Two drivers leave a party equally drunk. Both choose to drive home. Both run the same red light. In one street, nobody is crossing; in the other, a child steps into the road. One driver reaches home shaken. The other kills someone and may spend years in prison. Their choices were morally alike up to the point where chance entered the scene. Our judgments usually are not.
That discomfort is the quickest route into the problem of moral luck. The problem is not that luck sometimes changes what happens. Everyone knows that. The harder claim is that luck appears to change what we are entitled to blame or praise a person for, even when the luck itself was outside that person’s control. The modern debate took its canonical form in the paired work of Bernard Williams and Thomas Nagel in the late 1970s. The Stanford Encyclopedia of Philosophy describes the puzzle as a collision between a control principle — we should assess people only for what is under their control — and ordinary moral practice, which repeatedly violates that principle.
THOUGHT EXPERIMENT: Hold intention and recklessness fixed. Change only the world’s response. If blame changes, you have already entered the problem of moral luck.
The control principle is attractive — until we use it consistently
Most moral systems contain some version of a control intuition. We distinguish murder from accidental death, deliberate cruelty from clumsiness, coercion from voluntary choice. Courts do it; everyday blame does it; philosophical theories do it. If an earthquake makes a promise impossible to keep, we do not normally treat the failure as if it were a voluntary betrayal. For a secondary map of the argument and its disputes, see PhilPapers — Moral Luck bibliography.
Nagel’s pressure point is that control shrinks as soon as we look closely. Temperament is partly inherited. The situations we face are not chosen from a menu. Outcomes depend on other people, weather, timing and institutions. Even the sequence of experiences that forms our practical judgment is heavily contingent. If responsibility must be purified of every influence of luck, there may be far less left to assess than moral life seems to require.
Nagel’s four kinds of moral luck
Resultant luck
How an action turns out: the reckless driver who happens to hit someone versus the equally reckless driver who does not.
Circumstantial luck
The situations one happens to face: whether a person is ever placed under conditions that reveal courage, cowardice, complicity or resistance.
Constitutive luck
Traits, dispositions and capacities a person did not fully choose: temperament, upbringing, sensitivity, impulsiveness.
Causal luck
The causal history leading to an action — a category that threatens to swallow the others if every choice is embedded in prior causes.
The taxonomy matters because it blocks an easy escape. It would be convenient to say that only outcomes are luck-dependent while character and intention remain safely “ours.” Nagel’s scheme shows why that boundary is unstable. Character is formed under conditions we did not select, and the occasions on which character becomes visible are also distributed unevenly.
Williams is asking a different question
Williams uses the painter Gauguin as a deliberately uncomfortable case. Imagine a person who abandons serious obligations in order to pursue an artistic project. Whether we can later understand — not necessarily excuse — that choice seems partly dependent on whether the project succeeds. Failure can retroactively make the decision look like self-deception; success can make the same risk intelligible as a life-defining commitment. Williams’s point is not that success magically erases duties. It is that the shape of a life cannot always be evaluated from a standpoint scrubbed clean of contingency.
This is where Williams’s argument is more existential than Nagel’s. Nagel worries that responsibility dissolves when control is examined too severely. Williams worries that a morality obsessed with immunity from luck misdescribes human agency. Our projects matter precisely because they are exposed to the world. A life whose value could never be changed by what actually happened would be strangely detached from action, history and commitment.
The debate is not permission to blame people for bad luck
Moral luck is sometimes misunderstood as a philosophical defense of hindsight punishment: if the outcome is worse, blame more. That is too quick. The debate instead asks which parts of our reactive practices are justified and which merely reflect emotional or legal convenience. The drunken-driver case can support a distinction between culpability for risk and additional responsibility for harm, but philosophers disagree about whether the second layer is genuinely moral or only legal, prudential, expressive or victim-centered.
A useful way to resist confusion is to separate four questions: What did the agent choose? What risk did the agent knowingly create? What harm actually occurred? What response — blame, punishment, repair, regret — is appropriate? Moral luck becomes most illuminating when these are not collapsed into one score.
Do not confuse moral luck with the free-will problem
Moral luck can arise even if we assume that an agent chose voluntarily and is ordinarily responsible for choosing. The free-will debate asks whether actions can be free enough for responsibility at all. Moral luck asks a different question: once agency is already in view, how far should praise and blame track outcomes, circumstances, character formation and causal histories that the agent did not control?
The distinction matters because the two debates can overlap without being identical. A compatibilist and an incompatibilist can still disagree about resultant or circumstantial luck; likewise, a theory can defend responsibility in general while remaining uneasy about why two equally reckless agents attract different judgments when only one causes harm.
Why this problem keeps returning
The issue appears anywhere outcomes outrun intentions: medical decisions, wartime command, investing, parenting, professional negligence, political leadership, historical judgment. It also complicates praise. A scientist may make a discovery partly because the right instrument, mentor or anomaly appeared at the right time. A rescuer may become heroic because chance placed danger within reach. If luck contaminates blame, it also complicates credit.
That symmetry is one reason the problem is durable. We want to hold people responsible without pretending that they authored the whole world in which they acted. We want moral judgment to track agency, but agency is always exercised through circumstances it did not create. Moral luck is the name for the friction between those two truths.
How to stage the disagreement when you read them
Read Nagel first if you want the architecture of the paradox. He begins from the control principle and expands the domain of luck until responsibility itself looks unstable. Read Williams next if you want to see why eliminating luck may be the wrong ambition. His interest is not only in blame but in the relation between identity, projects and the retrospective meaning of a life.
The strongest reading is not to choose a winner too quickly. Nagel makes the demand for consistency harder to ignore; Williams shows why a perfectly purified morality may purchase consistency by losing contact with the texture of action. Between them, moral luck stops being an exotic puzzle and becomes a question about what kind of beings moral agents actually are. For a book-length route through Williams, Nagel, and later responses, BookInLight’s moral luck reading guide extends the debate beyond this conceptual map.
How commentators frame the dispute
The argument is clearer when it is treated as a dispute in the secondary literature rather than as a sealed doctrine. First, Stanford Encyclopedia of Philosophy — Moral Luck maps the main positions, terminology, and bibliography. Second, Bernard Williams, “Moral Luck” — Cambridge Core provides a more specialized scholarly treatment and makes the interpretive stakes visible. Finally, Internet Encyclopedia of Philosophy — Moral Luck maps the main positions, terminology, and bibliography. Read together, these sources show which distinctions are standard, which reconstructions are later scholarly moves, and where apparently simple formulations become controversial.
That source-led method changes the tone of the analysis. It separates the primary argument from later reconstruction and from criticism of that reconstruction. Where commentators disagree about what follows from the canonical text, the disagreement is reported explicitly rather than smoothed into one voice. The result is closer to a guided report through an existing philosophical debate than to a self-contained theoretical essay.
Continue the argument on BookInLight
Primary texts and scholarship
Stanford Encyclopedia of Philosophy — Moral Luck — Updated overview of the control principle, the four forms of moral luck, and the Williams–Nagel debate.
Internet Encyclopedia of Philosophy — Moral Luck — Accessible survey of responsibility, luck and major positions in the literature.
Bernard Williams, “Moral Luck” — Cambridge Core — Publisher page for Williams’s classic essay in Moral Luck.
PhilPapers — Moral Luck bibliography — Research bibliography and map of the contemporary literature.
Mark B. Anderson, “Moral Luck as Moral Lack of Control” — PhilPapers — A later contribution clarifying the relation between luck and control.
