1. Axioms, Exceptions, and Admonitions

Before discussing the central Winchian challenge to Sidgwick's principle, and by way of stage-setting, it is worth analyzing Sidgwick's own treatment of the principle. Even for him – generally seen as an unproblematic champion of the principle – matters are not quite as simple as they first appear. Although he sees the principle as axiomatic, Sidgwick admits an exception to it: situations in which what is right for one agent is not right for all other similarly situated agents. These are cases in which ‘my circumstances include (1) the knowledge that the rule is not universally accepted; and (2) the conviction that my act will not tend to make it so, to any extent worth considering’ (Sidgwick, Reference Sidgwick1877, p. 293).

Sidgwick's first clause admits of two interpretations, neither of which seems particularly promising for the principle. On one interpretation of ‘not universally accepted’, all that would be needed in order to cast doubt on the universality of the principle is knowledge of one person (presumably different from me) who does not accept it. This would of course make it exceedingly easy to deny its alleged axiomaticity. A facially universal principle will not apply to me if I know of someone – anyone – who does not accept it. Locke's famous skepticism about innate ideas can do all necessary work here: surely there are some who would disagree with any principle whatsoever, including not only moral principles, but mathematical and logical principles too – say, Locke's ‘children and idiots’ (Locke, Reference Locke1824, p. 14).

It may thus be better to assume that Sidgwick's point is that we need to know that a number of people reject the principle. But this looser interpretation problematically introduces vagueness regarding how many people need to reject the principle. This vagueness, I think, also casts doubt on its alleged axiomaticity. Even more vagueness would be introduced – in either interpretation – if we seize Sidgwick's shift from ‘knowledge’ into ‘conviction’ in the second clause: how strong must our conviction be?

It takes no great perspicuity to realize that Sidgwick's exception – and the problems it generates – derive from his utilitarianism: when the consequences of following the principle are worse than the consequences of following his cherished moral doctrine, then the allegedly axiomatic principle does not in fact apply. While utilitarianism, in its kaleidoscopic malleability continues to hold sway amongst moral philosophers, I will not here focus on the ways in which it may permit the violation of any principle that conflicts with its utility-maximizing ethos. I will henceforth focus on a challenge to Sidgwick's principle that does not rely on any criticism of utilitarianism.

Sidgwick also qualifies the principle in a more general way. Merely a few pages after presenting the principle in the straightforward words I quoted at the outset, Sidgwick reintroduces it in a more complicated fashion: ‘if a kind of conduct that is right (or wrong) for me is not right (or wrong) for someone else, it must be on the ground of some difference between the two cases, other than the fact that I and he are different persons’ (Sidgwick, Reference Sidgwick1877, p. 353). Sidgwick's rarely discussed change in formulation is revealing.

There is, most obviously, a recognition of the fact that what appear to be the ‘same circumstances’ may not be quite the same. To be sure, this warning is not meant to reject the importance of moral (or legal) precedent, or of analogical reasoning; neither is it a blanket denial of the possibility of relevant similarities guiding our thought. Rather, this salutary warning seeks to highlight how frequently cases that on first approximation appear to be ‘relevantly identical’ can, on closer inspection, turn out to be relevantly different.

But Sidgwick's second formulation of the principle contains an additional admonition: in trying to distinguish between two putatively ‘relevantly identical’ cases we should not appeal to differences concerning the agents in each. Again, this seems salutary: for without this admonition there simply would never exist two similar cases (in the required sense), since, ex hypothesi, there being two different agents would, eo ipso, render the cases different. It is of course crucial to keep in mind that in our context ‘similar’ and ‘like’ (etc.), mean ‘relevantly identical’. Endorsing Sidgwick's principle does not entail accepting the (absurd) view whereby there could exist two cases which, although containing different participants, are, strictly speaking, identical. This would be an uninteresting way of giving the game away from the start, by rendering the principle necessarily inoperative. But the importance of Sidgwick's admonition goes further, underscoring how much work needs to be done regarding the notion of relevance invoked here. Since this admonition will play an important role throughout, I will henceforth refer to it as ‘Sidgwick's admonition’.

2. From Perplexity to Self-Realization

In a deservedly celebrated article, Winch (Reference Winch1965) challenges Sidgwick's principle by relating it to a famous novel written by one of Sidgwick's contemporaries across the Atlantic: Herman Melville (Reference Melville1986). It is impossible to do justice to the moral complexity that Melville explores in Billy Budd without actually reading it. But a brief summary of its main theme will have to do here.

Melville's masterpiece takes place in the late eighteen-century, onboard the HMS Bellipotent. The novel's eponymous protagonist is painstakingly described as an extraordinarily morally pure young man. Two other characters are of great significance. One is John Claggart, the master-at-arms. Claggart is presented as a rather complex man, though his relationship with Billy is rather simple: he hates Billy. The third (and perhaps most) important character in Billy Budd is Captain Vere. Vere is presented as eminently just and honest, as a man of ‘exceptional character’, with a ‘marked leaning to everything intellectual’ – doubtless a genuinely moral man (Melville, Reference Melville1986, p. 311).

Claggart's hostility climaxed when he approached Vere to falsely accuse Billy (of mutiny, no less). Truly a good judge of character, however, Vere was not merely incredulous, but downright suspicious; he warned Claggart about the dire consequences of perjury. Almost as if to expose Claggart's perfidy, but certainly also because this was the proper thing to do regardless of the verisimilitude of the accusation, Vere summoned Billy, and ordered Claggart to repeat his accusation to Billy's face – an order which Claggart diligently obeyed.

