A Shelter from Luck: The Morality System Reconstructed

Author: Matthieu Queloz
Published in: In *Morality and Agency: Themes from Bernard Williams*. András Szigeti and Matthew Talbert (eds.), 184–211. New York: Oxford University Press. 2022. [doi:10.1093/oso/9780197626566.003.0009](https://doi.org/10.1093/oso/9780197626566.003.0009)
DOI: 10.1093/oso/9780197626566.003.0009
Canonical entry: https://www.matthieuqueloz.com/entries/a-shelter-from-luck-the-morality-system-reconstructed/
Published PDF: https://philpapers.org/archive/QUEASF.pdf

Machine-readable text companion generated from the PDF. Page markers follow the journal pagination.

[p. 185]

A Shelter from Luck criticism of the particular forms that these building blocks assume within the system. When Williams remarks that we would be better off without the system, he does not mean that we should completely jettison the conceptual material it draws on. His aim is the more constructive one of making “some sense of the ethical as opposed to throwing out the whole thing because we can’t have an idealized version of it” (2009, 203).

By juxtaposing Williams’s vindicatory explanations with his critique of the ideas making up the morality system, I aim to reveal the connections between them; in particular, I shall argue that understanding what these ideas do for us when they are not in the service of the system is just as important to leading us out of the system as the critique of that system because it offers us an alternative and more reflectively stable way of making sense of them in vindicatory terms: it offers us vindicatory explanations of why we came to live by these ideas that strengthen our confidence in those ideas and the reasons we take to justify their application because they suggest that these ideas are not just products of deception or holdovers from the enchanted world, but ideas that it makes sense for us to cultivate given our needs and concerns.2 It is crucial to Williams’s critique that the system’s “idealized version” of the ethical is not a philosopher’s fantasy that does little harm outside the seminar room, but “the outlook, or, incoherently, part of the outlook, of almost all of us” (2011, 194). It is incoherently part of our outlook because we also have, alongside the system’s rarefied conceptions of moral value, obligation, voluntariness, and blame, more everyday conceptions of these things, and, much of the time, it is these more relaxed conceptions that we act on. Hence, when we talk simply of “our concepts” in the coarse-grained way that ignores finer distinctions between different conceptions of the same thing, there is an important sense in which “our concepts” are not those of the system. This explains how Williams can say that “[w]e have fooled ourselves into believing that we have a more purified notion of moral responsibility than we have” (1999a, 163), and why he is drawn to the distinction between what we think and what we merely think that we think (1993, 7, 91). But, as this formulation itself brings out, the boundary between what we think and what we think that we think cannot ultimately be a sharp one: even what we merely think that we think will often have very real effects—not just on what else we think, but on how we end up living. The morality system may be an 2 I elaborate on the idea of vindicatory explanation in Queloz (2021a, 38–41, 98, 178–187, 213–221).

[p. 186]

unrealizable vision, but it is not, for all that, unreal. It is the real problem of an unrealizable vision.

To understand how the system’s conceptions and our more everyday conceptions can be seen as conceptions of the same things at all and how they relate to each other, Williams approaches them in light of a tertium quid: maximally generic conceptions of ideas and practices that are schematic, underdetermined, and probably fictional, but that nevertheless present us with helpful models of the conceptual practices we really do find in human societies. For Williams, the point of considering these prototypes lies in identifying defeasible reasons to think that any society will develop conceptual practices along these lines because they answer to needs of a very basic sort. To keep track of which of these three conceptions of a given conceptual practice is at issue, I shall use the subscript (GEN) to mark the generic conceptions that are too underspecified to be situated in space or time, (UND) to mark the undemanding conceptions we actually live by a lot of the time, and (MS) to mark the morality system’s more demanding conceptions. My argument falls into three parts. The first section of this chapter considers vindicatory explanations, in terms of generic needs, of what will turn out to be the four crucial building blocks of the morality system: the moral/nonmoral distinction(GEN), the idea of obligation(GEN), the voluntary/involuntary distinction(GEN), and the practice of blame(GEN).3 This part performs a double function: it explains why these conceptual practices are there to be harnessed by the system in the first place, and it offers us a way of making sense of them that is independent of the system. The second section is a vindicatory explanation, relative to the need for ultimate fairness, of the way in which the system combines and refines these building blocks into the moral/ nonmoral distinction(MS), the idea of obligation(MS), the voluntary/involuntary distinction(MS), and the practice of blame(MS) in order to provide a shelter from luck. Reconstructing the system in light of this organizing ambition gives us a good grasp on why it has the shape it has and what the different components of the system contribute. The third section is a critique of the resulting construction: I argue that, on Williams’s view, the ultimate problem 3 A fuller treatment than I have room for here might add guilt, which Williams contrasts in particular with shame (1993, 1997), and which he describes as “the characteristic first-personal reaction within the system” (2011, 197). The combination I repeatedly explore here, of a vindication of the generic form of X with a critique of the refined form it takes within the system, can be found also in Williams’s treatment of guilt. In its generic form, guilt helpfully “turns our attention to the victims of what we have wrongly done” (1993, 222). But this virtue is lost once the conception of guilt is elaborated into something more abstract in the morality system.

[p. 187]

A Shelter from Luck with the system is its frictionless purity. It robs valuable concepts of their grip on the world we live in and, by insisting on purity from contingency, it threatens to engender nihilism about value and skepticism about agency. To overcome these problems, it is not enough to accept that contingency pervades human life. We also need to revise our understanding of what the fact of contingency entails. In particular, we need to abandon the purist attitude that blinds us to alternative ways of making sense of human values and agency—alternatives that naturalistic but vindicatory explanations can provide.

8.1 Vindicatory Explanations of Four Building Blocks of the Morality System 8.1.1 The Moral/Nonmoral Distinction Let us begin with what is arguably the most basic building block of the morality system: the moral/nonmoral distinction(GEN). Critical as Williams may be of the particular form which the distinction between the moral and the nonmoral takes within the morality system, he still has a vindicatory story to tell about our need to draw some distinction along those lines. One of the aims of his 1972 book Morality, Williams declared in the preface he added in 1993, was the “placing of morality in relation to other ethical considerations and to the rest of life” (2001, xiv). The book achieves this, notably by examining “what the distinction between the ‘moral’ and the ‘nonmoral’ is supposed to do for us” (2001, xiii). Not yet observing the distinction he later came to draw between the “moral” as construed by the morality system and the broader notion of the “ethical,” his 1972 inquiry concerns the moral/nonmoral distinction(GEN), of which the moral/nonmoral distinction(MS) is a particular sociohistorical elaboration. Williams suggests that the moral/nonmoral distinction(GEN) is one we should expect to find in any human society. It may be possible for an individual to live outside the ethical life, but no community can get by without some minimal ethical consciousness which stakes claims against selfinterest.4 To stake claims against self-interest, Williams contends, is “one basic and universal function of morality” (1973b, 250). Any morality, in 4 See Williams (2011, 32, 51).

[p. 188]

order to count as a morality at all, needs to involve some distinction between actions which only minister to the interests of the agent and actions which take the interests of others into account (2001, 66). In its most primitive form, the concept of the moral marks a distinction between these two kinds of action and selects the latter for approval.

