic properties of virtue and vice. Instead they rely on pre-theoretic understandings of these character traits. This is less than we could hope for from a theory of virtue and vice. It would be like going to cooking class and being instructed to make foods that people enjoy. The obvious question to ask is ‘what is it about certain foods that people enjoy?’ We want to know about the intrinsic properties of enjoyable dishes; how to prepare them, what spices to use, etc. Similarly with a theory of virtue and vice, we want to know what characteristics character traits have that make them pleasing or good-producing. But consequentialist theories do not answer these questions because they do not tell us about the intrinsic properties of the virtues and vices.

Driver's response to this objection is that her theory is not intended to give details about individual virtues and vices. It is only meant to give a general account ‘leav[ing] room for the analysis of particular virtues in terms of the characteristic psychological states that underlie them’.Footnote ²⁴ However, this response is not wholly adequate. Although it may be unfair to reject outright a theory such as Driver's, because it fails to give insight into individual virtues and vices, it certainly is an advantage if our theory does so. There are two distinct advantages to having a general theory of virtue and vice that gives us insight into the intrinsic properties of individual virtues and vices. First, as we will see in the next section, when our characterization of individual virtues and vices follows from a general theory of virtue and vice, we get insight into the common structure of virtue and vice more generally, and we learn about structural similarities and differences between various virtues and vices. Consequentialist theories tell us nothing about the common intrinsic properties of virtues and vices.

Second, when there is a theoretical connection between a general theory and individual characterizations of objects covered by that theory, the general theory and the individual characterizations are mutually supportive. For instance, we would be more confident about a neurophysiological account of the mind if we could characterize individual mental states neurophysiologically. In the same way, we would be more certain about our characterizations of individual mental states, such as loving our neighbour, if they followed from principles constitutive of our general theory of mind, such as that certain chemicals are necessary for emotion. We would not get this mutual support if we had a neurophysiological theory of mind but characterized individual mental states solely in terms of folk psychology. Similarly, a general theory of virtue and vice gets additional support if it gives us plausible and insightful characterizations of individual virtues and vices, just as our characterizations of individual virtues and vices get additional support if they follow from our general theory. This is an advantage enjoyed by intrinsic theories that is not shared by consequentialist accounts.

V. INTRINSIC THEORIES

Intrinsic accounts of virtue and vice avoid the implausible instability objection and are more complete and satisfying than consequentialist accounts. Consider, for example, Thomas Hurka's recursive account. According to Hurka, virtue consists in attitudes towards good and bad states of affairs that are themselves intrinsically good, while vice consists in attitudes towards good and bad states of affairs that are intrinsically bad.Footnote ²⁵

The account begins with two base clauses which list the base level intrinsic goods and bads. Next, recursion clauses tell us which attitudes towards intrinsic goods and bads are themselves intrinsically good or bad. Hurka's base clauses list pleasure, knowledge and achievement as intrinsically good, and pain, false belief and failure in the pursuit of achievement as intrinsically bad.Footnote ²⁶ Here are some of the recursion clauses he gives:

Combined with his list of base level intrinsic goods and bads, these recursion clauses yield various virtues and vices. For example, we get malice when we replace another person's pain for x in (LE). That is, malice consists in loving another person's pain for its own sake.Footnote ²⁸ Similarly, we get benevolence when we replace another person's good for x in (LG). That is, benevolence consists in loving someone else's good for its own sake.Footnote ²⁹

Since Hurka's account characterizes virtue and vice in terms of the intrinsic properties of these character traits, it is conceptually impossible that changes to the instrumental, or consequential, properties of the virtues and vices (i.e. changes from being instrumentally good or bad to being instrumentally bad or good) could change virtues into vices or vice versa. For instance, malice is a vice because it consists in loving what is intrinsically bad and (LE) says that loving what is intrinsically bad is a vice no matter how instrumentally good it happens to be. Thus, Hurka's account avoids the implausible instability objection.

