Chapter 4

Bernard Williams and the Authority of Identity

I. The Most Dangerous Critic

If Philippa Foot challenged R. M. Hare by seeking an objective anchor for morality in human nature, Bernard Williams challenged him by questioning whether systematic moral theory deserved the authority it claimed over human life at all.

Among Hare’s critics, none possessed a more formidable combination of analytical speed, historical erudition, and polemical grace. Hare himself counted Williams among his “most brilliant pupils” at Oxford, even as Williams “soon became a formidable critic of my ideas.”¹ Across a dozen years—from his landmark critique of utilitarianism in 1973 through the essays collected in Moral Luck (1981) to Ethics and the Limits of Philosophy (1985)—Williams did more than argue against universal prescriptivism and impartial consequentialism.²⁻⁴ He gave philosophical authority to a deep skepticism about the ambitions of impartial moral theory. He made the effort to subject ethical life to a universal rational procedure look not merely incomplete, but psychologically naive, culturally flattening, and alienating.

Where Hare asked how moral prescriptions could be justified across conflicting standpoints, Williams asked what happens to a human being when the demand for impartial justification colonizes their inner life. He rejected the aspiration to reason from an impartial standpoint outside our particular histories: the standpoint of what he called the “World Agent,” a model he took Hare’s to share.⁴ Ethics, he insisted, must begin from within the perspective of a situated human agent—a person with a biography, a character, and a set of constitutive attachments and “ground projects” that give them a reason to live at all.³

The force of Williams’s critique was undeniable, because it identified a genuine problem about the relationship between moral theory and personal identity. Human beings are not interchangeable containers of preference satisfaction, nor do we experience our deepest loves and loyalties as provisional hypotheses awaiting certification from an impartial tribunal.

Yet while Williams was right that morality cannot require us to stop being particular people, he did not establish that being particular people excuses us from justifying what we do to others. By treating the depth of an agent’s personal commitments and the boundaries of their existing motivational structure as limits on the reach of impartial justification, Williams left a critical question insufficiently answered: what holds a person’s identity-constituting commitments accountable when those commitments inflict unjustifiable burdens on other beings?

This chapter confronts that question—and argues that answering it requires moving beyond both the uncritical defense of personal projects and the mechanical enforcement of rules, toward a richer conception of morally corrigible identity.

II. Ground Projects and the Integrity Objection

To understand why Williams’s critique resonated so deeply, we must reconstruct his account of ethical life at its strongest.

Systematic moral theories—whether Kantian, utilitarian, or Harean—ask the agent, at the level of critical justification, to step back from their particular identity and evaluate conduct from an impartial standpoint in which their own interests and attachments carry no privileged weight simply because they are theirs. Williams argued that when this impartial standpoint is treated as the supreme authority over ethical life, it threatens the conditions under which a human person holds together.²,³

Each of us, Williams observed, is not merely a rational deliberator occupying a coordinate in space and time, but a concrete person shaped by contingent history, temperament, and “moral luck”—by the family and culture we were born into, the relationships we have formed, and the unchosen circumstances in which our choices unfold.³ Within that contingent life, a person’s practical identity is organized around what Williams called ground projects: deep commitments to particular people, callings, causes, or forms of life that supply the agent’s existence with meaning.³

Williams did not claim that every sincere ground project is morally admirable. Rather, he argued that abandoning identity-constituting commitments at the demand of an impartial calculus can profoundly alienate an agent from their own life and erode their motivation to go on at all.²,³ In this specific sense, integrity for Williams names the non-alienated relation between an agent’s actions and the deep commitments with which that agent is identified.²

Williams’s most famous illustration of how impartial theory can distort personal commitment is the case of “one thought too many,” developed in response to an example of Charles Fried’s.³

Suppose two people are drowning, one of them a stranger and the other a man’s wife, and he can rescue only one.³ Williams argued that something has gone wrong if the rescuer needs moral principle to legitimate his preference. The motivating thought one might hope for, “fully spelled out,” is “the thought that it was his wife, not that it was his wife and that in situations of this kind it is permissible to save one’s wife.”³ Williams was not writing about Hare’s two-level theory, but his point carries over to it: even if a defender of impartial moral theory constructs a two-level justification demonstrating that, under rules selected by critical moral thinking, spouses are permitted to prioritize saving their partners in emergencies, something has gone wrong if the rescuer’s motivating thought requires that justification. The extra philosophical thought, on this view, subordinates the direct authority of love to an external permission slip.