Of course, this primitive distinction between two classes of actions is still too primitive; if we are to make sense of anything like our concept of the moral, we need to understand why it involves discriminating not just between different kinds of action, but also between different kinds of motivation. The explanation that Williams gives turns on the example of “a self-interested business man who writes a cheque to famine relief ” (2001, 66), thus doing a good thing, but whose concern is for his own reputation rather than for the relief of famine. “What,” Williams asks, “is the point and content of saying that we do not morally approve of the self-interested donor to charity, or that, though he does a good thing, he does not act morally?” (2001, 67). Why refine the moral/nonmoral distinction far enough to discriminate not just between actions, but also between the motives from which they spring? The answer, Williams suggests, is that we have reason to discriminate not just between self-interested and non–self-interested actions, but also between (a) non–self-interested actions done out of self-interested motives and (b) non–self-interested actions done out of non–self-interested motives. Giving money to charity out of concern for one’s reputation is still better than a self-interested action—the self-interested donor will, after all, help relieve famine, and this is surely “better than that another combined cocktail cabinet and TV set should be bought” (2001, 66). Yet there is still “a very good point” (2001, 67) in withholding moral approval in the case of the self-interested donor and reserving it for cases of type (b). While cases of type (a) only yield non–self-interested actions when these happen to align with the agent’s self-interest, the motivations in cases of type (b) are steadier because they are general dispositions to do things of the non–self-interested sort: motivations grounded in principle, as a Kantian emphasis would have it, or in sympathy with others, as a Humean emphasis would have it. “This must surely,” Williams remarks, “have something to do with the point of selecting certain motives for moral approbation: we are concerned to have people who have a general tendency to be prepared to consider other people’s interests on the same footing as their own” (2001, 68).5 5 There is a parallel here to Kitcher’s (2011) claim that the most basic point of morality is to remedy “altruism failures.” See also Williams (1980).

[p. 189]

A Shelter from Luck The most basic point of having a concept of the moral that ties moral action to moral motive, then, is to steady tendencies toward selflessness by selecting for special approval “general dispositions to do things of the non– self-interested sort” (2001, 69–70). So far, this is a vindicatory story. We get a vindicatory explanation, first, of why any community would need to draw some kind of distinction between moral and nonmoral actions, and then of why this distinction would need to be focused further to make a moral action one that stems from a certain kind of motive.

8.1.2 The Idea of Obligation Another crucial source material for the morality system that Williams aims to achieve an independent grip on is the idea of obligation(GEN). Why do we have it in the first place? Williams goes some way toward offering an explanation that might “help us to understand the point and value of living a life in which obligations counted as ethical reasons” (2006e, 73). The idea of obligation(GEN) (along with its correlate, the idea that those toward whom one has an obligation have a right)6 is grounded in the basic interest of human beings everywhere in being able to rely on certain things (e.g., that they will not be killed, assaulted, or arbitrarily expropriated; Williams 2011, 205). The idea of obligation(GEN) works to secure reliability by helping to create “a state of affairs in which people can reasonably expect others to behave in some ways and not in others” (2011, 208). In particular, it works to ensure that considerations of importance are given high deliberative priority. This, according to Williams, is the most basic point of the idea of obligation(GEN). A consideration enjoys such priority for us if, first, it appears in our deliberations, and, second, it is given heavy weighting against other considerations (2011, 203). The concept of obligation(GEN) is like the special email format which ensures that important emails are flagged as “high priority” when they appear in the recipient’s inbox: it provides a format for ethical considerations that lends them particular prominence and weight in people’s deliberations.

Using the idea of obligation, we can try to make sure that considerations of basic and standing importance are reflected in “settled and permanent pattern[s] of deliberative priorities” (2011, 206). But obligations can also be 6 See Williams (2011, 206).

[p. 190]

more context-sensitive. If the person next to me suffers a stroke, a “general ethical recognition of people’s vital interests” becomes “focused into a deliberative priority by immediacy” (2011, 206). Immediacy to me generates an obligation for me to help. Even more conditional are obligations generated by promises. The institution of promising “operates to provide portable reliability,” as Williams puts it, “offering a formula that will confer high deliberative priority on what might otherwise not receive it” (2011, 207–208). In light of this, it is not altogether surprising that the morality system should have grown around the notion of obligation rather than around some other category of ethical thought. If Williams is right, the notion of obligation is a device which originally serves to acknowledge and reflect, at the level of deliberation, the overriding urgency and demandingness of our most basic needs and our needs in situations of emergency. These practical origins explain why the device is so demanding—it is “just because the needs involved are so elementary that the psychological mechanisms designed to meet those needs are demanding” (1995g, 205). Yet it is also “because those mechanisms are demanding that the theory which grows around them becomes so dense and oppressive” (1995g, 205). Taken beyond its proper remit, the device of obligation soon seems absurdly overpowered, rather like a Roman dictator who retains his emergency powers beyond the fulfilment of his mandate.

8.1.3 The Voluntary/Involuntary Distinction The third crucial building block is the voluntary/involuntary distinction(GEN). Williams thinks that it can be constructed already out of distinctions that humans everywhere are bound to find worth having. “All conceptions of responsibility make some discriminations” (1993, 66) between what is voluntary and what is not. Though “no conception of responsibility confines response entirely to the voluntary” (1993, 66), the voluntary/involuntary distinction, and the concomitant notion of “the will” in a correspondingly unambitious sense, are universally worth having. His idea is that the need for the voluntary/involuntary distinction(GEN) grows out of the need for some practice of recognizing responsibility(GEN). Already in his essay for the BBC’s Third Programme, Williams observes that this is a practice which the Greeks shared with us in some form (1963, 1–2). Like us, they recognized that to be responsible for a state of affairs is not just a matter of being the cause of it through some bodily movement. To determine

[p. 191]

A Shelter from Luck whether someone carries responsibility for that state of affairs in the full sense which makes them a proper subject of blame, we want to know more about the bodily movement. Was it just a nervous twitch, or did they really act? Did they intend to bring about that state of affairs? And what state of mind were they in when they did so? In Shame and Necessity, Williams spells out four “basic elements of any conception of responsibility” (1993, 55): Cause: the idea that someone brought about a bad state of affairs in virtue of what they did; Intention: the idea that they intended that state of affairs; State: the idea that they were in a normal state of mind when they brought it about (i.e., not sleepwalking or subject to extreme incident passions); Response: the idea that this calls for some response on their part, that they need to make up for it.

Out of these four basic elements, a great many different conceptions of responsibility can and have been constructed by interpreting the elements in different ways and varying the emphasis between them. These are “universal materials” (1993, 56) because the need for them follows “simply from universal banalities” (1993, 55). It may be plausible enough that the ideas of Cause and Response are universal, but why should we expect the ideas of Intention and State to be universal? Williams’s answer is that we are bound to be interested in drawing distinctions between what is intended or done in a normal state of mind and what is not, because these distinctions are crucial to understanding how an action relates to an agent’s plans and character. Williams gives an example from the Odyssey, where Odysseus and Telemachus confront Penelope’s suitors and find, to their great alarm, that the suitors are handing out the weapons that Telemachus was supposed to have hidden in a storeroom. Odysseus angrily wonders who opened the storeroom, and Telemachus explains that it was his mistake and that no one else is to blame—he left the door of the storeroom ajar, and one of the suitors must have been a better observer than he was (1993, 50). Telemachus is clearly drawing a distinction here between aspects of what he did that were intentional and aspects that were unintentional: it was he who left the door ajar, but he did not mean to. This, Williams contends, shows that although Homer had no direct equivalent for the word “intention,” he nevertheless had the concept of intention—not because we

[p. 192]

are disposed to draw on this concept in describing the situation, but because Homer and his characters themselves make distinctions which can only be understood in terms of that concept (1993, 50–51). For Williams, it is no surprise that they draw such distinctions: “it must be a possible question how the intentions and actions of an agent at a given time fit in with, or fail to fit in with, his intentions and actions at other times,” he writes, because “under any social circumstances at all, that is a question for other people who have to live with him” (1993, 56). Being sensitive to which aspects of an action are intentional helps us understand what kind of action it is and what else to expect from someone who, in that situation, intends those things. If Telemachus had intentionally left the door ajar, this would have disquieting implications for Odysseus, suggesting that Telemachus was not, after all, on his side. Here, the question whether Telemachus meant to do what he did is a matter of life and death.