Utterly stunned by the malicious slander, Billy became almost paralyzed and could only manage ‘a strange dumb gesturing and gurgling’, after which Billy punched Claggart in the face (Melville, Reference Melville1986, p. 349). When the ship's surgeon pronounced Claggart dead (as a result of Billy's blow), Vere exclaimed: ‘struck by an Angel of God! Yet the Angel must hang!’ (Melville, Reference Melville1986, p. 352). Vere summarily set up a drumhead court and convicted Billy to death (ignoring the pleading of other members of the court).

The main reason Winch disagrees with Sidgwick's principle is that while he is convinced that Vere's actions were right for him, if he were to imagine himself in Vere's position, he would not convict Billy. Thus, Winch believes that what was right for Vere is not right for a similarly situated Winch. Winch admits that he would have been (morally) unable to look past Billy's peculiar innocence, and that it would have been morally right for Winch not to convict Billy. But he believes that this is nonetheless consistent with it being right for Vere to have convicted Billy. Winch sees Vere (and his own imagined self deciding what to do in this case, for that matter) as instantiating a ‘completely morally serious [man], who fully intends to do what he ought to do but is perplexed about what he ought to do’ (Winch, Reference Winch1965, p. 205).Footnote ¹ Winch's point is that when, after deliberation, a morally serious man concludes ‘“this is what I ought to do”, there is nothing in the meaning or use of the word “ought” which logically commits him to accepting as a corollary: “and anyone else in a situation like this [i.e., relevantly identical] ought to do the same”’ (Winch, Reference Winch1965, p. 206).

Despite Winch's emphasis on moral seriousness, it is hard to agree with him about the dispensability of the corollary – something I am inclined to do – without feeling that one is thereby embracing a form of moral skepticism – something I am not inclined to do. Some clarifications are thus in order. Winch himself gives us reasons for resisting interpreting his position as a form of (Protagorean) relativism. He engages with Billy Budd not in order to cast doubt on either the reality of moral experience or the rigor of our theorizing about it, but rather to approach ‘the facts of moral experience with any sensitivity’ (Winch, Reference Winch1965, p. 213). He happens to believe that what Vere did was right for Vere and would not be right for a similarly situated Winch – but it was not made right for Vere by the mere fact that Vere thought it was right. (Nor would it be right for Winch to do what he thinks he should do simply because he thinks it right.) As Winch notes: Vere ‘did not think that whatever he thought would be the right thing would in fact be so’ (Winch, Reference Winch1965, p. 209). Similarly, Wiggins underscores that Winch's position ‘is no threat to truth’, and if it is to be deemed a form of subjectivism ‘it is only subjectivism in the good sense of introducing a subject or agent’ (Wiggins, Reference Wiggins2002, p. 170).

Winch also rejects the possibility that amoralism may lurk in the background: Vere, after all, was consumed by the moral choice – by the tragic moral choice – he faced. Indeed, Melville depicts Vere on his deathbed, a considerable time after the events onboard the Bellipotent took place, murmuring Billy's name: ‘Billy Budd, Billy Budd’ were Vere's last words (Melville, Reference Melville1986, 382). Vere resembles neither the psychopathic nor the parasitic versions of the amoralist that, say, Bernard Williams has so famously dismantled (Williams, Reference Williams1993, pp. 1–13). As the reader may anticipate, Winch's position does invite a close look at what moral particularism has to offer – a position whose non-amoralist credentials are, I think, well understood.

Albeit brief, these clarifications should suffice to dispel any notion that Winch wishes to cast doubt on morality as such. Any sort of nihilism that may unceremoniously jettison a principle as facially fundamental as Sidgwick's is beyond Winch's aims, and beyond mine. Rather, what we both wish to do is to highlight the fact that morality is complicated, in ways that affect something as apparently straightforward as Sidgwick's principle. Even if two cases appear to be relevantly identical, the perspectives of each of the agents cannot be easily stipulated away. And perspectivism should not be seen as a metaethical challenge to morality either. Again, Wiggins helpfully reminds us of a ‘platitude’: ‘perspective is not a form of illusion, distortion, or delusion. All the different perspectives of a single array of objects are perfectly consistent with one another. Given a set of perspectives, we can recover, if only they be reliably collected, a unified true account of the shape, spatial relations, and relative dimensions of the objects in the array’ (Wiggins, Reference Wiggins2002, p. 108). Similarly we may create a unified true account of moral perspectives. While these clarifications underscore how much more interesting Winch's position is vis-à-vis those forms of moral skepticism with which it may be confused, they do not remove the perplexities caused by Winch's reflections on Billy Budd. (They were of course not meant to do that.) Clarifications aside, then, let us return to Winch's main argumentative line.

Winch posits that choices such as Vere's ‘seem to span the gulf between propositions and expressions of decisions’ (Winch, Reference Winch1965, p. 209). And he immediately rhetorically asks ‘how can a gulf like that be spanned?’ (Winch Reference Winch1965, p. 209). The gap is spanned because Vere – and indeed anyone facing choices as complicated Footnote ² as Vere's – ‘is not merely concerned to decide to do something, but also to find out what is the right thing for him to do’ (Winch, Reference Winch1965, p. 209). As it turns out, it is this picture of moral deliberation that constitutes the crucial difference between Winch's position and Sidgwick's: Winch claims that ‘Sidgwick would have to say that the decision is one thing, the finding out quite another’ (Winch, Reference Winch1965, p. 209).

Surely, however, the Sidgwickian agent also needs to ‘find out’ what the right thing to do is before deciding to do it. However, while the Sidgwickian agent (just as much as the Winchian agent) needs to find something out (say, which of her possible actions will generate the best consequences), these findings are not, except in rare circumstances, related to her own particularities. Consider famous trolley problems, for example: it is typically irrelevant who the characters therein happen to be. In contrast, the Winchian agent finds out something ‘about himself’ (Winch, Reference Winch1965, p. 212). The Sidgwickian ‘finding out’ is relatively mundane: again, simply a matter of identifying which option is optimal.

