I. INTRODUCTION
A number of ancient philosophers thought that being virtuous would necessarily and directly enhance your welfare.Footnote 1 Indeed some ancients went so far as to say that being virtuous would guarantee for you an on-balance welfare-good life. The moderns (especially those in the utilitarian tradition)Footnote 2 have rejected each of these theses, citing a range of cases where the alleged links between virtue and welfare break down.Footnote 3 They insist that any connections to be had are merely contingent and purely instrumental. Here, a person's virtue can pay off for her only if she has some personal interest in being moral.
But for all its intuitive support, the modern view is not entirely satisfactory either: as many have pointed out,Footnote 4 it contradicts what our own behaviors and fundamental life practicesFootnote 5 seem to indicate, namely, that being moral is a central and indispensable part of any life worth living.
In light of the difficulties facing each of these approaches, it is natural to suppose that the most plausible position regarding the virtue/welfare link can be found somewhere between the ancient and the modern. In a recent article Wayne Sumner defends one such view.Footnote 6 He argues that virtue is indeed intrinsically, prudentially valuable (at least if we assume that his happiness theory of welfare is true), but that its impact on the total welfare value of a life is rather modest: as a matter of necessity, virtue does make a life go better; but it neither guarantees an on-balance welfare-good life, nor is it required in order to live one. While I find Sumner's dialectical move alluring, I am not persuaded by the particular argument he offers in support of this thesis. Nor am I convinced of a second thesis he intends, in part, to derive from it, namely, that, necessarily, a virtuous person is intrinsically, prudentially better-off than his non-virtuous counterparts (all else being equal). In this article, I try to say precisely where and how I think Sumner's arguments for these theses fail.
I will proceed as follows. In section II, I present interpretations of Sumner's arguments for his first and second thesis. In section III, I offer several criticisms of the argument for the first thesis, each of which makes the same basic point: certain contingent facts over and above a person's being virtuous have to obtain if virtue is to issue in any prudential reward. If my criticism is sound, then Sumner's first thesis turns out to be indistinguishable from the moderns'. In section IV, I raise an objection to Sumner's second thesis by arguing that moral neutrality, i.e. the state of being neither virtuous nor vicious, can in some cases be at least as welfare-enhancing as moral virtuosity. In the fifth and final section, I present three different cases that attempt to illustrate that even if virtue is intrinsically rewarding, the potential external effects it may have on a person's welfare make it impossible to determine a priori whether adopting a virtuous lifestyle would be prudentially rational for that person. These cases seem to undermine virtue's distinctive prudential appeal.
II. SUMNER'S ARGUMENTS
A. Clarification of Sumner's First Thesis
Sumner states one of his goals to be
to show that virtue is one prudential good among others, and thus that its presence in a life (any life) to that extent makes that life go better, not for any instrumental reason (such as external rewards, in this life or the next, accruing to the righteous) but intrinsically, in itself.Footnote 7
Though this passage suggests that Sumner is going to argue for a claim about all human persons, later passages indicate that he wants (1) to exclude from consideration any individual ‘whose pathology is precisely the absence of empathy or fellow-feeling’Footnote 8 and (2) to restrict his claim to those ‘whose sociality has not been corrupted by ignorance or indoctrination’.Footnote 9 Let us call the class of socialized persons who satisfy these conditions normal. We can use this terminology to state officially what I understand to be Sumner's first thesis:
(ST1) Necessarily, for any normal human person, P, virtue is welfare-enhancing for P.
Before moving on to the argument for ST1, a couple of clarifying remarks are in order. First, I am presenting ST1 as expressing a de dicto necessary proposition. Formulating the thesis in this way acknowledges that ST1 is intended to be ‘more than a contingent fact’Footnote 10 and that it is to preserve an ‘internal’ link between virtue and welfare.Footnote 11 Second, the kind of virtue whose presence in a life is alleged to make that life intrinsically, prudentially better is other-regarding. On its standard “Humean” analysis, and on the analysis Sumner seems to prefer, a virtue is other-regarding iff its exercise generally promotes the welfare of others.Footnote 12 Some paradigm examples include justice, benevolence and fidelity.Footnote 13 For the remainder of this article, by ‘being virtuous’ I shall mean ‘being other-regardingly virtuous’.