Similarly, we are bound to care whether actions are done from a normal state of mind, because this is crucial to figuring out what to expect from people: if someone acts intentionally but in a strange state of mind, we know that their actions probably do not stand in a regular relation to their plans and character. Williams gives the example of Agamemnon, who took Briseis from Achilles and did so intentionally, but in a strange state of mind: the gods cast ate (delusion) on his wits, so that he was in a state of blind madness (1993, 52). Being sensitive to the state of mind from which people act helps us see how their intention and action on one occasion fits into the broader pattern of their intentions and actions on other occasions. It helps us separate the exceptional from the expectable. Like the capacity to separate the intentional from the unintentional, this is a capacity that people living together are bound to have an interest in possessing.

This brings us to the key point for our purposes: namely, that if the notions of Cause, Intention, State, and Response are available, one already has all the material necessary to construct the notion of the voluntary(GEN). It earns its keep in virtue of the need to recognize responsibility for certain actions and to understand the place of intentions and actions in people’s plans and characters.7 The notion of the voluntary(GEN) picks out all those actions we are left with once we have filtered out things done unintentionally or in an abnormal state of mind: “a certain thing is done voluntarily if (very roughly) it is 7 Williams puts it even more strongly in a footnote to “Moral Luck: A Postscript”: “the idea of the voluntary . . . is inherent in the concept of action” (1995e, 247 n. 4). See also Williams (2002, 45, 1995c, 1995f, 1999a; Magee and Williams 1971).

[p. 193]

A Shelter from Luck an intentional aspect of an action done in a normal state of mind” (1993, 66). Williams sees a similar notion at work today. Speaking about our presentday version of the voluntary/involuntary distinction(UND), he writes: “ ‘A does X voluntarily’ is equivalent to ‘A does X intentionally in a normal state of mind’ ” (2006a, 120).8 These notions of the voluntary—voluntary(GEN) and voluntary(UND)—may strike one as superficial: push beyond a certain point questions such as what exactly someone intends, what makes it true that they intend it, and whether they intended to become the kind of person who can intend such a thing, and this notion of the voluntary gives out. But for Williams, its superficiality is precisely what makes it worth having: “if voluntariness is to do its work such questions cannot be pressed beyond a certain point” (2006a, 124); it is “an essentially superficial notion, which works on condition that one does not try to deepen it” (1995i, 495). A useful notion of the voluntary is one that helps us capture such obviously important differences as that between intentionally turning on the stove and somnambulism. Distinctions at this superficial level do nothing to settle the problem of free will (nor, indeed, do they generate that problem in the first place).9 But it is by doing work at this level that the notion of the voluntary helps us to live. It is (in Nietzsche’s phrase which Williams quotes more than once) superficial out of profundity.10 8.1.4 The Practice of Blame The last building block I want to examine, which also grows out of the practice of recognizing responsibility, is the practice of blame(GEN). Here also, Williams’s critique of blame(MS) is rooted in a more charitable account of blame(GEN), an account which presents blame(GEN) as a valuable, if peculiar, “instrument of social control” (1995c, 15). It is an instrument of social control because it helps sustain communities by inducting new individuals 8 Another formula he uses is: “an agent does X fully voluntarily if X-ing is an intentional aspect of an action he does, which has no inherent or deliberative defect” (1995h, 25). 9 See Williams (1993, 67–68): “It is a mistake to suppose that the notion of the voluntary is a profound conception that is threatened only by some opposing and profound theory about the universe (in particular, to the effect that determinism is true). That supposition underlies the traditional metaphysical problem of the freedom of the will . . . there is a problem of free will only for those who think that the notion of the voluntary can be metaphysically deepened.” 10 Nietzsche himself uses the phrase more than once (2001, preface, §4; 2005, epilogue, §2). For further discussion of the superficiality of the notion of the voluntary, see Williams (1993, 67–68; 1995b, 127–128; 1995e, 243; 1995i, 495–496; 2006a, 124–125).

[p. 194]

(including notably children) into a shared ethical sensibility and secure some degree of realignment with that sensibility where individuals problematically deviate from it. It is a peculiar instrument, however, because as long as the participants in the practice of blame think of it in purely instrumental terms, as justified only by its efficacy as a tool of alignment, blame will fail to produce the desired effects—it will tend to produce resentment rather than remorse: Blame that is perceived as unjust often fails to have the desired results, and merely generates resentment. This shows that the idea of blame’s justification is not the same as the idea of its efficacy. When a recipient thinks that blame is unjustified, the content of his thought cannot be that the blame will be ineffective. This does not show that the purpose of blame may not in fact lie in the modification of behaviour; it means only that if this is true, it cannot be obvious to those who are effectively blamed. (1995c, 15) The practice of blame can be efficacious only insofar as it is understood by participants to be more than just a regulative device, because only then will recipients of blame be suitably moved by the normative demands that blame expresses. Consequently, no account of blame that bases its justification merely on its efficacy can be adequate, because “it collides with one of the most obvious facts about blame, that in many cases it is effective only if the recipient thinks that it is justified” (1995c, 15).

At first pass, the conclusion that blame cannot wear its function on its sleeve seems subversive, calling into doubt my claim that Williams has a vindicatory account of blame in its generic form. But there is another way of reading the claim that blame cannot wear its function on its sleeve. Taking our cue from Truth and Truthfulness, we can see blame as a practice exhibiting what I call self-effacing functionality11: the practice of blame is functional, but only insofar as and because it is sustained by motives and reasons that are autonomous (i.e., not conditional on the practice’s functionality). As a result, the functionality of blame is effaced—it is not the primary consideration for participants as they engage in the practice, but for benign functional reasons: we reap the benefits of blame only if we are not benefitminded about it.12 11 See Queloz (2018).

12 In Queloz (2021b), I elaborate on this account of blame as a self-effacingly functional practice and on why it really is vindicatory and reflectively stable in a way that other instrumentalist or consequentialist accounts of blame are not.

[p. 195]

A Shelter from Luck This explains why blame’s functionality is not obvious to participants: that functionality itself requires them to be motivated by something other than the practice’s functionality. However, participants can become conscious of this without destabilizing the practice, because not only blame, but also the practice of thinking of it noninstrumentally are vindicated in terms that (unlike the sorts of vindications envisaged by indirect consequentialism) do not undermine the authority of the noninstrumental reasons for blame. In Truth and Truthfulness, the insight into the functionality of thinking in nonfunctional terms is offered as something that strengthens our confidence in the practice.13 For Williams, then, the four building blocks I have focused on are not the invention of the morality system. Nor are they inextricably linked to each other and to that system in a way that would preclude our making sense of them outside the system. All four building blocks can be made independently intelligible, because they have deeper roots in independent generic needs.