I thus suggest interpreting Winch's ‘finding out’ expansively. The ‘finding out’ which will ensnare anyone facing choices such as Vere's is not mundane at all: it involves discovering, at the very least, something interesting or deep, something not meant to be quantifiable. This ‘finding out’, I think, is not a bare finding, but a much richer figuring out, or indeed a creating. It is this richness that seems missing in Sidgwick, for whom moral deliberation is not merely less agent-centered than it is for Winch, but, consequently, also much more passive and contextless. If the consequences of veering right are better than veering left (in a typical trolley problem, say), determining who is steering adds very little complexity: the consequences are what they are independently of who reflects about them, or brings them about. Not so in the case of my figuring out the best thing for me to do in morally complicated situations.

Interestingly, it is not clear that Winch would agree with my expansive reading of his position. After all, Winch appears to have tried to preempt readings of his rejection of Sidgwick's principle as leading to the sort of creative figuring out I have just sketched. He thought he needed to emphasize that nothing in his penetrating article should be construed as an endorsement of ‘any “self-realization” theory of morality’ (Winch, Reference Winch1965, p. 212). He further stressed that the ‘finding something out’ he had in mind concerns only (1) discoveries about ourselves, and (2) that are expressible ‘only in terms of the moral ideas by consideration of which [we] arrive at our decision’ (Winch, Reference Winch1965, p. 212).

But it is not clear, either, why Winch thinks that the ethics of self-realization calls for special precaution. On one reading, the ethics of self-realization can be seen as a form of perfectionism (Aristotelian or otherwise), and Winch provides no reason why we should avoid that at the ethical level. On another reading, however, something like the ethics of self-realization can evoke what Jon Elster dubbed a ‘narcissistic theory of politics’ (Elster, Reference Elster1983, p. 98), and which he so compellingly suggested is self-defeating (Elster, Reference Elster1983, pp. 91–108). (Roughly: this sort of theory defeats itself because some of its ‘goals’ cannot be goals: they necessarily are by-products.) I will return to narcissism in the last section of the paper, but the current point is that perhaps what Winch fears is a certain unmanageability that may be generated by readings such as mine, or by the (mysterious) specter of the ethics of self-realization.

Evidently, I do not think that reading Winch's rejection of Sidgwick's principle as expansively as I do is worrisome. One perfectly general reason for not over-worrying may be gleaned from an example. Sometimes we figure out what our view is, say, when writing a philosophy paper, in the process of actually writing the paper. There really is no view, or no clear view, until the paper gets written (or otherwise expressed). (I am not suggesting that all philosophy papers are like that – only that some are.) In those cases, the process of figuring out our views is indeed inseparable from the process of writing (or otherwise expressing) them, from our creating said views. That cases of this sort may be deeply interesting nowise renders them problematically mysterious or unmanageable. Similarly for some moral situations: sometimes the process of figuring out our moral views is inseparable from the process of creating them. Another more specific reason for not over-worrying is that the figuring out of which I speak does not affect the whole of morality. Admittedly, I do believe that cases in which this creative figuring out obtains are much more common than typically acknowledged. Still, some moral situations are not very complicated at all. In those uncomplicated situations Winch's reflections – or my Winch-inspired reflections – as to the creative figuring out ‘what we ought to do’ have a rather small role to play, if a role at all.

What remains to determine is what to make of Winch's rejection of Sidgwick's principle in those complicated cases in which the principle is supposed to apply. I will in due course return to this issue. For now I wish only to register a certain hesitation in Winch: he wants to reject Sidgwick's principle by appealing to a sensibility that the particular agent contributes, but without going too far, without thereby lapsing into either moral skepticism or narcissism.

3. Moral Impossibility and the Varieties of Self-Discovery

Winch's mobilization of Billy Budd as a counter to Sidgwick's principle has generated a considerable number of reactions. Many of these seek to show that Winch cannot possibly be right: he misunderstands Sidgwick's principle, misreads Vere's situation, misconstrues the process of putting himself in Vere's shoes, or fails to draw the right conclusions from his own work. For example, Samuel Guttenplan is interested in showing that the existence of moral dilemmas is no threat to moral realism. Guttenplan makes a very good case for that, but he surely goes wrong when he suggests that Winch's ‘target’ is the ‘usefulness’ of universalizability in moral reasoning (Guttenplan, Reference Guttenplan1979–1980, p. 74). While, like me, Wiggins disagrees with this aspect of Guttenplan's take on Winch, he (unlike me) nonetheless reproaches Winch for dismissing ‘the thought that war makes men like Captain Vere selectively but dangerously mad’ (Wiggins, Reference Wiggins2002, p. 172).Footnote ³ Others, such as Roger Montague (Montague, Reference Montague1974), John E. Atwell (Atwell, Reference Atwell1967), and Lilian Alweiss (Alweiss, Reference Alweiss2003), offer various arguments seeking to expose what they see as Winch's mistakes. But, before I turn to an overlooked mistake in Winch's position, I will discuss what I consider the most probing and promising – and peculiarly charitable – reaction to Winch: Raz's.

Raz locates Winch's rejection of Sidgwick's principle squarely within the contemporary debate between particularism and generalism. Given my early clarifications, it should be obvious that I am sympathetic to this framing – at least insofar as it helps distinguishing Winch's position from the sorts of skeptical positions from which I, too, have tried to distinguish it. Jonathan Dancy, one of contemporary particularism's most compelling proponents has put it succinctly, ‘moral judgment can get along perfectly well without any appeal to principles’ (Dancy, Reference Dancy2004, p. 1). Particularism has no truck with ethics as such – only with the role of principles within it. Sidgwick's principle may simply be one of those many principles to which, according to the particularist, we need not appeal. Importantly, the focus on particularism has the additional advantage of bolstering my expansive reading of Winch's ‘finding out’: pushing principles aside opens up theoretical space for creative figuring out to enter complex moral deliberation.