B. The argument for ST1
Sumner begins his argument by noting that ‘our own moral decency is something that seems to matter deeply to most of us’,Footnote 14 at least for those of us who are socially normal in the sense I have introduced above. The source of our concern for our own decency can be found in its relation to social feelings like empathy and sympathy. If these social feelings have been properly cultivated, we will inevitably care about our own moral worth – nature has simply designed us in this way.Footnote 15
But caring about our own decency is only one among a number of our natural tendencies. According to Sumner, we are inclined also to think well of ourselves in some overall sense, or to have what he calls self-worth.Footnote 16 Sumner characterizes self-worth as a psychological condition that requires for its sustenance the positive appraisal of one's own value, in particular, one's own moral value:
Precisely because assurance of our own decency is such an important component of our sense of self-worth, we all conceptualize our lives under the form of virtue: we all want to believe that we are good and decent folk, however wildly wide of the mark this belief might be.Footnote 17
So while the social feelings of empathy and sympathy prompt us to care about our own decency, our sense of self-worth prompts us to evaluate it both routinely and favorably. If we add to these considerations a modal qualification strong enough to help yield ST1, we can formulate what I take to be the first premise of Sumner's argument.
(1) Necessarily, for any normal human person, P, P cares about and frequently, positively evaluates her own decency.
The second step of the argument for ST1 tries to show how Sumner's own theory of welfare can afford the virtuous some special prudential reward. To make this line of reasoning clear, we need first to consider some of the details of Sumner's happiness theory of welfare.
For Sumner, happiness is a matter of being satisfied with one's life. Satisfaction is a complex psychological state comprised of both a cognitive and an affective component: ‘[The] cognitive aspect consists in a positive evaluation of the conditions of one's life, a judgment that, at least on balance, it is measuring up favorably against one's standards or expectations for it’.Footnote 18 Of course, the conditions that we are directly concerned with in this article are those pertaining to a person's moral decency. We can say that a person is cognitively satisfied with his moral decency provided that he judges himself to be “good enough” relative to his own moral standards. Regarding the affective component of satisfaction, Sumner maintains that it ‘consists in what we commonly call a sense of well-being: finding one's life (or some aspect of it) enriching or rewarding, or feeling satisfied or fulfilled by it’.Footnote 19 This phenomenology is familiar enough; for our purposes, we needn't say more about it.
Sumner uses this notion of satisfaction to state the relevant portion of his theory of welfare: a condition of someone's life counts as an intrinsic source of well-being for her if and only if she authentically endorses it, or finds it satisfying for its own sake.Footnote 20 To say that a person is authentically satisfied with a condition of her life is to say that her satisfaction is both informed and autonomous. A person's satisfaction is informed when it is either (a) based entirely on true empirical information, or (b) based on false (or missing) empirical information but would nonetheless survive confrontation with the truth.Footnote 21 A person's satisfaction is autonomous when it expresses values that have not been ‘externally manipulated’ and preserve the capacity for critical reflection.Footnote 22 The information condition is meant to ensure that a person's welfare be determined in part by an accurate or at least personally acceptable picture of her life. The autonomy condition is meant to ensure that a person's welfare is determined in part by a set of values that are her own.