However, it might be argued that, while avoiding the implausible instability objection, intrinsic accounts, such as Hurka's, fall prey to another objection that consequentialist theories avoid: the implausible rigidity objection.Footnote ³⁰ According to this objection, by characterizing the virtues and vices purely in terms of their intrinsic properties, intrinsic accounts imply that virtues remain virtues even if circumstances change so dramatically that they systematically lead to bad or disastrous consequences. For this reason intrinsic theories of virtue and vice are implausibly rigid. Consider, for example, that in Victorian times chastity was thought to be an important virtue for women since unchaste women would invariably end up with unwanted children, ruining their prospects for good marriages and happy adult lives. However, times have changed, at least in Western democratic countries. Effective forms of birth control make it much less likely that unwanted children will result from extramarital sex, and marriage and future happiness are no longing incompatible with single parenthood. Thus, chastity is no longer considered a significant virtue, if it is thought be a virtue at all. According to the implausible rigidity objection, since intrinsic accounts, such as Hurka's, imply that virtues remain virtues no matter what the consequences, these theories cannot account for the varying moral status of virtues such as chastity.

But this is not quite correct. Intrinsic accounts, such as Hurka's, can account for the varying moral status of chastity. They can do so by pointing out that while in Victorian times chastity was a form of loving one's own good and the good of those who would be adversely affected by an extramarital pregnancy, this is no longer the case. Women can now show concern for their good, and for the good of others, without avoiding extramarital sex. Thus, given the change in social circumstances, chastity is no longer a form of loving the good, and so it is no longer a virtue on Hurka's intrinsic account. Chastity is what we might call a contingent virtue since whether it is a form of loving the good depends on the social context.

Similarly, what counts as benevolence depends on the circumstances in which we find ourselves. For instance, it might be benevolent to encourage lesbians to ‘come out’ within liberal communities since being honest and open about their sexual orientation might have a liberating effect on their lives. It would not be similarly benevolent to encourage lesbians to ‘come out’ in repressive communities where lesbians are persecuted. Unlike chastity, benevolence is a non-contingent virtue since it is always a virtue to love other people's good. It is just that what counts as loving other people's good is different in different contexts. Thus, intrinsic accounts, such as Hurka's, can explain how changes in circumstances can change the status of contingent virtues such as chastity and how non-contingent virtues such as benevolence require us to act differently in different contexts.

However, proponents of the implausible rigidity objection might find other cases more troubling for intrinsic theories. Imagine a society where people are generally so hostile to each other that the only way to avoid being sabotaged is to lie about one's true purposes, methods and motivations. It might be argued that such a society is no society at all, but rather, a state of nature. But I think we can imagine that, with the help of a system of laws and a police force, the inhabitants of such a population would be able to cooperate for their mutual benefit to some extent, provided they are not too truthful. Proponents of the implausible rigidity objection contend that the problem with intrinsic theories is that they imply that truthfulness would be a virtue in such a population even though it would be disastrous for those who had the trait and of no overall benefit. For instance, for Hurka, truthfulness is a virtue by (LE), i.e. it is an instance of hating the bad, because it ‘involves a desire that others not believe or come to believe what is false’.Footnote ³¹ Since by (LE) hating what is bad is a virtue no matter what the consequences, truthfulness would be a virtue in the population discussed above even though it would be disastrous for those who had the trait and of no overall benefit. According to the implausible rigidity objection it is implausible to maintain that truthfulness would be a virtue in such a population. But is it really implausible to believe that truthfulness remains a virtue even when it leads to disastrous results? Again, I think the right thing to say is that truthfulness is a virtue even when it leads to disastrous results, but that a virtuous person would refrain from exhibiting this virtue in order to exhibit other, morally more important, virtues such as prudence. Truthfulness remains a virtue in populations such as the one discussed above because it is intrinsically good to have this sort of attitude toward false beliefs, even if it is so instrumentally bad that we ought to be deceptive. We can see that truthfulness maintains its intrinsic value in circumstances where we ought to be deceptive because being truthful would make our characters better in such circumstances if it did not conflict with other more pressing virtues such as prudence.

The recursive account also avoids the incompleteness objection by describing in some detail the intrinsic properties of virtue and vice rather than simply describing them as character traits productive of good or bad states of affairs. For instance, the recursive account describes virtue as an attitude towards good or bad states of affairs. The relevant attitudes are further described by the recursion clauses such as loving the good, hating the bad, loving proportionately, etc. Further, Hurka provides insight into the nature of individual virtues and vices in terms of the particular attitudes in which they consist. For instance, by the proportionality principle we learn that selfishness is a vice that involves loving one's own good more than its intrinsic value warrants when compared to other peoples’ good.Footnote ³² In addition, we see that different virtues and vices have similar structures. For instance, like selfishness, cowardice is a vice by the proportionality principle that involves loving one's own comfort and safety more than its intrinsic value warrants when compared to some greater good that could be achieved by risking them.Footnote ³³ Thus Hurka's intrinsic account gives us much more insight into the nature of virtue and vice than do consequentialist accounts.