According to Raz, ‘typically, cases like that of Winch and Vere occur when impersonally judged there is no answer to the question of which set of considerations must prevail’ (Raz, Reference Raz, Hooker and Little2003, p. 72). In other words, when Vere deliberates about his predicament, he realizes that none of the conflicting considerations confronting him is ‘superior’ or ‘more stringent than the other’ – and this is exactly what Winch realizes too. Each of them realizes that their options are ‘incommensurate, or ‘underdetermined [by reasons, or by principles]’. Raz then probingly asks: ‘if impersonally the conflicting considerations are incommensurate what is there for agents to discover through their decisions?’ (Raz, Reference Raz, Hooker and Little2003, p. 72). Raz's answer is that what each of them discovers is ‘in a sense about himself’ (Raz, Reference Raz, Hooker and Little2003, p. 72), and it is foregrounded by paying very close attention to the exact words Winch uses when he admits that had he been in Vere's shoes: he would ‘have found it morally impossible to condemn a man “innocent before God” [as Billy was]’ (Winch, Reference Winch1965, p. 208).

In what surely will (initially) look like a weak move, I suggest not reading Winch's specific words as literally as Raz does. I think that Winch did not mean impossibility – moral impossibility that is – in anything like a literal sense. When Winch claims that he discovers that it would have been ‘impossible’ for him to convict Billy, he means something looser, similar to what we mean when we say (even if only to ourselves) when a friend who has helped us in the past asks for a small (albeit slightly inconvenient) favor, ‘How could I say no?’. Evidently, we could say no; we just think we ought to say yes. (Incidentally, this non-literal use of putative impossibility is rather common: we frequently use unimaginable, unforgivable, unforgettable, etc., in just this fashion.) Similarly, Winch could have convicted Billy, but he thinks that he should not have done so. What Winch means (or, at any rate, what he should have meant) is that there are strong grounds for him not convicting Billy – and that in deciding not to convict, he was in a sense (my sense) creatively figuring out those grounds.

Raz's puzzlement at the possibility of being ‘determined by what is not possible for one to do’ (Raz, Reference Raz, Hooker and Little2003, p. 73) dissipates if we understand that we can be determined by what (we have strong grounds to think) we ought not to do. What Raz and Winch call an impossibility, I would call a difficulty: again, it is morally difficult to say no to a friend, but it is not ‘morally impossible’. The very idea of moral impossibility may well be a category mistake.

Raz is aware that the impossibility on which he focuses should not be the result of ‘a feeling that one cannot perform the act’ (Raz, Reference Raz, Hooker and Little2003, p. 72), or of lack of resolve or guts (since Winch, heeding Sidgwick's admonition, explicitly excludes that possibility). But I do not see how Raz's specification of what may or may not be part of ‘moral impossibility’ would help solve the problem at hand. Granted, by specifying ‘moral impossibility’ as he does, Raz avoids grounding the moral quality of Winch's choice on one of the ways proscribed by Sidgwick's admonition. But this in no way precludes distinguishing between Vere and Winch by appealing to who they are: Vere is right in executing Billy because he is Vere, and Winch would have been right in doing otherwise, because he is Winch. In other words, discovering that something is ‘morally impossible’ (whatever that means) for me, whereas not morally impossible for someone else, is no solution to our problem. The problem, again, is to heed Sidgwick's admonition: to account for the possibility of a decision being right for you and wrong for me in relevantly similar situations without appealing to the (obvious) fact that we are different people, with different psychologies, characters, abilities, etc. On its face, this seems to be all there is to Raz's emphasis on this differential impossibility: Vere can (psychologically) convict Billy and Winch cannot.

Despite this crucial reservation, I believe that Raz's attempt to explain this ‘impossibility’ is promising. Raz suggests that it is through decisions of this sort that people ‘reveal’ and ‘create’ their personality, their character, and their traits and dispositions. ‘One way or another’, Raz tells us, ‘our past actions and decisions form us. They make us into who we are’ (Raz, Reference Raz, Hooker and Little2003, p. 74). In cases that are not underdetermined, when we act (more or less unproblematically) in accordance with reason ‘we reveal and we mould [merely] our executive virtues or failings’ (Raz, Reference Raz, Hooker and Little2003, p. 74). But in cases which are (or we think are) underdetermined by reason, we instead ‘reveal and mould our distinctive individuality, our tastes, our imagination, our sociability, and many of our other, including our moral, characteristics’ (Raz, Reference Raz, Hooker and Little2003, p. 74).

As an example of underdetermination, Raz offers the competing demands of justice and mercy. Faced with this incommensurability ‘some prove themselves, and make themselves, merciful by generally choosing the side of mercy. Others turn into [prove themselves, and make themselves,] stern and unforgiving people’ (Raz, Reference Raz, Hooker and Little2003, p. 74). When people contemplate acting in a way that goes ‘against the grain [of their personality]’ and that ‘offends their moral character’, they do violence to their ‘integrity and self-respect’ (Raz, Reference Raz, Hooker and Little2003, p. 75).

Raz thus appears more receptive to the possibility of self-realization from which Winch attempted to distance himself. I think he successfully captures that complexity whereby our actions simultaneously reveal and create (or ‘mould’) who we are. We have a say in what we become: living our lives involves exercising, as Raz puts it, ‘the self-determining and self-creating aspect of decision and action’ (Raz, Reference Raz, Hooker and Little2003, p. 76); and as he puts it elsewhere: ‘we are part authors of our own character’ (Raz, Reference Raz, Hooker and Little2003, p. 76). This is very much the sort of creativity that (expanding on Winch) I suggest is inseparable from choosing in complicated moral cases.