8.2 A Vindicatory Explanation of the Morality System With these vindicatory explanations of the building blocks in place, we can now turn to the system itself. To what end does it incorporate and reshape just these building blocks in the way it does? What concerns explain the “particular development of the ethical” (2011, 7) that is the morality system, and what is its point?14 Williams gives his most pointed answer to this question in “Moral Luck: A Postscript”: The point of this conception of morality is, in part, to provide a shelter against luck, one realm of value (indeed, of supreme value) that is defended against contingency. (1995e, 241) Our demand for a shelter from luck grows out of a longing for “ultimate justice” (2011, 43) or “fairness” (1995f, 75). In the face of the fact that 13 Nor does the fact that blame sometimes overstretches the idea that one had reason to act otherwise count against it, since even the fiction that one had reason to act otherwise has a valuable tendency to instill in the blamed just the sensitivity to reasons that it pretends they already possess (1995d, 41–44). On this “proleptic” function of blame, see Fricker (2016). 14 A related question is what drives the systematization of ethical thought that ultimately issues in ethical theory. See Cueni and Queloz (2021) for a reconstruction of Williams’s answer.

[p. 196]

“most advantages and admired characteristics are distributed in ways that, if not unjust, are at any rate not just, and some people are simply luckier than others,” the morality system expresses “the ideal that human existence can be ultimately just” because it offers a special kind of value, moral value, that outshines every other kind of value and “transcends luck” (2011, 217).15 On Williams’s view, the concern that has to be factored in to get from the four building blocks in their generic form to the moral/nonmoral distinction(MS), the idea of obligation(MS), the voluntary/involuntary distinction(MS), and the practice of blame(MS) is the longing for ultimate justice, which they serve by providing a shelter from luck. Is that longing for ultimate justice a more sociohistorically local concern than the generic human needs we considered earlier? Williams’s position on this is difficult to pin down. He describes the morality system as a historical phenomenon, connected to Platonism and Christianity, that finds its purest expression in the moral philosophy of Kant.16 But he is also sympathetic to the idea that the sense of fairness and the resentment of unfairness have deep naturalistic roots in the social character of our species and might even be innate.17 As Williams also writes, however, what desires we have depends on what we deem possible (1973a, 147), and perhaps the thought is that while the longing for fairness is old, the sense that ultimate fairness is possible is not—that, roughly speaking, had to wait for Plato, whose development of Pythagorean ideas provided the required dualism of soul and body and, in particular, the idea of a “featureless moral self ” (1993, 160).18 The Socratic dictum that “the good man cannot be harmed, since the only thing that could touch him would be something that could touch the good state of his soul” (2011, 39) articulates the animating idea of the morality system. It was also Plato who imposed a stark division between “rational concerns that aim at the good, and mere desire” (1993, 42), thereby providing the strategy of ethicizing psychology, as Williams calls it—the strategy of fitting psychological ideas to moral demands instead of trying to fit moral ideas into an antecedent 15 Many have since highlighted the importance of fairness as a motivation for the immunization of morality against luck; see Levy (2011, 9–10), Otsuka (2009, 374–75), and Sher (2005, 180). For a critical discussion of fairness-based arguments in favor of luck-free morality, see Hartman (2016, 2017). 16 See Williams (2014, 86). See Louden (2007) for an assessment of Williams’s equation of the morality system with Kant’s system which, barring some “subsidiary aspects,” finds it “substantially correct” (126–127). 17 See Williams (1999b, 248).

18 See Williams (2006d, 16).

[p. 197]

A Shelter from Luck understanding of human psychology.19 Thus, while the concerns driving the construction of a shelter from luck may be old, the construction itself had to wait for suitable material to come together, for only then could the ambition to “make the world safe for well-disposed people” (2006f, 59) gain a foothold.20 Refracted through Platonic ideas, the primitive concern with fairness was elaborated and focused into a demand for ultimate fairness. In this “strong form,” the concern to resist luck is one of the “idiosyncrasies” of the “local species of the ethical” (1995e, 242) that is the morality system. Once we see the system as organized by a concern to deliver ultimate justice by providing a shelter from luck, we can try to reverse-engineer the distinctive contributions of its components and explain—in terms that will be vindicatory relative to the demand for ultimate justice—their elaboration into the moral/nonmoral distinction(MS), the idea of obligation(MS), the voluntary/involuntary distinction(MS), and the practice of blame(MS). To this end, our guiding question must be: How does one construct a shelter from luck?

First, one needs a special kind of value that is not “merely a consolation prize you get if you are not in worldly terms happy or talented or goodhumoured or loved,” but the “supreme” form of value: it “has to be what ultimately matters” (2011, 217) if it is to eclipse any lack of luck in other respects. This is where the moral/nonmoral distinction(GEN) comes in. In its generic form, Williams agrees with Hume, the distinction is not very sharp.21 But it can be elaborated into a stark distinction to provide a special kind of value— moral(MS) value—whose importance can be dialed up to the point where it drowns out any other kind of value. The supremacy of moral(MS) value helps shut out luck by ensuring that misfortunes along other dimensions do not count. The most effective way of achieving this is for moral(MS) value to be supreme not just by carrying more weight than other kinds of value, but by forbidding comparison altogether. It is not that moral(MS) value ends up outweighing other kinds of value—there is not even a competition. This is what Williams refers to when he notes the tendency of the system to close in on itself, so that it comes to seem an “indecent misunderstanding” (2011, 19 In Williams’s terms, to ethicize psychology is “to provide a psychology that gets its significance from ethical categories” (1993, 43) or to define “the functions of the mind, especially with regard to action . . . at the most basic level in terms of categories that get their significance from ethics” (1993, 160). See also Williams (2006b, 78).

20 A phrase Williams used in the referenced passage to describe “the tireless aim of moral philosophy.”

21 See Williams (1995c, 20 n. 12) and Hume (1998, appendix IV).

[p. 198]

217) to ask, as Nietzsche did, what the value of that system is.22 The system’s “purism and its self-sufficiency mean that it is structured not to hear any considerations that might limit its own” (1995g, 204). From the point of view of the system, nothing outside the system really matters.

Second, one needs to ensure that the point of view of the system and the demands it makes on us are truly inescapable. The demands raised by the system therefore have to combine two aspects: ubiquity, to ensure that there is no domain where the demands of the system do not arise, and stringency, to ensure that these demands are forceful enough to take precedence over other demands. This is where the notion of obligation(GEN) comes in as the ideal format in which to couch moral thought. For reasons we considered, obligations are stringently demanding and designed to intrude into deliberation and impose themselves at the top of the priority list. Moreover, as Williams notes, “if obligation is allowed to structure ethical thought, there are several natural ways in which it can come to dominate life altogether” (2011, 202). If only an obligation can overrule an obligation, the felt need to resist a given obligation will invite one to look for ways in which the need to resist it can itself be rationalized as expressing a general obligation; and, the more this happens, the more general obligations multiply, so that they end up providing “work for idle hands” (2011, 202) across all aspects of life. The notion of moral obligation(MS) is thus perfectly suited to the task of ensuring that the system comes as close as possible to being truly inescapable. The proliferation of moral obligations(MS) means that there is always something that one is under an obligation to do, leaving no part of life free of the system’s influence, and the stringency of moral obligation(MS) ensures that what the system demands is what one really must do. The notion of obligation(GEN), once elaborated into the notion of obligation(MS), allows the demands of the morality system to become—in Kant’s term that usefully combines the two aspects of ubiquity and stringency (2011, 198)—categorical. Our construction thus far already looks well defended against contingency. Insofar as agents manage to stand in the right relation to their moral obligations, they will be living well in the only respect that ultimately matters: insofar as they manage to enter the shelter, they will be safe from 22 Nietzsche asks after “the value of morality,” urging that “we need a critique of moral values, for once the value of these values must itself be called into question” (1998, preface, §§5–6). Williams strikingly echoes Nietzsche when he notes that “the principal aim of all moral philosophy” is that of “truthfully understanding what our ethical values are and how they are related to our psychology, and making, in the light of that understanding, a valuation of those values” (1995a, 578).