With Sumner's theory in hand, we are in a position to see why he thinks that virtue is welfare-enhancing. For the virtuous person, ‘the conviction that one has tried to be a good and decent person can be a source of great fulfillment’.Footnote 23 And since the fulfillment (or satisfaction) of the virtuous person must be authentic, it will necessarily enhance her welfare.Footnote 24 We can help ourselves to a hypothetical example to make this reasoning more clear. Imagine a virtuous person, V, who cares about and frequently, positively evaluates her own decency. Consider an arbitrary evaluation. V will be cognitively and affectively satisfied by what she finds, since, by hypothesis, she is a socially normal person inclined to favorably evaluate her own moral worth. Since V is virtuous, her satisfaction will be both informed and autonomous. It will be informed because it will be based on true empirical information. When she goes to measure how well she is living up to her moral standards, she will appeal to a set of facts concerning her own past behavior; these facts will indicate to her that she has in fact behaved virtuously. What is more, V's satisfaction will be autonomous; it will be based in part on a set of values fit for a virtuous person.Footnote 25 Since V's satisfaction will be both informed and autonomous, it will be authentic; and since authentic satisfaction makes a life go better, V's virtue will intrinsically pay off for her.
Of course, we can generalize from V's condition since we didn't derive our result from any special features of her case: any person with a virtuous disposition will enjoy authentic satisfaction with respect to her own decency. So for any normal person who is both concerned with her own moral decency and capable of being virtuous, virtue will have the power to enhance that person's welfare. If we condense this line of reasoning and add to it the appropriate modal qualification, we have what I understand to be the second premise of Sumner's argument.
(2) Necessarily, if P is virtuous, then P's virtue is welfare-enhancing for P.
It should be clear that (1) and (2) together entail ST1. If Sumner is right, it turns out that many of us have selfish reasons for being moral.
C. Clarification of Sumner's Second Thesis
Sumner asks, ‘Other things being equal, are nice people happier than nasty ones?’Footnote 26 The conclusion of the last section does not, by itself, guarantee a positive answer to this question; for it might be the case that an argument similar in certain respects to the one given above can be offered for the conclusion that the moral self-satisfaction of a deluded, non-virtuous person is also intrinsically, prudentially rewarding. If such an argument were available, then it wouldn't necessarily be the case that the virtuous fare better than the non-virtuous (other things being equal). Since Sumner answers his own question affirmatively, his task is to show how moral self-satisfaction can pay off only if ‘it is based on genuine moral merit and not its counterfeit’ [my emphasis].Footnote 27 In conjunction with ST1, this last claim would secure the truth of what I take to be Sumner's second thesis:
(ST2) Necessarily, for any virtuous person, V, V is intrinsically, prudentially better-off than her non-virtuous counterpart(s) – other things being equal.
Like ST1, ST2 purports to express a de dicto necessary truth. Any weaker interpretation of the thesis would make it indistinguishable from the moderns'.
D. The argument for ST2
As with the argument for ST1, I will elaborate on Sumner's reasoning and help myself to a hypothetical example in order to draw out the subtleties of his argument.
Since ST2 employs a ceteris paribus clause, we need to set up the proper testing conditions. First, we need to consider two situations, one in which a person is virtuous and one in which a person is not virtuous. Second, we need to hold a couple of relevant factors fixed in these situations and compare the impact each person's satisfaction has on his own welfare. To this end, let's consider a time t, and an arbitrary pair of individuals, A and B, such that A is virtuous at t and B is not virtuous at t. Further let's suppose that, at t, A and B are alike in each of these respects: (a) they enjoy equal amounts of total welfare; and (b) each sincerely cares about and is eager to evaluate his own moral decency. Now consider a later time, t1, such that at t1, A and B evaluate their own decency and are satisfied by what they find. Assuming that A's satisfaction is authentic for the reasons we discussed in a previous section, we ask whether B's satisfaction is also authentic. If it is not authentic, then ST2 will be vindicated, since, at t1, A will be better-off than B.