Unfortunately, however, no sooner does Raz make these remarks about the complex interrelation between our choices, our selves, and our values, than he proceeds, with a hesitation that evokes (and actually surpasses) Winch's own, to cabin their scope and reduce their importance. First, Raz suggests that the domain in which these remarks would have any relevance is considerably ‘restricted’: these considerations only have purchase when ‘impersonal reasons are incommensurate’ (Raz, Reference Raz, Hooker and Little2003, p. 77). And Raz refers to these situations as constituting a ‘limited range of cases’ (Raz, Reference Raz, Hooker and Little2003, p. 76). This is a peculiar position for Raz to take, and not just because it appears to be in tension with other remarks he makes. For example, Raz rightly points out how infrequently philosophers acknowledge ‘the many occasions in which the demands of morality themselves are indeterminate’, as he rightly emphasizes that ‘the morality of right and wrong is not exhaustive’ (Raz, Reference Raz, Hooker and Little2003, p. 74). Furthermore, and as we have just seen, Raz mentions the competing demands of justice and mercy as an example of ‘incommensurate’ choices – and cases in which the tension between justice and mercy manifests itself are extraordinarily numerous.Footnote ⁴ I think Raz is importantly right when he stresses the prominence of these sorts of cases. But then the domain of self-creation does not appear so ‘limited’ after all.

More importantly, these limitations appear in tension with the most profound and ambitious aspects of Raz's discussion of Winch's rejection of Sidgwick's principle in particular. For if these cases were really as restricted as Raz (intermittently) suggests, then there would be a corresponding small number of occasions for us to have a say in who we are. Again: ‘it is primarily where matters are underdetermined [or incommensurate]’ that we get to ‘reveal and mould’ our selves (Raz, Reference Raz, Hooker and Little2003, p. 74). It would be anticlimactic – and at any rate dispiriting – if it turned out that we only had occasion to have a say on who we are in such a ‘limited number of cases’.

One may thus be tempted to overlook the passages in which Raz restricts the scope of these cases. Alas, this is not easy to do, for Raz undermines their importance in yet more ways. For example, he deems Vere's case unusual because in ‘usual’ cases of underdetermination ‘people just follow their inclination, or follow a momentary desire, or just choose’ (Raz, Reference Raz, Hooker and Little2003, p. 73). Few of us would ever face choices as momentous as the choice Vere faces in Billy Budd, but the universal appeal of the novel is (partly) the result of the way in which it resonates with us, since it reveals that the complexity of these dilemmas is common enough. Dilemmas as complex as Vere's are an irreducible component of ordinary moral experience. And Raz's suggestion here seems to me to flatten that moral experience. An imagined Vere who would ‘just choose’ or would lightly follow a ‘momentary desire’ would cast not only a depressing figure, but an alien one.

Finally, Raz further reduces the importance of these sorts of cases by suggesting that the process of discovery of which we are speaking may not be that interesting or creative after all. He insists that sometimes what we discover (or create) need not be something unknown or surprising: sometimes it can be simply ‘reaffirming what one thought to be the case anyway’ (Raz, Reference Raz, Hooker and Little2003, p. 76). Raz even compares these situations to ‘many scientific discoveries’ in which we obtain ‘experimental confirmation of a theoretically predicted result’ (Raz, Reference Raz, Hooker and Little2003, p. 76). This is a feeble sense of ‘discovering’ and of ‘self-creation’. In these cases, Raz's ‘discovery’ and ‘self-creation’ are demoted to mere ‘reaffirmings or to mere ‘confirmations’ – to double-checkings of sorts. Raz does admit that there may be other cases in which what we discover during our deliberating about a complicated (underdetermined) moral situation may indeed ‘surprise’ us – but he is quick to assert that such ‘possibility’ is ‘unlikely’ (Raz, Reference Raz, Hooker and Little2003, p. 76).

4. Reasons and the Wrong Perspective on Perspectives

In addition to these specific ways in which Raz cabins the creativity provoked by the Winchian challenge to Sidgwick's principle, two general methodological considerations lead Raz toward a restrictive stance. First, the contemporary ascendancy of the talk of reasons; second, an exaggerated confidence in the explanatory power of the difference between first- and third-person perspectives.

In recent years, philosophers have grown increasingly preoccupied with reasons. T.M. Scanlon begins his 2009 John Locke lectures by asserting that one of the most salient characteristics of contemporary metaethics is its focus on questions related to ‘reasons for action, and, even more broadly, reasons for belief and other attitudes’ (Scanlon, Reference Scanlon2014, p. 1). Scanlon welcomes this development, and in fact champions what he calls ‘reasons fundamentalism’: truths about reasons are primitive, ‘in the sense that the truths about reasons are not reducible to or identifiable with non-normative truths’ (Scanlon, Reference Scanlon2014, p. 2). Similarly, Raz claims that ‘life is activity and we are active in so far as, as it seems to us, we function well, that is in so far as, as it seems to us, our moods, emotions, beliefs, desires, etc., are properly responsive to reason’ (Raz, Reference Raz1999, p. 20).Footnote ⁵ Furthermore, Raz claims that ‘the normativity of all that is normative consists in the way it is, or provides, or is otherwise related to reasons’ (Raz, Reference Raz1999, p. 66). John Broome's pithy gloss on Raz's passage – ‘all is reasons’ – is scarcely an exaggeration; and Broome is right, too, when he notes that ‘reasons have come to dominate thinking about normativity’ (Broome, Reference Broome, Jay Wallace and Pettit2014, p. 28).