[p. 199]

A Shelter from Luck luck. And everywhere, the shelter is there to be entered, because morality makes not just a claim on one, but the claim: what morality demands is what one really must do. We thus have a robust shelter which ubiquitously invites us in and promises to shut out the forces of contingency. But, though ubiquitous and contingency-proof, our shelter does not yet fully serve people’s concern to escape contingency, because it still suffers from an unequal access problem: some may find it easier than others to align their lives with the demands of the system. To eliminate contingency even here, entry to the shelter needs to be regulated in terms that guarantee equal access. Clearly, for instance, moral value must not be tied up with the consequences of actions, because this would render agents vulnerable to luck. As a Medieval proverb has it, when the flung stone leaves the hand, it belongs to the devil. To truly discount luck, moral value needs to retreat into the agent, to lie “in trying rather than succeeding, since success depends partly on luck” (2011, 217). Harnessing the emphasis on moral motivation which we noted in the moral/nonmoral distinction(GEN), the system is thus driven to focus exclusively on moral intentions and motives.23 Yet if the basis on which we allocate moral worth is to be ultimately just, we still need to eliminate various contingencies within the agent, because the “capacity to try,” or to act from moral motives, “is itself a matter of luck” (2011, 217). Various contingencies at the level of natural endowments, socialization, education, and other biographical and historical circumstances may make it easier for some to develop the right kind of motivation. Moral motivation must therefore itself be understood in terms that insist on purity from contingency. It must not be conditional on contingent desires or motives. It must be a form of motivation that the agent has anyway already— for instance, in virtue of being a rational agent.24 The requirement that makes itself felt here is that in order to guarantee equal access to the shelter from luck, the system must base itself solely on what any agent has complete control over no matter their circumstances. This is where the voluntary/involuntary distinction(GEN) comes in. Even in its generic form, it already does some of the work required by separating out what the agent did unintentionally or in an abnormal state of mind. But the 23 Attempts to render utilitarianism actionable issue in a notably similar structure by enjoining the agent to maximize expected rather than actual utility to the best of the agent’s knowledge: although actual utility is to a substantial degree a matter of luck, agents escape blame as long as they maximize expected utility (though see Monton [2019] for a discussion of exceptional cases). 24 Williams (1984) offers an exegetical reconstruction of Kant’s version of this idea.

[p. 200]

“search for an intrinsically just conception of responsibility” (1993, 95) leads one to push responsibility even further back, to a purified form of trying that is not conditioned by any disposition or desire or any other trait that the agent contingently has: what Williams calls “utter voluntariness” (2011, 218) or, following Kant, “the unconditioned will” (1981, 20). To be truly unconditioned, this purified conception of “the will” must not be in any way empirically determined by what the agent contingently is. The locus of the will in this demanding form cannot therefore be the socially situated and contingently constructed self. It has to be the featureless moral self that lies beyond all determination by empirical circumstances (in Kantian terms: the noumenal self). The intrinsically just basis on which moral responsibility(MS) is allocated in the system must be the unconditioned will of the characterless self. Williams acknowledges that the idea of the unconditioned will also has other roots rendering that aspect of the system “overdetermined” (1995g, 204).25 Williams notes, for example, that the “phenomenology of bodily movement and the notion of trying” (1999a, 149) already invite, via the observation that one can will a movement without it actually ensuing, the distinction between the self qua locus of action and the self qua locus of the will. He also remarks that we want there to be something over which we have complete control, and we want that because we feel the need for “real authorship of our actions” (1999a, 149). A further driving force is the resentment we feel when others wrong us. Williams here takes a leaf out of Nietzsche’s Genealogy (1998, I, §13), which describes the idea of a featureless moral agent—an agent lying beyond all determination by circumstance who can will to actualize his contingent dispositions or not—as being motivated by the felt need to blame not just the nature of things in general, but those who wronged us in particular. Connected to this is the human tendency to indulge in a “fantasy of retrospective prevention” (1995f, 73), where the victim fantasizes about replacing the wrongdoer’s action with an acknowledgment of the victim. This fantasy again motivates thinking of the agent in isolation from the network of circumstances, as an autonomous entity capable of willing to act otherwise than the agent in fact did.

Last, the system must also allocate moral blame in a way that shields the agent from luck. This is easily achieved at this point, since moral blame(MS) 25 The idea has a rich history in the theory of action; see Glock (1996, “will”), Hyman (2011, 2015), Candlish and Damnjanovic (2013), and Queloz (2017).

[p. 201]

A Shelter from Luck only needs to track utter voluntariness to ensure that the agent is blamed only “on the ultimately fair basis of the agent’s own contribution” (2011, 216). This is what Williams calls the “purified conception of blame” (1995f, 72). By tying blame to utter voluntariness, the system ensures that agents are blamed “for no more and no less than what is in [their] power” (1995f, 72). Given the moral demands on the will(MS) to align with obligations(MS), the purified conception of blame ties blame(MS) to the purely voluntary breaking of obligations. Because the system focuses on blame(MS) at the expense of other reactive attitudes and links blame(MS) to the purely voluntary breaking of obligations, “the thought I did it has no special significance” within the resulting picture of the ethical life; the only question is “whether I voluntarily did what I ought to have done” (2011, 196). This leads to a blinkered disregard for what Williams insists is an important dimension of ethical experience, namely “the distinction simply between what one has done and what one has not done” (2011, 196).26 The purified conception of blame comfortingly shields one from two kinds of blameworthiness that would otherwise render one vulnerable to luck: it shields one from being blamed for what one does involuntarily (this is what George Sher calls the “Searchlight View” (2009, ch. 1) of responsibility, on which agents are responsible only for those features and results of their acts of which they are aware when they act), and it shields one from being blamed when one does something as the lesser of two evils. Choices between wrong and wrong—the stuff of tragedy—lose their sting in the morality system, because if blameworthiness is tied to broken obligations, and if ought implies can—one can only be under an obligation to do what one can do—one is not blameworthy when one does something as the lesser of two evils. There might have been what W. D. Ross (1930) calls a prima facie obligation not to do what one ended up doing, but this obligation was eventually defeated by the consideration that the alternative would have been worse. Once the practice of blame(MS) is appropriately purified to be sensitive only to purely voluntary acts, our moral agents longing for ultimate justice are finally home and dry. The only thing that ultimately matters—moral(MS) value—is now completely within their control, for it depends only on whether they choose, from motives they all equally have anyway, to align their unconditioned will with their categorical obligations.

26 This is one of the main points of “Moral Luck” (1981). See also Williams (2011, 43–44).

[p. 202]

8.3 Critique of the Morality System: Frictionless Purity I take it that Williams is not just being sarcastic when he says of the morality system that it expresses a “moving” (2011, 217) ideal in its quest for ultimate justice. Moreover, the system appears well-tailored to its task. So what is wrong with it?

Williams finds numerous things wrong with it, and much of the criticism he levels at moral philosophy in general, or at Kantianism and utilitarianism in particular, applies also to the morality system. One line of criticism is that once our ethical thought has been subjected to the demands of coherence and systematicity to the degree required for it to become a system, that system leaves us with too few ethical thoughts and feelings to be true to ethical experience: like a color filter laid over the ethical landscape, it masks all but a few morally relevant features of it.27 Another line of criticism is that the morality system alienates us from our projects and hence from what sustains the possibility of a meaningful life—the system leaves no-one in particular for me to be (2011, 78, 224). In his critique of utilitarianism, Williams notes that this amounts, “in the most literal sense,” to an attack on my “integrity” (1973a, 116–117). What he means is not that some of my actions fail to fit in with my character as expressed in my other actions. It is rather that the very connection to what makes my various actions mine threatens to be lost. My integrity as an agent depends on the fact that my actions can be seen as actions that flow from the projects and convictions with which I am most closely identified. By asking me to step away from these projects and do whatever the utility calculus requires, utilitarianism alienates me as an agent from that which allows me to see my actions as mine. The resulting actions lack any unifying connection to the projects and convictions of a particular agent. In that sense, they are no-one’s actions in particular.