Let's consider, then, B at t1. Since B is satisfied by her moral decency, she must believe that she is living up to her moral standards. But since she is not virtuous, she has to be mistaken about her own moral decency – hence she must be mistaken about some of the conditions of her life that bear directly on the quality of her character. Perhaps she has false beliefs about the extent to which she has benefited those in need, or perhaps she is culpably unaware of some of the harmful consequences of her past actions.Footnote 28 Whatever the case may be, since, by hypothesis, B wants to be decent, and she in fact is not decent, her satisfaction would not survive confrontation with the truth; so B must fail the information condition of Sumner's theory of welfare. Thus, B's satisfaction, unlike A's, does not increase her welfare. This gives A the prudential edge. Since this conclusion has been reached on the basis of very general considerations that would apply equally well to any pair of individuals in A and B's situation, we may infer ST2: other things being equal, the virtuous will be better-off than the non-virtuous.
III. CONTINGENT REWARDS
In the following subsections I try to make the case that (pace Sumner) merely being virtuous is not sufficient for welfare-enhancement; a number of other contingent facts have to obtain if virtue is to issue in any prudential reward.
A. Criticism of premise (1) of the argument for ST1
The first contingent fact that has to obtain is this: each person who is normal in the current sense has to care about and frequently, positively evaluate his own decency. I don't think that being normal requires anything quite this strong. Recall what the concept of normalcy was introduced to do: restrict Sumner's generalization to those for whom virtue is not entirely lost, namely, those who are equipped – at least to some degree – with the pro-social feelings of empathy and sympathy. I think that a bit more reflection on the nature of these emotions reveals why premise (1) is false.
Empathy and sympathy are essentially “other-directed” – as opposed to “self-directed” – dispositions; being empathetic is a matter of being disposed to identify with another person's emotional condition; being sympathetic is a matter of being disposed to share in or directly experience for oneself another person's feelings. I don't see why a person who possesses these other-directed dispositions must also possess the self-directed disposition to monitor his own moral worth. On the face of it, it seems that these dispositions are entirely distinct. We can bring this conceptual point out more clearly by considering an example. Imagine a non-self-monitoring person who out of sheer convenience takes a job working nights at a children's shelter. Suppose that over time, this person starts to cover various day shifts. Suppose that as she does this, she begins to work with the children more closely and, as a result, becomes more actively engaged in their lives. Unreflectively, we might suppose, she begins to develop a genuine concern for the children's welfare. As time goes on she remains so busy handling the affairs of the shelter that she can reserve little time for anything other than her work, let alone time for self-reflection. Thus, although she never develops a disposition to monitor her own decency, she remains emotionally available to every child in her shelter; she feels their sorrows along with them and she is appropriately moved by the painful experiences they share with her. Given the way I described the social worker's emotional evolution, it seems clear that she has – at least to some degree – the normal-making dispositions of empathy and sympathy but does not have the self-monitoring disposition that Sumner's premise insists she must. Since there seems to be nothing incoherent or inconsistent about the depiction of this case, I think it provides good reason for thinking that being normal in the relevant sense does not entail caring about and frequently, positively evaluating one's own decency.
One might be tempted to counter by saying that premise (1) might still be true as a matter of nomological necessity, citing the fact that the case just presented seems highly unlikely for any real-world person. But if we were to weaken the first premise so that it purports to express a proposition true in all the worlds with laws like ours, we could only derive from it a conclusion of equal or lesser modal strength. In either case, the revised conclusion would hold merely contingently. I doubt this is something Sumner would want to do, given that he intends to distinguish his thesis from the moderns'.
B. Criticism of premise (2) of the argument for ST1
Another contingent fact that has to obtain if virtue is to issue in any prudential reward is this: a virtuous person has to be satisfied – as opposed to dissatisfied – with his own decency.
Imagine a virtuous person who cares about and frequently evaluates his own decency. Suppose this virtuous person has unusually high standards for being a morally good person. Suppose these standards are so high that he consistently judges himself to fall short of them despite the fact that he really is quite benevolent, compassionate and just. On each occasion of evaluation, this person believes that he should be giving a little more to charity than he has been giving, that he should be acting a little more kindly than he has been acting, and that he should be managing his business a little more fairly than he has been managing it. We might imagine that each of these judgments frustrates him and leaves him dissatisfied with his own moral value, thereby preventing his welfare from being enhanced by his virtue. Of course, it will not do to say in response that this person must be at least somewhat satisfied by his virtue even if he does not completely live up to his unusually high standards; for we can imagine that this person has an all-or-nothing kind of attitude with respect to his own virtue – to his eccentric mind, anything less than perfect virtue is as worthless and morally reprehensible as vice.