Neither Winch nor Raz should simply say – and, to their credit, they do not simply say – that the difference between Vere's decision and Winch's (imagined) decision is that the reasons each of them has are different. For if they said that, then they would be rather obviously violating Sidgwick's admonition, and, as we have seen, that would rob our discussion of its appeal. Raz summarizes the point nicely: ‘if what makes the action right for me is something about me, it is trivial that it is not necessarily right for everyone’ (Raz, Reference Raz, Hooker and Little2003, p. 71). But how can the triviality be avoided? How can anyone accept Winch's distressing conclusion without either ignoring Sidgwick's admonition or else unwittingly conceding that the two cases are different after all?

Raz's answer, nestled within the thicket of reasons, is complicated. He is intent on establishing the following difference: ‘it was not that Vere's character or personality, or moral sensibilities, or anything like that figures among his reasons. At least it does not figure as such, under that description. It was merely that it is part of what makes the decision right’ (Raz, Reference Raz, Hooker and Little2003, p. 71). Clearly, had any of these characteristics figured as reasons for Vere then the case would violate Sidgwick's admonition, and the discussion would indeed become trivial at best. Raz rather wishes to insist that while the agent does not invoke these characteristics as reasons, these characteristics remain somehow about the agent.

Raz's strategy involves two crucial moves. The first is to highlight that ‘our knowledge of [our] reasons exceeds our ability to articulate them’, or, in other words, that ‘we know more than we know we know’ (Raz, Reference Raz, Hooker and Little2003, p. 77). Winch's process of discovery, which in some cases underdetermined by reasons is inseparable from choosing what to do, involves drawing from an ‘inarticulate fund of knowledge at our disposal’ (Raz, Reference Raz, Hooker and Little2003, p. 77). This fund contains all sorts of reasons for action, even if we are (literally) unaware of their existence, and even if we are (literally) unable to appeal to them in explaining our decisions.

But explain we must – and here enters the second Razian move: he posits a sharp difference between reasons and explanations – both at the level of accounting for our actions and at the level of accounting for their moral status.Footnote ⁶ Reasons are not universalizable, and this allows Raz to agree with Winch and to in fact concede quite a bit to the particularist (again: albeit only within a ‘limited’ or ‘restricted’ domain). Explanations, on the other hand, are universalizable, and they do guarantee ‘the intelligibility of reason’. In the case at hand: ‘what makes it right for Vere to decide as he did, and for Winch to decide contrariwise, can be explained. And the explanation, relying on the difference in their moral character, and in the concrete fact that they did decide as they did, is universalizable’ (Raz, Reference Raz, Hooker and Little2003, p. 77). Thus, for Raz, while there is no universal truth of the matter as to what the right thing to do about Billy's fate is, there is a universal truth of the matter as to the explanation of how this is so.

This is all unquestionably subtle. I am not sure, however, that it is very stable. For example, Raz admits that sometimes we should trust agents based on their track-record, ‘even when they cannot explain their judgment’: again, because they may have more reasons than they are aware or can articulate, or because the fact that a reason is not (or cannot be) fully articulated does not show that it does not exist (Raz, Reference Raz, Hooker and Little2003, p. 77). Though I find this a bit obscure, I do suspect that Raz is right about at least some of it. A pressing problem facing Raz's position, however, cannot be easily dismissed: How exactly are we to universalize what we cannot even articulate? Once we enter the realm of the ineffable, it becomes hard to maintain our grip on what ‘universalization’ may even mean.

Moreover, and for our purposes more importantly, I am not sure that Raz's position addresses the pertinent problem of universalizability under consideration. What the discussion of Sidgwick's principle problematizes is the universalizability of (first-order) moral principles, not of (second-order) explanations. So even if a satisfactory account of the universalizing of explanations ranging over ineffable reasons could be given, that is not what we needed. What we needed was an account of the universalizability of reasons, not of explanations (to use Raz's terms). Raz's distinction between reasons and explanations does not, then, help avoid the lurking sense of triviality. It does not, that is, manage to avoid either violating Sidgwick's admonition or else (unwittingly) conceding that the cases are indeed different after all.

As for Raz's second move, centered on the distinction between the first-person and the third-person perspective, I find the appeal to it within the context of our current discussion somewhat disconcerting. For it strikes me as obvious that the force of Sidgwick's principle would remain in cases in which there is but one single imagined agent, who would say (to herself) ‘what is right for me on this occasion must be right for me in all relevantly identical occasions’. The force of the principle would seem to hold, too, in cases in which one single agent is comparing not herself against someone else (or herself in two relevantly identical occasions), but two other agents – living or dead – against each other (or another single agent at two different times). And in all these cases the problematics of Sidgwick's principle would obtain. In other words, versions of Winch's challenge can be easily generated when the two cases being compared belong to the first-person perspective, and when the two cases being compared belong to the third-person perspective.

The scenario last mentioned – whereby a person evaluates two situations in which she plays no role whatsoever – highlights also the limitations of the potentially excessive focus on actions, on ‘what we ought to do’. For in this scenario there simply is nothing for our agent to do (other than to evaluate someone else's actions – not the requisite sort of ‘doing’) though the force of Winch's rejection of Sidgwick's principle remains the same. Focusing too much on actions is a widespread tendency in moral philosophy, conspicuous in both utilitarian (except G. E. Moore's ideal utilitarianism) and deontological ethical systems.Footnote ⁷

In the end, then, all Razian moves face difficulties; I do not see how they can prevent him from running afoul of Sidgwick's admonition. In the end, Raz, just like Winch, seems too interested in limiting the consequences of the rejection of Sidgwick's principle. In so doing, he, again like Winch, has missed an important opportunity to further advance our understanding of the richness of moral experience.