Yet I believe that the ultimate problem with the morality system, for Williams, is its frictionless purity: it robs valuable concepts of their grip on the kind of world we live in and, by insisting that true value and free agency be pure of any contamination by contingency, it threatens to engender nihilism about value and skepticism about agency. This critique has two strands: the No-Friction critique and the Purist-Attitude critique.

27 See Williams (2011, 130). For a valuable discussion of this point, see Krishna (2014). The idea that the demand for systematicity and coherence comes in degrees is explored in Cueni and Queloz (2021).

[p. 203]

A Shelter from Luck To illustrate the No-Friction critique, let us focus on blame(MS). The criticism then is that blame(MS) fails to recruit people into a shared ethical sensibility or to bring deviators back into it because it fails to get a grip on the kind of world we actually live in, where an agent’s character and circumstances largely lie outside that agent’s control. A conception of blameworthiness that tracks purely voluntary acts is “frictionless” or “fails to get a grip in the world we live in” insofar as its extension in the world we actually inhabit is empty, so that acts of the sort we do in fact encounter fail, upon closer scrutiny, to qualify as proper targets of blame. A conception of blame that is frictionless in this sense cannot properly do any work for us, for it lacks an empirical basis in the world we live in. This is a criticism we find already in Williams’s “The Idea of Equality” (first published in 1962), where he insists that “the concept of ‘moral agent,’ and the concepts allied to it such as that of responsibility, do and must have an empirical basis” (1973c, 235–236, emphasis added). To be concepts worth having, our concepts must allow us to make discriminations within the empirical world we live in rather than only between that world and something beyond it. This is not the case with blame(MS) and its attendant conception of voluntary(MS) action. To be voluntary(MS), actions must not reflect anything that agents involuntarily and contingently are, but this means that actions can only be voluntary(MS) insofar as agents have chosen all the circumstances that shape their lives. Needless to say, hardly any action will pass this test. We did not choose our circumstances—or, if we did, that choice was likely itself a reflection of prior circumstances we did not choose. Voluntariness cannot extend all the way back. Indeed, we could not have chosen our circumstances all the way through life, because at the beginning of this process, there would have to be the pure, characterless self envisaged by the morality system, and this characterless self would lack the basis to make such a choice: it would be too unencumbered by commitments and attachments to get an adequate view of the value of anything.28 If, as a matter of fact, the “machinery of everyday blame” (2011, 214) does any work for us, this is because it “attempts less than morality would like it to do” (2011, 215). We operate, much of the time, by the lights of blame(UND) rather than blame(MS). This is true more broadly: “If our modern ethical understanding does involve illusions, it keeps going at all only because it is supported by models of human behaviour that are more realistic than it acknowledges” (1993, 11). Blame(UND) works with a conception of 28 See Williams (1993, 158–159; 2011, ch. 6).

[p. 204]

voluntariness that is less demanding. In inquiring whether an action was voluntary in this undemanding sense, we typically seek to determine only whether people really acted, knew what they were doing, and intended such and such aspect of what they did (2011, 215–216). Many actions will meet this standard. But if we consistently allocated blame according to the demanding conception of voluntariness(MS), our blaming practices would cease to serve our need to align our ethical sensibilities and fail to discharge a function that we need to see discharged. The No-Friction critique of the morality system is thus that the purity of blame(MS) and its concomitant ideas robs them of their much-needed friction with the empirical world: too purified to achieve a grip on the rough ground we live on, they become pointless.

It is tempting to conclude that there is a simple remedy: we need only learn to accept that the requirements of utter voluntariness cannot be met. We need more truthfulness and knowledge about the world we inhabit so that we come to see that we are “building ethical life around an illusion” (2011, 212). But there is a more insidious problem here, which brings us to the PuristAttitude critique. For Williams, the problem with the morality system is not just that we fail to find actions that plausibly fall under its purified conceptions; the problem is also that these conceptions shape our attitude toward what we do find. The system encourages overblown normative expectations about what shape the world can properly have if value and agency are to have a place in it—the attitude that Williams labels its “purity” (2011, 216). This purist attitude “abstract[s] the moral consciousness from other kinds of emotional reaction or social influence” (2011, 216) and conceives of moral value as lying “beyond any empirical determination” (2011, 217). Its purism lies in its insistence on stark contrasts between the purity of moral values and free agency and the natural, emotional, and social forces pervading human life. “In truth,” however, “almost all worthwhile human life lies between the extremes that morality puts before us” (2011, 216). The demands that the system’s conceptions make on moral motivation and voluntariness cannot in fact be met. “This fact,” Williams writes, is known to almost everyone, and it is hard to see a long future for a system committed to denying it. But so long as morality itself remains, there is danger in admitting the fact, since the system itself leaves us, as the only contrast to rational blame, forms of persuasion it refuses to distinguish in spirit from force and constraint. (2011, 216)

[p. 205]

A Shelter from Luck This is where the Purist-Attitude critique proves crucial: there is danger in admitting to what extent contingency pervades human life as long as one remains attached to the morality system’s outlook, because that outlook blinds people to the forms of value and freedom that really are to be found in the world we live in.29 As a result, disenchanting our view of the world through truthful naturalistic enquiry risks making things worse rather than better. It risks exacerbating the sense that there is no room for moral value in a world thus understood, resulting in a nihilism that maintains that nothing has value. Moreover, because the system entrained an ethicized psychology reflected in our conceptions of free and rational agency as something that excludes the influence of mere desires and emotions, nihilism about value will be accompanied by skepticism about agency. This is why the system encounters the problems of free will and determinism “in a particularly acute form” (2011, 195). It “makes people think that, without its very special obligation, there is only inclination; without its utter voluntariness, there is only force; without its ultimately pure justice, there is no justice” (2011, 218). The result is a bleak and flattened vision of life that elides all difference between rational persuasion and manipulation, convincing and coercing, the force of the better reason and the force of a punch in the face. This danger, which was a central concern of Nietzsche’s,30 and to which Williams gave pride of place in the resounding final lines of Ethics and the Limits of Philosophy, shows that merely facing up to the world we live in—merely revising our beliefs—is not enough. The Purist-Attitude critique is therefore this: in the face of a naturalized view of the world, the attitudes cultivated in us by the morality system threaten to turn us into skeptics or nihilists who see no room for real value or real agency. Take the example of blame again. Once one admits that the demand for utter voluntariness cannot be met, there are two ways one can go: one can renounce the demanding conception of voluntariness in favor of a less demanding one and confidently allocate blame on that basis; this is the exit from the system that Williams recommends, and it is what Paul Russell (2017) calls the strategy of the pessimist. But it is at least equally tempting to reason in a different direction and to conclude, with the skeptic, that if the 29 The system “conceals” all the “options for ethical thought and experience” that there are outside itself—“Kantian associations constantly work to short-circuit our understanding” (1993, 77) of those other ways of making sense of things as valuable. This is why, from the perspective of morality, the “Greeks do emerge as premoral” (1993, 77).