A similar but perhaps less contentious example might be given by the following. Imagine a devoutly religious person who sincerely desires to be benevolent and just. Suppose that by any reasonable standards, this person is benevolent and just – she gives quite a bit of her time and money away to worthy and charitable causes. Suppose that when the annual church appeal comes around and she is asked to stretch her means even further than she normally stretches them, she finds herself unwilling to give the amount of money she believes that she ought to give. Suppose that later in the week after reflecting on her perceived weakness she feels like a total moral failure and experiences a tremendous bout of guilt. When she considers her own decency, she is quite dissatisfied by it. I take it that this dissatisfaction is entirely compatible with her being virtuous in the other-regarding sense.Footnote 29
A final problem for the second premise of the argument for ST1 surfaces when we consider the affective fulfillment requirement of Sumner's happiness theory of welfare. To undergo an increase in welfare, a person must not only cognitively endorse the conditions of her life; she must also be made emotionally satisfied by her endorsement of those conditions. But this requirement is too strong: a person could have the other-regarding virtues without ever experiencing deep fulfillment. Imagine a virtuous person who, by nature, has a depressive or gloomy temperament and who cares about and frequently evaluates her own decency. She might be cognitively satisfied with her moral development, but she needn't be affectively fulfilled by it as well. Her depressive condition might put this kind of emotional experience out of her reach.
I think the considerations of this section show that virtue, by itself, is not necessarily welfare enhancing. If it were, every virtuous person would have to (a) know that he is virtuous; (b) have a reasonable set of personal moral standards that permit him to take satisfaction in his own decency; and (c) have the psychological capacity for experiencing affective fulfillment. I contend that merely possessing the other-regarding virtues does not require that (a)–(c) be true.Footnote 30
IV. THE MORAL NEUTRALITY OBJECTION
Despite Sumner's clever argument for ST2, I think it is false that a virtuous person is always better-off than her non-virtuous counterpart (other things being equal). To see why, let's consider a kind of person that Sumner never discusses, namely, a morally neutral person, i.e. one who is neither virtuous nor vicious. How might he fare compared to his virtuous counterpart? It seems clear to me that a morally neutral person could enjoy the same level of welfare as his virtuous counterpart; he need only (i) be satisfied by his decency to the same degree as his virtuous counterpart; (ii) not base his satisfaction on any false empirical information whose correction would undermine his satisfaction; and (iii) possess a set of autonomous values. I think a morally neutral person could satisfy all three of these conditions. To illustrate, let's imagine a morally neutral person who doesn't think too hard about the extent to which his decency matches up to his personal moral standards and who is not concerned all that much about being morally virtuous (though of course, since he does not want to be evil, he is concerned about his moral decency at least to some degree). Imagine that when this person goes to evaluate his own decency, he does not base his judgment on any substantial review of the relevant facts; instead, this person makes a relatively automatic and superficial judgment to the effect that he is a morally good person. Though he is mistaken in this regard, merely being wrong is not enough to make his satisfaction inauthentic; the information condition of Sumner's theory was – as Sumner puts it – ‘formulated to handle problems of empirical, not moral, mistake’.Footnote 31 Given that this person has no significant interest in being moral, we can assume that his satisfaction wouldn't be withdrawn if he were informed of his moral neutrality. Accordingly, this person satisfies the information requirement of Sumner's theory of welfare. And since he is not morally vicious, it would be implausible to say that he must therefore have non-autonomous values. So it seems reasonable also to assume that he satisfies the autonomy requirement of Sumner's theory of welfare. If I have not overlooked anything important here, and this kind of case is in fact logically possible, then moral neutrality can be at least as welfare-enhancing as moral virtue.