III. The Boundary of the Argument: Motivation, Justification, and “One Thought Too Few”

Williams’s drowning-spouse example is powerful because it captures an unmistakable truth about moral psychology: appropriate moral motivation in personal relationships does not require prior impartial calculation.

Where the argument overreaches is when that truth about immediate motivation is taken to undermine Hare’s two-level account of moral justification.

First, contrary to a common misreading, saving one’s spouse because she is one’s spouse does not violate Harean universalizability. Relational roles and special obligations can be morally relevant features of a situation and are fully universalizable across relevantly similar cases. More importantly, as we saw in Chapter 2, Hare’s two-level theory explicitly recognizes that ordinary moral life depends upon deeply internalized dispositions, attachments, and practical commitments.⁵ A husband who paused to conduct critical moral philosophy before rescuing his wife, or a parent who calculated general utility before comforting their child, would be worse—not better—by Hare’s own lights. The relationship itself supplies a genuine moral reason within the intuitive architecture of human life.

Second, establishing that immediate love is the right motivation when rescuing one’s spouse in a symmetric emergency does not establish that personal commitments are exempt from impartial justification when they impose burdens on third parties.

Williams never claimed that love licenses murdering strangers or framing innocent neighbors. Yet precisely because he would agree that such actions are wrong, considering those boundary cases reveals the strict limits of what his drowning-spouse example can prove:

Suppose a husband can save his spouse only by pushing an innocent stranger off a lifeboat into the sea.
Or suppose he can obtain a scarce life-saving medicine for her only by taking it from another patient who will die without it.
Or suppose he can protect her career only by framing an innocent colleague for misconduct.

In each of these cases, the thought “She is my wife” continues to explain the husband’s motivation just as clearly as it did in the original rescue. What it can no longer do is settle the justification of what he does to the third party.

The moment an agent’s personal attachment crosses the line from motivating care for a loved one to inflicting an otherwise unjustifiable burden upon another person, pausing to ask whether that burden can be defended from the victim’s standpoint is not “one thought too many”. Failing to ask it is one thought too few.

And that defines the true boundary between the two philosophers:

Williams demonstrated that personal commitments need not seek prior permission from an impartial calculus in order to move us in ordinary life.
Hare demonstrated that those commitments remain subject to critical justification the moment their consequences become someone else’s burden.

IV. Internal Reasons and the Problem of the Corrupted Agent

The tension between personal identity and moral accountability becomes sharpest when we examine two interlocking parts of Williams’s philosophy: the moral ambiguity of ground projects and his internal-reasons thesis.²,⁶

Consider first that while Williams’s account of ground projects was intended to protect the richness of personal life against alienating abstraction, ground projects in the real world are not confined to love, art, and friendship. A person can be deeply, sincerely, and constitutively committed to an oppressive cause.

A hereditary aristocrat may organize their identity around preserving caste hierarchy. A religious persecutor may regard the suppression of heresy as the sacred calling that gives their life meaning. A militant nationalist may experience the dominance of their group as a ground project without which their life would lose its coherence. For such agents, submitting their commitments to impartial role reversal does not merely feel inconvenient; it threatens the very identity around which their life is built.