30 See Clark (2015) and Queloz and Cueni (2019).

[p. 206]

demands of the system cannot be met, no act of blaming is ever truly justified, and all we are left with is people being coerced by their circumstances. Note that the difference between the pessimist and the skeptic is not a difference in knowledge. They both agree that no act is ever voluntary(MS) in the way that blame(MS) requires. It is just that while the pessimist takes this to speak against that conception of blame, the skeptic takes it to speak against the hope that blame might ever be justified. The skeptic thereby betrays a continued adherence to the system in concluding that since no act is ever voluntary(MS), no acceptable form of blame is ever justified—in much the same way that atheists betray a residual religiosity if they believe, with Ivan Karamazov, that since God does not exist, everything is permitted. The skeptic’s position involves a counterfactual adherence to the system which is structurally analogous to the “counterfactual scientism” (2006c, 187) that Williams accuses Hilary Putnam and Richard Rorty of: the skeptic believes that blame is in fact never justified, but that, if it were, this would have to be due to there being utterly voluntary acts.

The pivotal question that separates the skeptic from the pessimist is therefore this: What does the fact that luck and contingency pervade human life entail? In drawing from it the conclusion that nothing has value and no-one is free, one betrays a residual commitment to the system. One betrays a commitment to the purist pattern of reasoning encoded in the conceptions of that system—a pattern that notably licenses inferences such as the following: If anything has value, it is the moral value of things done from moral motives.

If an action is done from a moral motive, it is a voluntary action. If an action is voluntary, it is not conditioned by anything that is contingent or lies beyond the agent’s control.

Via the contrapositives of those claims, one quickly gets from the realization that every action is somehow conditioned by things that are contingent or lie beyond the agent’s control to the conclusion that no action is ever voluntary and nothing has value. But as the different pattern of reasoning exemplified by the pessimist shows, one might also take the same realization to entail nothing of the sort. Drawing on conceptions of voluntariness, moral motivation, and value that are more tolerant of contingency and draw contrasts within the empirical world, one can also endorse a pattern of reasoning which allows us to accept that no moral motivation is ever fully pure

[p. 207]

A Shelter from Luck of contingent desires, and no action ever fully pure of the influence of unchosen circumstance, and still recognize value and freedom in the world. To endorse the first of these two patterns rather than the second is not to fall prey to cognitive error; it is to evince a bad attitude, an attitude whose badness is ethical rather than cognitive. It is an attitude that does not help us to live. As Nietzsche would have put it, it is a life-denying attitude. It emerges that the journey out of the system involves not just the first, epistemic step of facing up to reality, but also the second, more radical step of liberating oneself from overblown normative expectations about just how pure of contingency the world would have to be in order to contain things of value and responsible agents. The conclusion that we are left with nothing turns out to depend on an overblown conception of what counts as something.31 In light of this, emancipation from the system can be thought of as involving three stages.32 At the initial stage, one is still immersed in the illusion that the “rationalistic metaphysics of morality” (1993, 159) correspond to something in reality. Through truthful reflection and inquiry, one then moves to a transitional stage, where one realizes that those metaphysics do not correspond to anything, but one retains the idea that they would have to correspond to something if the world were to contain true value and free agency. This entrains nihilism and skepticism. Finally, on being shown that there are other ways of making sense of values and agency in naturalistic but nonetheless vindicatory terms, one can move out of the system altogether. One “resituate[s] the original opposition[s] in a new space, so that the real differences can emerge” between contingent desires that are moral and contingent desires that are not, between conditioned actions that are voluntary and conditioned actions that are not, and “between the force which is argument and the force which is not—differences such as that between listening and being hit, a contrast that may vanish in the seminar but which reappears sharply when you are hit” (2002, 9). At this third stage, one is liberated from the system’s constraining conceptions and capable of affirming one’s values on grounds different from before. There is no guarantee that all of our ideas will survive a truthful understanding of them, but the threat of nihilism and skepticism will have been averted if some of them do. (Arguably, Ethics and the Limits of Philosophy, with its critique of the “philosophical errors” 31 Williams diagnoses an analogous problem in the view of some ethical theorists that if we give up on ethical theory, we are left with nothing (2011, 223). 32 A comparable schema sheds light on Nietzsche’s conception of the process by which European morality collapses in the wake of the “Death of God.” See Queloz and Cueni (2019).

[p. 208]

[2011, 218] involved in the system, primarily helps move its readers from the initial to the transitional stage, whereas Shame and Necessity and Truth and Truthfulness, with their naturalistic but vindicatory explanations of ideas of agency, responsibility, and intrinsic values like truthfulness, are more concerned to move their readers out of the transitional stage by giving them somewhere outside the system to stand.)

The morality system thus turns out to merit its name: it systematically harnesses and adapts to its own ends a variety of initially helpful ideas to hold out the ultimately illusory promise of a shelter from luck. Combining a vindicatory understanding of why we have these ideas with an initially vindicatory but ultimately critical understanding of why they take the form they do in the morality system can provide us with a nuanced sense of what we need them to do for us and what kind of friction with the world they need in order to do that. But it is not enough simply to admit that no action is ever voluntary in the sense demanded by the system because contingency and luck pervade life. We also need alternative ways of making sense of value and agency. Only then can we really throw off the powerful misconceptions which, by shaping our sense of how much contingency our ideas of value and agency can tolerate, determine what the pervasiveness of contingency entails.

REFERENCES

Candlish, Stewart, and Nic Damnjanovic. 2013. “Reason, Action and the Will: The Fall and Rise of Causalism.” In The Oxford Handbook of the History of Analytic Philosophy, edited by Michael Beaney, 689–708. Oxford: Oxford University Press.

Clark, Maudemarie. 2015. “On the Rejection of Morality: Bernard Williams’s Debt to Nietzsche.” In Nietzsche on Ethics and Politics, 41–61. Oxford: Oxford University Press.

Cueni, Damian, and Matthieu Queloz. 2021. “Whence the demand for ethical theory?” American Philosophical Quarterly 58 (2): 135–45.

Fricker, Miranda. 2016. “What’s the point of blame? A paradigm based explanation.” Noûs 50 (1): 165–183.

Glock, Hans-Johann. 1996. A Wittgenstein Dictionary. Oxford: Wiley-Blackwell.

Hartman, Robert. 2016. “Against luck-free moral responsibility.” Philosophical Studies 173 (10): 2845–2865.

Hartman, Robert. 2017. In Defense of Moral Luck: Why Luck Often Affects Praiseworthiness and Blameworthiness. New York: Routledge.

Hume, David. 1998. An Enquiry Concerning the Principles of Morals, edited by Tom L. Beauchamp. Oxford: Clarendon Press.

Hyman, John. 2011. “Action and the Will.” In The Oxford Handbook of Wittgenstein, edited by Marie McGinn and Oskari Kuusela, 451–471. Oxford: Oxford University Press.

Hyman, John. 2015. Action, Knowledge, and Will. Oxford: Oxford University Press.

[p. 209]

A Shelter from Luck

Kitcher, Philip. 2011. The Ethical Project. Cambridge, MA: Harvard University Press.

Krishna, Nakul 2014. “The Morality of Common Sense: Problems from Sidgwick.” PhD thesis. Oxford University.

Levy, Neil. 2011. Hard Luck: How Luck Undermines Free Will and Moral Responsibility. Oxford: Oxford University Press.

Louden, Robert B. 2007. “The Critique of the Morality System.” In Bernard Williams, edited by Alan Thomas, 104–134. Cambridge: Cambridge University Press.