5. Ways of Lifemaking

I submit that neither Winch nor Raz succeeds in showing that there can exist two ‘relevantly identical’ cases without smuggling – in admittedly sophisticated and insightful ways – a violation of Sidgwick's admonition. The work expected of the term ‘relevant’ in the expression ‘relevantly identical’ is quite literally an impossibility: anything facially irrelevant can turn out to be relevant (Zaibert, Reference Zaibert and Kaspar2020). When Sidgwick's admonition is heeded, his principle is revealed as nothing more than a natural extension into morality of Leibniz's principle concerning the identity of indiscernibles (Leibniz, Reference Leibniz1902, p. 14). Just as it is always possible to find some similarity between any two entities or phenomena, it is also always possible to find some difference between them. Otherwise, we are not in the presence of two different entities or phenomena – and this holds as much within morality as it does anywhere else.

Insofar as Sidgwick's admonition has – despite appearances – not really been heeded, we must conclude both that Winch has not refuted Sidgwick's principle, and that Raz is mistaken in accepting Winch's refutation (in the limited number of highly restricted cases in which he accepts it). Surprisingly, however, this is not good news for Sidgwick's principle.

Sidgwick's principle can only survive at the expense of a rich understanding of moral experience. And this survival strikes me as a Pyrrhic victory – so Pyrrhic that it invites skepticism whether it should count as a survival at all. If recognizing the richness of moral experience is necessary for good moral philosophy (necessary, that is, for approaching ‘the facts of moral experience with any sensitivity’ (Winch, Reference Winch1965, p. 213)), and if the survival of Sidgwick's principle is inimical to that richness, then in what sense has the principle ‘survived’?Footnote ⁸ The principle ‘survives’ in that the efforts to show that it is false fail; the principle remains true. Truth, of course, matters – but some truths can be uninteresting, or worse. This is the case of Sidgwick's principle: rather than illuminating the rational basis of morality – as Sidgwick believed, and as it is so natural to assume – the discussion here reveals that the principle, itself, really is remarkably trivial. Although trivial, moreover, the principle is also misleading.

Sidgwick's principle primes us to expect moral experience to be much simpler than it is. The conjunction of this priming effect and an exaggerated emphasis on our reasons and our actions really obscures quite a bit. As we have seen, moral experience can involve cases in which no action of ours is required. Moral evaluation, for example, is an important dimension of moral experience: we can morally evaluate agents in cases in which we have no role to play beyond the evaluation itself. We can agonize over these evaluations: they matter to us, and we care about getting them right.

It should by now be clear that, despite my reservations regarding one aspect of his reading of Billy Budd, I have benefited enormously from Wiggins's impressively rich and probing understanding of morality. He captures nicely part of what the conjunction just noted obscures: ‘philosophy has put happiness in the place that shall be occupied in moral philosophy by meaning’ (Wiggins, Reference Wiggins2002, p. 88). Neither Vere nor Winch deliberates exclusively about which of their choices is likelier to generate more happiness, or will have better consequences, at least not in a narrow, standard sense of ‘consequences’ (e.g., the forward-looking sense of utilitarianism, and of much of contemporary moral theory – and the sense in which I use the term here). Rather, they agonize over the complexity of their choice, and that complexity forces them to also think about things other than the consequences of their actions: meaning is part of what they are after.

This consideration of meaning offers a much more capable gloss on Winch's talk of impossibility than the fraught sense of ‘moral impossibility’. The reason why Winch finds it so difficult to convict Billy is that it is hard to make sense (to ‘make meaning’ if the phrase be allowed) of such action. The meaning to be made here is related to the value-rich, organic whole that is created by our choices – a whole that prominently contains us. This whole constitutes the always-evolving entity that is our moral life. It is possible that the best whole that Vere can create is value-wise different from the best whole that Winch can create. But just as this better not be baldly based on the fact that Vere is Vere and Winch is Winch, it better not be exhausted by the consideration of the consequences of their actions either.

The creative process of figuring out what we are supposed to do relates to what we want to make of the world around us – a world which of course includes our own moral lives. Often, the action with the best consequences is also the action that generates both a better world and a better version of ourselves. Alas, this is not always so: potential divergences between these dimensions highlight the importance of, say, Williams's views on integrity (Williams, Reference Williams, Smart and Williams1973, pp. 109 ff.), the discussion of agent-centered morality, the distinction between doing and allowing, etc. We may, on occasion, refuse to do the action with the best consequences, because we do not want the world to contain us doing that (even if we know someone else will do it), or because we do not want to be the sort of person who does that. Of course, sometimes consequentialist considerations are dominant, and refusing to act in order to avoid taints in those cases may be morally obtuse. And yet, a morally serious person who chooses suicide over doing a truly horrible thing (even if we stipulate that doing the horrible thing has the best consequences) is not only not inconceivable but could be quite admirable.

I do not wish to overstate my case. What we actually do (as opposed to what we otherwise make of the world and of our lives) is unquestionably very important. While I am not here interested in ranking these different dimensions of moral experience, it is likely that, in general, the active dimension is most essential. Despite Winch's efforts to put himself in Vere's shoes, it was (the fictional) Vere – or at any rate, a real navy captain – who really faces the complex choice, not Winch. Real choices may provide us with both a lens and a chisel into who we are that mere imaginings and extrapolations, no matter how earnestly and judiciously pursued, just cannot afford. We may, perhaps, never be as certain as to how we may have acted in a given situation unless we actually were in it – and acted. Furthermore, we may be haunted by having been unable to face a particularly difficult and important choice, which in turn may have provided us with an opportunity to prove ourselves – and to figure out who we really are.Footnote ⁹

Despite the undeniable importance of actual choices and actual actions, however, the non-action-oriented dimensions of moral experience I have here emphasized deserve attention too. Note that we may deeply (morally) respect – the haunted agent just described, even if we stipulate that her anguish has no consequences (say, she only mentioned it to me, and I do nothing about it, etc.). Non-action-oriented dimensions of moral experience are perfectly real, and they are also important. They add a type of complexity to moral experience that Sidgwick's principle tends to obscure.