Another problem for ST2 arises when we consider again dissatisfied virtuous agents. If we compare the welfare level of a dissatisfied virtuous person to the welfare level of his equally dissatisfied morally neutral counterpart (holding other things equal), we won't find the difference that ST2 predicts we should since neither person is made better-off by his dissatisfaction. This kind of case seems perfectly common, and it provides additional reason for thinking that even if other things are equal, those with virtue do not necessarily fare better than those without it.
V. THE APPEAL OF VIRTUE
At the end of his paper, Sumner conjectures that if he is right about the intrinsic rewards of virtue, then it should have appeal for anyone whose sociality has not been corrupted.Footnote 32 In one sense of ‘appeal' – and in the sense he most likely intended – Sumner is surely right: if virtue is intrinsically rewarding, then we all have some prudential reason for pursuing it. But in the motivational sense of the word, I think virtue's appeal is unfortunately lost; for in many cases, adopting (or maintaining) a virtuous lifestyle and enjoying its benefits would come at too high a price. This can make virtue an irrational choice for such individuals. Consider the following three kinds of cases to demonstrate this point.
Case A: Imagine a virtuous person who would gain as much authentic moral self-satisfaction from being neutral as he currently gains from being virtuous, but would also see a net gain in welfare if he were to make the transition to neutrality. This might happen if this person were to trade in his job as an underpaid hospice worker for a less demanding one that allowed him much more time and money for recreation. For this person, virtue might not be prudentially appealing.
Case B: Imagine a morally neutral person who would gain as much authentic moral self-satisfaction from being virtuous as he currently gains from being neutral, but would suffer a net loss in welfare by adopting a more virtuous lifestyle. This kind of situation might apply to a person who had a chance to leave his high-paying but morally suspect job for a low-paying and emotionally draining job teaching children in a crime-ridden school district. As with (A), virtue might not be prudentially appealing for this person.
Case C: Finally, imagine a vicious person who would gain as much authentic moral self-satisfaction from becoming virtuous as he would gain from becoming neutral. Suppose further that he would suffer a net loss in welfare if he were to become virtuous and would enjoy a net gain in welfare if he were to become neutral. This kind of situation might apply to a criminal who has the chance to abandon his gang and either (a) take up an average job in business, or (b) take up the estimable but often dangerous job of an emergency aid worker. For this person, virtue might also lack prudential appeal.Footnote 33
So it seems that the rationality of a virtuous lifestyle is also a contingent matter: even if virtue is intrinsically rewarding, we cannot judge a priori whether it is prudentially worth pursuing; in order to make this kind of determination, we would need to examine the circumstances in which our virtue is to be exercised.
VI. CONCLUSION
In closing, I want to note that nothing I have said in this article rules out the possibility that individuals with perfect virtue benefit intrinsically from their own virtue. If we assume that the other-regarding virtues come in degrees so that (e.g.) a person could be more or less benevolent and more or less just, and we also assume that for each of these virtues there is a maximum degree to which a person can possess that virtue, then at the limit, I suspect that the perfectly benevolent and perfectly just person might have to have the kind of self-knowledge that would make it impossible for her not to know about and take satisfaction in the extent of her own virtue. And I suspect (though with much less confidence) that this person couldn't suffer from long-term depression or any other psychological condition likely to interfere with being affectively fulfilled by her own decency. Perhaps the concepts of perfect benevolence and perfect justice secure the desired link between virtue and satisfaction. But working out the details of these hypotheses and subjecting them to the scrutiny they deserve is far too great a task to take up here.
In this article, I have argued that Sumner's arguments go wrong in two different ways: first, by mistaking various contingent matters of fact for necessary ones, and second, by failing to acknowledge the possibility of satisfied morally neutral agents. For these reasons, I believe that we ought to reject those arguments. In spite of my objections, however, I remain hopeful that something like Sumner's strategy will prove that perfect virtue is indeed its own reward.Footnote 34