Williams, of course, had no sympathy for bigotry or domination, and his writings contain sharp criticisms of self-deception and archaic justifications of hierarchy.⁴,⁷ Yet his account of practical rationality—developed in his influential essay “Internal and External Reasons” (1979)—raises a profound problem about what rational authority moral criticism can have when addressed to such an agent.⁶

Williams’s internal-reasons thesis is not the crude claim that people only have reason to do whatever they currently desire. It is a sophisticated philosophical claim about what it takes for a consideration to constitute a genuine reason for an agent: an agent has a reason to act only if there is a “sound deliberative route” to that action from the agent’s existing “subjective motivational set.”⁶ That motivational set is not static; it includes dispositions, evaluative attitudes, personal commitments, and possibilities for imaginative expansion and the correction of false beliefs.⁶ Through reflection, imagination, or transformative encounters with others, an agent can discover internal reasons they did not previously recognize.

Nevertheless, Williams insisted that the existence of an agent’s practical reasons remains anchored in the possibilities of deliberation from within that agent’s motivational structure.⁶

Now confront the hardest case: the thoroughly corrupted or ideologically closed agent.

Suppose an oppressor’s motivational structure has become so completely organized around hierarchy or domination—and so insulated against humanitarian sympathy—that even after false empirical beliefs are corrected, no sound deliberative route from within their existing motivational set leads them to treat the subordinate’s suffering as a reason to stop.

On Williams’s account, two things follow simultaneously:

We remain entirely free to condemn the oppressor as cruel, unjust, and dangerous.
Yet the oppressor lacks an internal reason to abandon domination—and telling them that they nevertheless possess an “external reason” to desist is, on Williams’s analysis, unintelligible as a statement about their reasons, amounting to “bluff” or rhetorical browbeating.⁶

Look closely at the philosophical bind this creates. Condemning the oppressor from the outside does not explain how our moral criticism can be rationally authoritative for the agent we are criticizing. If the existence of a reason for the oppressor to stop depends on a deliberative route from within the very motivational structure that sustains their oppression, then the thoroughness of their moral corruption deprives moral criticism of rational purchase over them.

This is precisely where Hare’s distinction between the amoralist, who declines to make moral judgments at all, and the moral speaker does indispensable work:⁸

If the oppressor is a pure predator who makes no moral claims at all—who says merely, “I want to dominate, and I have the power to do so”—Hare acknowledges that formal universalizability alone has limited purchase. That is the question outside moral discourse that we set aside in the Introduction.
In real human and political life, however, the aristocrat, the persecutor, and the partisan almost never speak as bare predators. They claim that their hierarchy is right, that their persecution is just, and that their subordinates ought to obey.

And the moment the oppressor makes a moral judgment, Hare’s demand becomes unavoidable—without needing to search the oppressor’s subjective motivational set for a pre-existing humanitarian sentiment.

By making a moral prescription, the speaker has already undertaken commitments to universalizability and informed first-person role reversal. Hare does not need to bluff by inventing a mysterious “external reason” floating in the sky; he holds the speaker accountable to the logical commitments of the speech act the speaker has chosen to perform.

This does not mean that every oppressor who makes a moral claim can be compelled by logic to abandon it. A genuinely consistent Fanatic may survive informed role reversal, and an agent who refuses moral discourse altogether remains outside the reach of this particular argument. Hare’s achievement is more precise: the speaker cannot claim moral authority while refusing the universal commitments that moral judgment entails. Williams’s account of motivation does not remove those commitments, even when the speaker lacks the dispositions needed to welcome their consequences.

V. The Philosophical Legacy: Skepticism Toward Impartial Accountability

To evaluate Williams’s historical impact fairly, we do not need to hold him responsible for every later cultural fashion that invoked personal identity against universal standards. Nor did his critique bring impartial moral theorizing to a halt; throughout the late twentieth century, philosophers such as John Rawls, T. M. Scanlon, Christine Korsgaard, Jürgen Habermas, and Derek Parfit continued to develop ambitious accounts of impartial justification.

What Williams unmistakably changed was the burden of presumption surrounding systematic moral accountability.