VI. MIXED VIEWS

Not all theories of virtue and vice can be neatly classified as either consequentialist or intrinsic. Driver suggests that Aristotle and Rosalind Hursthouse have mixed views, having both consequentialist and intrinsic requirements for virtue and vice.Footnote ³⁴ I agree that Hursthouse's neo-Aristotelian view should be interpreted as a mixed view but I question whether Aristotle's theory is a mixed account. According to Aristotle ‘virtue. . . is a characteristic involving choice. . . [that] consists in observing the mean relative to us, a mean which is defined by a rational principle, such as a man of practical wisdom would use to determine it.’Footnote ³⁵ Aristotle's theory of virtue and vice is often called the doctrine of the mean since it defines virtue as a mean between vices of excess and defect. For instance, acting virtuously with regard to the giving and taking of money, i.e. generously, consists in observing a mean between giving money away too freely, which is the vice of extravagance, and not being free enough with one's money, which is the vice of stinginess.Footnote ³⁶ Similarly, courage is a mean with regard to feelings of fear and confidence. The courageous person feels fear and confidence to just the right amount given the circumstances. The vice of recklessness consists in an excess of confidence while cowardice consists in having too little confidence and an excess of fear.Footnote ³⁷

The property of observing a mean between excess and defect is not a consequential or instrumental property of a character trait since it does not consist in bringing about some other object or state of affairs. It is an intrinsic property. Aristotle also puts further intrinsic conditions on virtue, in particular, that virtuous acts must be done for their own sakes and that they must follow from practical wisdom.Footnote ³⁸ Thus, it seems that Aristotle's account characterizes virtue and vice in terms of the intrinsic properties of these character traits and not in terms of their consequential properties.

However, Driver believes that Aristotle's account is a mixed view since in addition to the intrinsic conditions he puts on virtue, he also seems to say ‘that for a trait to really be a virtue, it must somehow contribute to human flourishing’.Footnote ³⁹ There are two reasons to think that Aristotle makes contributing to human flourishing a precondition for a trait being a virtue. First, he argues that ‘the good of man [which Aristotle equates with flourishing or eudaimonia] is an activity of the soul in conformity with excellence or virtue.’Footnote ⁴⁰ This seems to say that for a trait to be a virtue it must be good for the individual in the sense that it contributes to her flourishing. Second, Aristotle states that ‘the man who does not enjoy performing noble actions is not a good man at all. . . actions performed in conformity with virtue are in themselves pleasant.’Footnote ⁴¹ Driver interprets this to mean that character traits are virtues only if they contribute to personal flourishing.Footnote ⁴²

It is beyond the scope of this article to give a fully adequate defence of the view that Aristotle's theory is an intrinsic theory rather than a mixed view.Footnote ⁴³ However, I will sketch briefly why I think his theory is best interpreted as an intrinsic theory. I will do so by showing that the two quotes mentioned above do not really show that Aristotle's view is partly consequentialist. First, when Aristotle says that ‘the good of man is an activity of the soul in conformity with excellence or virtue’ he seems to me to be making a conceptual claim rather than a causal claim. In other words, Aristotle seems to be offering a moralized conception of flourishing in which virtue is, at least partly, constitutive of flourishing rather than a cause of flourishing.Footnote ⁴⁴ But Aristotle's theory is only a mixed view if we interpret him as making the causal claim that virtues lead to, or cause, flourishing, since mixed views make the promotion of flourishing partly constitutive of virtue. Thus, if we interpret Aristotle as having a moralizing conception of flourishing, as I have suggested, then the claim that flourishing is an activity of the soul in conformity with virtue does not support a mixed view since then it would not say that flourishing is an effect of virtue but that flourishing and virtue are in some sense the same thing.

It might be objected that virtue cannot be fully constitutive of flourishing on Aristotle's account since he says quite explicitly that we cannot truly flourish unless we also have some external goods such as friends, wealth and good children.Footnote ⁴⁵ But even if virtue is only partly constitutive of flourishing it might still be that the connection between virtue and flourishing is conceptual and not causal: that virtuous character traits do not cause flourishing, but rather that part of what we mean by flourishing is having virtuous character traits.