This complexity brings me to the title of this final section. The obvious invocation of Nelson Goodman's views on worldmaking does not signal adherence to his thoroughgoing metaphysical irrealism. Dazzling as Goodman's thought unquestionably is, it is difficult not to agree with Israel Scheffler when he objects to Goodman's unbridled worldmaking powers noting that it is ‘obvious that we made the word “star”, and equally obvious that we did not make the stars’ (Scheffler, Reference Scheffler2001, p. 668). But if there is a context in which Goodman's views are attractive, it is ours. Many – if not most – of the problems facing Goodman's irrealism tend to either disappear or to radically diminish when we apply his views to our own lives. Admittedly, ‘our lives’, ‘lived lives’, ‘moral lives’ are sui generis entities. But they are entities, and creating our own lives is quite a different enterprise from creating our own (physical or metaphysical) world: we do create (aspects of) our character and (aspects of) our own moral landscape, say, in ways we evidently do not create stars. Jerome Bruner seems right when he casts doubt on ‘being a naïve realist about “life itself”’, even if one may rather part company when he further claims that ‘“world making” is the principal function of mind’ (Bruner, Reference Bruner1987, p. 13; p. 11). Again, I do not wish to overstate my case: something like worldmaking – lifemaking – clearly is a function of our mind.

As Goodman has put it, ‘much but by no means all worldmaking consists of taking apart and putting together, often conjointly: on the one hand, of dividing wholes into parts and partitioning kinds into sub-species, analyzing concepts into component features, drawing distinctions; on the other hand, of composing wholes and kinds out of parts and members and subclasses, combining features into complexes, and making connections’ (Goodman, Reference Goodman1988, p. 7). This, I submit, is an important dimension of moral experience: creatively figuring out the best moral world, and the best version of ourselves. Winch sees more meaning in the state of affairs in which he does not convict Billy; Vere sees more in one in which he convicts. The differential value in these complex states of affairs explains how their different decisions can be both right.

By way of conclusion, I return to Winch's worries as he challenged Sidgwick's principle: that it was a form of moral skepticism and that it was a form of narcissism. My appeal to lifemaking here commits us to neither of these.

First, my appeal is not a re-emergence of Protagorean relativism (or other forms of moral skepticism). It is not (baldly) because Winch is Winch and Vere is Vere that the meaning and value in these states of affairs can indeed be different. Rather, it is because the narrative that they cannot avoid constructing is different. Bruner again seems right when he claims that ‘we seem to have no other way of describing “lived time” save in the form of a narrative’ (Bruner, Reference Bruner1987, p. 12). And even Goodman admits that this narrative structure imposes ‘severe restraints’ to his radical relativism (Goodman, Reference Goodman1988, p. 94). It imposes at least equally severe restraints within the moral realm. Surely autobiography often is an important aspect of moral experience, but so is the coherence between the different aspects that make up the moral world (a world that includes more than our selves).

When facing complex moral choices, some of these restraints flow from looking back at our moral life up to the moment of this particular choice, so as to ascertain which choices may better fit narratively with what we have made of our selves – selves that will, as long as they exist, inevitably remain works-in-progress. Other restraints flow from looking around trying to ascertain which choices better fit narratively with the world that we continually contribute to make. This process is rather different – and rather richer – than merely looking forward and calculating the best consequences of our choices, and it may require deep examinations of our moral life and our moral world. The results of these examinations may on occasion be unexpected to the point of perplexity.

I suspect that this perplexity plays a role in the hesitation that both Winch and Raz evince: they want to challenge Sidgwick's principle and yet they strive to cabin the scope or the relevance of the results of said challenge. But I also suspect that both Winch and Raz are under the spell of the priming effect of Sidgwick's principle. If they did not expect moral experience to be as simple as Sidgwick's principle primes us to expect, they would not have so strenuously sought to show that while Vere's case and Winch's case are taken to be relevantly identical, their contradictory choices can both be morally correct. They could have instead admitted that the two cases are different (despite their hypothesized ‘relevant similarity’), and that this difference does not really save – in a robust sense of ‘saving’ – Sidgwick's principle. The difference between the two cases is not simply the result of latching onto the fact that, strictly speaking, no two cases can ever be identical. Rather, it is the result of paying attention to typically overlooked forms and loci of difference. And I have suggested here that Sidgwick's principle contributes to this overlooking.

Second (and lastly), it may be feared that my account cannot fully elude the charge of narcissism, for an excessive focus on our actions, our reasons, our perspectives, and our selves, does seem to rekindle Elster's fear of narcissism. But, again, this is not all that my agent does. It would have been a mistake to title this final section ‘Ways of Selfmaking’.Footnote ¹⁰

Relatedly, it could be argued that my account of lifemaking merely replaces the narcissism of self-realization with the narcissism of the moral aesthete. My agent appoints herself an arranger and re-arranger of lives and worlds, an appraiser and reappraiser of these lives and worlds. My agent, however, does these things in addition to deliberating about moral action (and its consequences) in traditional ways – although as she deliberates and acts in these traditional ways my agent often cares about how these actions fit into her moral life and her moral world. If positing that moral agents are complex and multidimensional is narcissism, then, echoing Wiggins, it may be only narcissism in the good sense of positing real human beings who care about moral lives and worlds – their own and others’ – instead of mere cutouts. If this is narcissism, I can certainly live with it, and I think we all should.Footnote ¹¹