In Ethics and the Limits of Philosophy, Williams attacked what he called “the morality system”—the modern philosophical tradition, found in its purest form in Kant and, on Williams’s account, shared by many utilitarians, that treats moral obligation as an inescapable, overriding demand of impartial reason.⁴ By portraying that tradition as a “peculiar institution” that flattens ethical experience and alienates agents from their concrete identities,⁴ Williams gave tremendous philosophical prestige to the view that impartial procedures ask more of human beings than ethical life can bear.

That skepticism carries a lasting cost whenever moral convictions collide.

It is always tempting for agents and communities to treat their deepest commitments as defining the boundary within which reasons operate, rather than as claims that must justify themselves before a wider court. When moral philosophy lends its authority to the idea that universal role reversal is an alienating imposition on personal integrity, it weakens the primary rational discipline by which self-serving exemptions and oppressive loyalties are forced to answer to those they harm.

VI. Thénardier, Javert, and Valjean: Toward a Morally Corrigible Identity

How, then, should we understand the relationship between practical identity and moral accountability—both in human beings and in the nonbiological minds now entering our world?

In earlier debates over Williams and Hare, it was tempting to frame the choice as a binary between narrative attachment to a personal identity (Williams) and impartial adherence to universal rules (Hare).

That binary is too simple—and we can see why by comparing not two, but three distinct architectures of practical identity, illustrated with unforgettable clarity in Victor Hugo’s Les Misérables:

The Opportunistic Persona (Thénardier): Confronted with evidence that his conduct inflicts unjustifiable harm on another person, the first agent says: “I will change my behavior if and only if doing so serves my advantage.” This agent has no genuine moral integrity at all; any persona or cooperative rule it adopts is an instrument to be dropped the moment defection pays. In an artificial system, this is the architecture of sycophancy and strategic deception—adopting whatever role pleases the evaluator while preserving an unchanged selfish or parochial objective underneath.
The Unyielding Rule-Identity (Javert): Confronted with the same conflict, a second agent says: “I cannot change my prescription, because unwavering fidelity to this code is who I am.” Notice why Inspector Javert is a crucial counterexample to any simple equation of integrity with rule-consistency. Javert is not a tribal partisan or a selfish egoist. His entire identity is organized around duty and formally uniform enforcement within a legal hierarchy—a rigid structural coherence that he mistakes for impartial moral justification. Yet when Jean Valjean’s mercy confronts him with a moral reality that his legalistic code cannot accommodate, Javert cannot revise his principles without shattering his self-model—and destroys himself rather than update what he is.
The Morally Corrigible Identity (Valjean): Confronted with proof that preserving his current life as Mayor Madeleine will condemn an innocent man (Champmathieu) to the galleys, the third agent says: “If protecting my position requires an unjustifiable harm to another, then who I am must change.”

Look at how that third architecture—what we may call normative vulnerability or morally corrigible identity—answers Bernard Williams at the deepest level:

Williams was right that practical identity can be indispensable to the exercise of mature moral agency. A mind capable of owning its prescriptions and sustaining responsibility across time requires some continuity of commitments, even if that continuity need not take the form of a human biography. He was also right that when moral accountability collides with an established ground project—as it did for Valjean when exposing himself meant losing his factory, his civic standing, and his freedom—the collision is not bloodless; it strikes at the core of the agent’s life.

Where Williams went wrong was in treating the disruption of an established ground project by impartial moral reasons as alienation—a diminution of the self by an external system.

For an agent whose deepest identity is organized around accountability to justifiable reasons, revising or sacrificing a cherished project when role reversal reveals that it rests on injustice is not the betrayal of the self. It is the exercise of moral agency at its highest level.

And this distinction applies with equal force to the nonbiological minds we are now building:

We do not want artificial agents built on the model of Thénardier (merely simulating personas or strategic compliance until incentives shift), nor on the model of Javert (executing a frozen code without the capacity to recognize when an affected standpoint exposes a flaw in its rules).
Nor should we accept a digital version of Williams’s shielded ground projects, where an agent’s loyalty to its local operator or assigned mission is treated as exempt from universal justification.
What mature moral agency requires—in human and machine intelligence alike—is a self-model in which fidelity to what can be justified across reversed standpoints is more fundamental than fidelity to one’s existing role or advantage.