Second, it does not follow from Aristotle's claim that ‘actions performed in conformity with virtue are in themselves pleasant’ that, for Aristotle, part of what makes a character trait a virtue is that it causes pleasure. As I interpret Aristotle, his claim is that it is an empirical fact about human beings that they get pleasure from what they love: ‘a lover of horses [gets pleasure] from horses, a lover of theatre [gets pleasure] from plays and in the same way a lover of justice [gets pleasure] from just acts, and a lover of virtue in general [gets pleasure] from virtuous acts.’Footnote ⁴⁶ Aristotle believes that this is just the way the human souls work. Thus, for Aristotle, it is an empirical fact about human beings that if we do not get pleasure from acting virtuously then we do not really love virtue, and thus, we do not want virtue for its own sake which is one of Aristotle's intrinsic conditions for virtue. It is for this reason that ‘[w]e may even go so far as to state that the man who does not enjoy performing noble actions is not a good man at all.’Footnote ⁴⁷ But for Aristotle to have a partly consequentialist theory of virtue, causing pleasure must be part of what makes a character trait a virtue and not merely a natural accompaniment for human beings given empirical facts about our nature. Consider the following analogy. Imagine that it is an empirical fact that male human beings cannot be happy unless married. It would follow from this that all bachelors are unhappy. But would that mean that part of what it is to be a bachelor is to be unhappy; that unhappiness is part of the very concept of bachelor? No, it would be a contingent empirical fact about human beings that bachelors are unhappy. The concept of bachelor would still be fully captured by the phrase ‘unmarried man’. Similarly, for Aristotle the concept of virtue is fully captured by the phrase: is a characteristic to choose, for its own sake, a mean between excess and defect using rational principles prescribed by practical wisdom.Footnote ⁴⁸ By saying that virtuous acts are pleasant in themselves Aristotle merely points out that, for human beings, it is an empirical fact that if someone does not take pleasure in performing a virtuous act then the act was not done for its own sake, and thus, it was not truly virtuous. At least that is how I interpret him.

Whether Aristotle's own theory is a mixed view, Hursthouse's neo-Aristotelian theory does seem to be partly consequentialist and partly intrinsic. She writes: ‘A virtue is a character trait a human being needs for eudaimonia, to flourish or to live well.’Footnote ⁴⁹ Unlike Aristotle, Hursthouse does not have a moralized conception of flourishing because she does not think that virtue is conceptually necessary for flourishing. She does not think that virtue is conceptually necessary for flourishing because she believes that it is possible for some wicked people, such as Nazis who escaped to South America after the Second World War, to flourish.Footnote ⁵⁰ Wicked people can flourish, on Hursthouse's account, since it is sufficient for flourishing that our lives have what she calls the ‘smile factor’ – the quality of being enjoyable from our own point of view.Footnote ⁵¹ So for Hursthouse, virtue is not conceptually necessary for, or constitutive of, flourishing, it is just the most reliable way to attain it. It is this ability to contribute reliably to flourishing that makes a trait a virtue.Footnote ⁵² Thus, Hursthouse's account is partly consequentialist since she makes a trait's ability to contribute reliably to flourishing a condition of it being a virtue.

However, Hursthouse's view is not entirely consequentialist since she also says that to be virtuous an agent must know what she is doing, act for the right reason, and have the appropriate feelings and attitudes when she acts.Footnote ⁵³ These are intrinsic conditions for virtue and so Hursthouse's theory is best viewed as a mixed account.

Driver points out that ‘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 ⁵⁴ However, Driver thinks that upon reflection mixed views are less tenable than either consequentialist or intrinsic views since they inherit the problems of both of these sorts of theories.Footnote ⁵⁵ I have a more positive assessment of mixed views: they avoid the objection raised by the incompleteness argument and fare better against the implausible instability argument than do consequentialist theories, but not as well as intrinsic theories.

Mixed accounts avoid the objection raised by the incompleteness argument by characterizing virtue and vice partly in terms of the intrinsic properties of these character traits. For instance, even if we interpret Aristotle's theory as partly consequentialist and partly intrinsic it still tells us something about the intrinsic properties of virtue and vice and gives us some insight into the nature of individual virtues and vices. The doctrine of the mean tells us that virtue consists in observing a mean between excess and defect while vice consists in acting or feeling immoderately. If correct, this is an interesting insight into the nature of virtue and vice. Furthermore, the doctrine of the mean tells us about individual virtues and vices, for example, that generosity consists in observing the mean with regard to the giving and taking of money, whereas extravagance and stinginess consist in acting excessively or defectively.