Magee, Bryan, and Bernard Williams. 1971. “Conversation with Bernard Williams: Philosophy and Morals.” In Modern British Philosophy, edited by Bryan Magee, 150–165. New York: St. Martin’s Press. 19 (18): 1–25.

Nietzsche, Friedrich. 1998. On the Genealogy of Morality. Translated by Maudemarie Clark and Alan J. Swensen. Indianapolis and Cambridge: Hackett.

Nietzsche, Friedrich. 2001. The Gay Science. Translated by Josefine Nauckhoff and Adrian Del Caro, edited by Bernard Williams. Cambridge: Cambridge University Press.

Nietzsche, Friedrich. 2005. “Nietzsche Contra Wagner.” In The Anti-Christ, Ecce Homo, Twilight of the Idols, and Other Writings, edited by Aaron Ridley and Judith Norman, 263–282. Cambridge: Cambridge University Press.

Otsuka, Michael. 2009. “Moral luck: Optional, not brute.” Philosophical Perspectives 23 (1): 373–388.

Queloz, Matthieu. 2017. “Two orders of things: Wittgenstein on reasons and causes.” Philosophy 92 (3): 369–397.

Queloz, Matthieu. 2018. “Williams’s pragmatic genealogy and self-effacing functionality.”

Queloz, Matthieu. 2021a. The Practical Origins of Ideas: Genealogy as Conceptual ReverseEngineering. Oxford: Oxford University Press.

Queloz, Matthieu. 2021b. “The self-effacing functionality of blame.” Philosophical Studies 178 (4): 1361–1379.

Queloz, Matthieu, and Damian Cueni. 2019. “Nietzsche as a critic of genealogical debunking: Making room for naturalism without subversion.” The Monist 102 (3): 277–297.

Ross, W. D. 1930. The Right and the Good. Oxford: Clarendon Press.

Russell, Paul. 2017. “Free Will Pessimism.” In The Limits of Free Will: Selected Essays, 243– 276. Oxford: Oxford University Press.

Sher, George. 2005. “Kantian fairness.” Philosophical Issues 15 (1): 179–192.

Sher, George. 2009. Who Knew? Responsibility Without Awareness. New York: Oxford University Press.

Williams, Bernard. 1963. “Freedom and the Will.” In Freedom and the Will, edited by David Pears, 1–13. New York: Palgrave Macmillan.

Williams, Bernard. 1973a. “A Critique of Utilitarianism.” In Utilitarianism: For and Against, edited by John Jamieson Carswell Smart and Bernard Williams, 77–155. Cambridge: Cambridge University Press.

Williams, Bernard. 1973b. “Egoism and Altruism.” In Problems of the Self, 250–265. Cambridge: Cambridge University Press.

Williams, Bernard. 1973c. “The Idea of Equality.” In Problems of the Self, 230–249. Cambridge: Cambridge University Press.

Williams, Bernard. 1980. “Conclusion.” In Morality as a Biological Phenomenon: The Presuppositions of Sociobiological Research, edited by Gunther Stent, 275–285. Berkeley: University of California Press.

[p. 210]

Williams, Bernard. 1981. “Moral Luck.” In Moral Luck, 20–39. Cambridge: Cambridge University Press.

Williams, Bernard. 1984. “Präsuppositionen der Moralität.” In Bedingungen der Möglichkeit: “Transcendental Arguments” und Transzendentales Denken, edited by Eva Schaper and Wilhelm Vossenkuhl, 251–261. Stuttgart: Klett-Cotta.

Williams, Bernard. 1993. Shame and Necessity. Berkeley: University of California Press.

Williams, Bernard. 1995a. “Ethics.” In Philosophy: A Guide Through the Subject, edited by A. C. Grayling, 545–582. New York: Oxford University Press.

Williams, Bernard. 1995b. “Formal and Substantial Individualism.” In Making Sense of Humanity, 123–34. Cambridge: Cambridge University Press.

Williams, Bernard. 1995c. “How Free Does the Will Need to Be?” In Making Sense of Humanity, 3–21. Cambridge: Cambridge University Press.

Williams, Bernard. 1995d. “Internal Reasons and the Obscurity of Blame.” In Making Sense of Humanity, 35–45. Cambridge: Cambridge University Press.

Williams, Bernard. 1995e. “Moral Luck: A Postscript.” In Making Sense of Humanity, 241– 247. Cambridge: Cambridge University Press.

Williams, Bernard. 1995f. “Nietzsche’s Minimalist Moral Psychology.” In Making Sense of Humanity, 65–79. Cambridge: Cambridge University Press.

Williams, Bernard. 1995g. “Replies.” In World, Mind, and Ethics, edited by J. E. J. Altham and Ross Harrison, 185–224. Cambridge: Cambridge University Press.

Williams, Bernard. 1995h. “Voluntary Acts and Responsible Agents.” In Making Sense of Humanity, 22–34. Cambridge: Cambridge University Press.

Williams, Bernard. 1995i. “What Has Philosophy to Learn from Tort Law.” In The Philosophical Foundations of Tort Law, edited by David G. Owen, 487–498. Oxford: Oxford University Press.

Williams, Bernard. 1997. “Shame, Guilt, and the Structure of Punishment.” In Scham und Schuld: Festschrift aus Anlass der Verleihung der Dr. Margrit Egnér-Preise 1997 zum Thema Scham und Schuld, 21–30. Zürich: Margrit Egnér-Stiftung.

Williams, Bernard. 1999a. “Bernard Williams.” In Key Philosophers in Conversation: The Cogito Interviews, edited by Andrew Pyle, 142–163. London: Routledge.

Williams, Bernard. 1999b. “Seminar with Bernard Williams.” Ethical Perspectives 6 (34): 243–265.

Williams, Bernard. 2001. Morality: An Introduction to Ethics. Cambridge: Cambridge University Press.

Williams, Bernard. 2002. Truth and Truthfulness: An Essay in Genealogy. Princeton: Princeton University Press.

Williams, Bernard. 2006a. “The Legacy of Greek Philosophy.” In The Sense of the Past: Essays in the History of Philosophy, edited by Myles Burnyeat, 3–48. Princeton: Princeton University Press.

Williams, Bernard. 2006b. “Moral Responsibility and Political Freedom.” In Philosophy as a Humanistic Discipline, edited by A. W. Moore, 119–125. Princeton: Princeton University Press.

Williams, Bernard. 2006c. “Pagan Justice and Christian Love.” In The Sense of the Past: Essays in the History of Philosophy, edited by Myles Burnyeat, 71–82. Princeton: Princeton University Press.

Williams, Bernard. 2006d. “Philosophy as a Humanistic Discipline.” In Philosophy as a Humanistic Discipline, edited by Adrian W. Moore, 180–199. Princeton: Princeton University Press.

[p. 211]

A Shelter from Luck

Williams, Bernard. 2006e. “The Primacy of Dispositions.” In Philosophy as a Humanistic Discipline, edited by A. W. Moore, 67–75. Princeton: Princeton University Press.

Williams, Bernard. 2006f. “The Women of Trachis: Fictions, Pessimism, Ethics.” In The Sense of the Past: Essays in the History of Philosophy, edited by Myles Burnyeat, 49–59. Princeton: Princeton University Press.

Williams, Bernard. 2009. “A Mistrustful Animal.” In Conversations on Ethics, edited by Alex Voorhoeve, 195–214. Oxford: Oxford University Press.

Williams, Bernard. 2011. Ethics and the Limits of Philosophy. London and New York: Routledge.

Williams, Bernard. 2014. “After Virtue: A Study in Moral Theory, by Alasdair MacIntyre.” In Essays and Reviews 1959–2002, edited by Michael Woods, 84–86.