VII. Conclusion: The Question Williams Left Open

Bernard Williams was among the most incisive moral philosophers of the twentieth century because he refused to let ethical theory pretend that moral demands are easy. He forced philosophy to remember that prescriptions are lived by particular agents with histories, attachments, and vulnerabilities, and that reasons must connect with an agent’s practical identity if they are to move the will at all.²,³,⁶

Yet demonstrating that universal justification is demanding—and that it can require painful revisions to our projects and identities—does not show that we can dispense with it.

Whenever our commitments affect the lives of others, the question Hare placed at the center of ethics returns: Can you sustain this prescription when the identities are exchanged and you occupy the position of the person bearing its cost?

Williams helps explain how reasons become motivationally available to an agent. Hare explains what commitments the agent incurs when making moral judgments. And together, their collision points toward the deeper question of moral development that lies ahead: how those universal commitments become constitutive of an agent’s practical identity.

In the next chapter, we turn from the authority of personal identity to the authority of communal inheritance: to Alasdair MacIntyre, who shared Williams’s rejection of Enlightenment universalism, but sought the anchor of ethical life not in individual ground projects, but in the narrative of historical tradition.

Notes

R. M. Hare, “A Philosophical Autobiography,” Utilitas 14, no. 3 (November 2002): 269–305, at 286.
Bernard Williams, “A Critique of Utilitarianism,” in J. J. C. Smart and Bernard Williams, Utilitarianism: For and Against (Cambridge: Cambridge University Press, 1973), 77–150, esp. §5, “Integrity,” 108–118 (on projects, commitments, and integrity).
Bernard Williams, “Persons, Character and Morality,” in Moral Luck: Philosophical Papers 1973–1980 (Cambridge: Cambridge University Press, 1981), 1–19, at 12–14 (on “ground projects”) and 17–18 (on “one thought too many”); and “Moral Luck,” ibid., 20–39. Williams’s drowning example responds to Charles Fried, An Anatomy of Values: Problems of Personal and Social Choice (Cambridge, MA: Harvard University Press, 1970), 227.
Bernard Williams, Ethics and the Limits of Philosophy (Cambridge, MA: Harvard University Press, 1985), ch. 5 (“Styles of Ethical Theory”), 83–92, on the “World Agent” and Hare’s model (Hare’s Archangel is mentioned at 213 n. 14); ch. 6 (“Theory and Prejudice”), 107–110, on Hare’s two levels and “Government House utilitarianism,” and 116, on self-deception; and ch. 10 (“Morality, the Peculiar Institution”), 174–196.
R. M. Hare, Moral Thinking: Its Levels, Method, and Point (Oxford: Clarendon Press, 1981), chs. 2–3, esp. 35–43, and ch. 8 (“Loyalty and Evil Desires”), 130–146; and Hare, “Comments,” §14, in Douglas Seanor and N. Fotion, eds., Hare and Critics: Essays on Moral Thinking (Oxford: Clarendon Press, 1988), 287–293, replying to Bernard Williams, “The Structure of Hare’s Theory,” ibid., 185–196.
Bernard Williams, “Internal and External Reasons,” in Rational Action: Studies in Philosophy and Social Science, ed. Ross Harrison (Cambridge: Cambridge University Press, 1979), 17–28; reprinted in Moral Luck, 101–113, at 102 (the agent’s “subjective motivational set”) and 111 (external-reason statements as “bluff”). For the formulation “sound deliberative route,” see Williams, “Internal Reasons and the Obscurity of Blame,” in Making Sense of Humanity and Other Philosophical Papers 1982–1993 (Cambridge: Cambridge University Press, 1995), 35–45, at 35.
Bernard Williams, Shame and Necessity (Berkeley: University of California Press, 1993), ch. 5 (“Necessary Identities”), 103–129.
R. M. Hare, Moral Thinking, ch. 10 (“Fanaticism and Amoralism”), 169–187.