Since Hursthouse's neo-Aristotelian account does not include the doctrine of the mean, it is less illuminating than Aristotle's. She tells us that virtue involves acting for the right reasons, in the right manner, while having appropriate feelings or attitudes.Footnote ⁵⁶ But she does not say exactly what this involves. At one point she says that acting for the right reason can involve a range of reasons depending on the virtue in question. For example ‘the sorts of reasons a courageous agent might have for performing a courageous act. . . [might include] “I could probably save him if I climbed up there”, “Someone had to volunteer”, “One can't give into tyrants”, “It's worth the risk”.’Footnote ⁵⁷ For Hursthouse, when a normal adult, in ordinary circumstances, acts for one of these reasons she acts for the right reason because she shows that she is ‘“really committed” to the value of her [courageous] acts. . . [that is] that the valuing of them goes “deep down”, we may say; that it governs and informs her whole life and conduct.’Footnote ⁵⁸ Hursthouse admits that this characterization of right reasons is not very perspicuous and I agree. She gives even less guidance about what it is to act in the right manner or from appropriate feelings or attitudes. At one point she says that honest people must do honest acts ‘readily, eagerly, unhesitatingly, [and] scrupulously’.Footnote ⁵⁹ But this is not meant to be a list of necessary or sufficient conditions for honesty nor does she offer similar lists for other virtues or provide a theoretical basis for the list she offers. The fundamental problem with Hursthouse's intrinsic conditions for virtue is that she does not offer a general scheme or principle that tells us what they require of us, both in the abstract and in particular cases. Thus, Hursthouse's theory of virtue and vice could be more complete and illuminating, but it is not in virtue of being a mixed view that it is deficient in this respect. At least it tells us something about the intrinsic properties of virtue and vice.

While mixed views, as a class, avoid the objection raised by the incompleteness argument, they do not avoid the objection raised by the implausibly instability argument. This is because, like consequentialist theories, mixed theories make the production of good consequences a necessary condition for virtue. For instance, Hursthouse's account of virtue states that, in addition to certain intrinsic conditions, the virtues must be ‘[needed for] eudaimonia, to flourish or to live well’.Footnote ⁶⁰ Thus benevolence and honesty would not be virtues in an evil-demon-world where they generally detracted from flourishing. But surely our characters would still be made better in such a world if we were able to exhibit these character traits without thereby neglecting other more pressing virtues such as compassion for a greater number. This counterfactual is true even though, by hypothesis, we cannot exhibit benevolence and honesty in these worlds without neglecting more pressing virtues. Yet as long as the counterfactual claim is true, benevolence and honesty, in themselves, are virtues, since by themselves they improve our characters. Thus, it is a conceptual failing of mixed views that they imply that virtues such as benevolence and honesty lose their status as virtues if circumstances are such that they generally detract from flourishing, or more generally, from the good.Footnote ⁶¹

Yet at least mixed views do not imply that paradigmatic virtues such as benevolence turn into vices in circumstances, or populations, where they generally detract from the good. Mixed views avoid this implication because to be a vice a character trait must not only detract from the good, it must also meet certain intrinsic conditions for vice as specified by the theory. Unfortunately Hursthouse does not specify the intrinsic conditions for vice directly, but it is reasonable to suppose that they would be something like acting for the wrong reasons, in the wrong manner, while having inappropriate feelings or attitudes.

The same arguments apply to Aristotle's theory of virtue and vice if we interpret it as a mixed view. However, if instead we interpret Aristotle's theory as an intrinsic account, as I have suggested, then benevolence remains a virtue no matter what its consequences, and it remains partly constitutive of flourishing and pleasurable to virtuous people who exhibit this character trait even if it generally leads to suffering. That said, it is unlikely that a virtuous person would exhibit this character trait in a world where it generally leads to suffering since, once again, doing so would require her to neglect other more pressing virtues such as compassion for the greater number. Since a truly virtuous person, on Aristotle's account, would act from practical wisdom she would not choose to be benevolent to a few in a world where doing so generally leads to suffering for the many since it is, quite plausibly, morally more important to be compassionate to the many. But that would not mean that for Aristotle benevolence is not a virtue in such a world. If we interpret his theory as an intrinsic view, it would mean that benevolence is a virtue that is typically passed over in these worlds so that competing virtues can be exhibited.