⁂ Concepts · Ethics · No. 133
Moral luck

Two drivers, one road, one evening, and two entirely different parts of the sentencing framework.

Moral luck

Judgement resting on factors outside anyone's control - which is supposed to be impossible and happens constantly.

Theme
Ethics and character
Tradition
Analytic
Term fixed
1976

In plain terms

Two drivers take the same risk. One arrives home. The other, through no difference in care or intention, hits a child who ran out.

We judge them differently. The second is a killer, will be prosecuted, and will regard themselves differently for the rest of their life. Nothing they did differed from the first driver in any way that was up to them.

That looks straightforwardly unfair, and the natural response is to say the judgement is confused. But almost nobody can hold that response steadily. The second driver cannot, and we would think worse of them if they could. Which leaves the uncomfortable position: our moral practice depends on factors we also believe should be irrelevant, and removing them would leave something we would not recognise as moral judgement.

As a philosopher would put it

Nagel's taxonomy distinguishes four kinds. Resultant luck in how actions turn out; circumstantial luck in the situations one faces; constitutive luck in the character and dispositions one has; and causal luck in how one is determined by antecedent circumstances.

The problem is that the control principle - that people are morally assessable only for what is within their control - is intuitively compelling and, applied consistently, eliminates almost everything we assess. Constitutive luck alone removes character from the picture, since nobody chose their temperament.

Williams's version is different from Nagel's and is often conflated with it. His interest is in retrospective justification: an agent who acts on a project and fails may be rationally subject to agent-regret that is not available to a spectator, and the rationality of a life's central choices can depend on outcomes not settled at the time. Kant is the target for both, since the good will is meant to be immune to fortune.

Where you meet it

Two drivers exceed the limit on the same road by the same margin, in the same conditions, on the same evening. One arrives without incident. The other strikes a pedestrian who steps out.

In England and Wales the first faces a fixed penalty and points. The second faces a charge of causing death by dangerous driving, with a maximum sentence that Parliament raised to life imprisonment in 2022. The sentencing guideline sets categories by culpability, and the difference in outcome moves the case into an entirely different part of the framework.

Every legal system does this, and every legal system has had the argument about whether it should. Criminal attempts are typically punished less severely than completed offences even where the failure was pure chance. Defenders argue that outcomes are what the law exists to prevent and that harm caused is a fact about the world; critics argue that punishing luck is punishing what nobody controlled. The practice has not changed in either direction.

What is actually disputed

Whether the control principle should be kept.

Rationalists hold that it should, and that our practice of judging by outcome is a mistake we could in principle correct - the luck is a distortion, and a fully reflective morality would assess only what was up to the agent. The cost is that assessment then applies to almost nothing, since character and circumstances are unchosen too.

Sceptics accept the elimination and conclude that moral responsibility as ordinarily understood does not survive.

Williams's position is that the practice is not correctable because morality itself is not the self-contained, luck-proof domain the tradition wanted. On his view the ambition to make ethics immune to fortune - so that the good person can never be harmed in what matters - is the deep error, running from Socrates to Kant, and moral luck is what shows it. That is a diagnosis of the whole subject rather than a solution to a puzzle, which is why the paper is still being argued about.

Commonly confused with

  • bad luck Ordinary misfortune affects what happens to you. Moral luck affects what you are judged for, which is a different and stranger claim.
  • blaming the victim Moral luck concerns the agent's assessability, not the sufferer's.
  • Determinism 1677 Causal luck is one of the four kinds and the free will debate is adjacent, but resultant and circumstantial luck arise whether or not determinism is true.

Defined against

  • Responsibility If assessment tracks what was not controlled, then either responsibility is not what it claims to be or the practice is unjust - and both options are hard to accept

Philosophers

  • Immanuel Kant The good will as immune to fortune, which is the position under attack
  • Aristotle External goods as necessary to a good life, an ancient acknowledgement of the problem
  • Friedrich Nietzsche The suspicion that moral judgement is not what it presents itself as
  • Marcus Aurelius The Stoic attempt to make virtue independent of circumstance

Works

Schools

Related concepts

Responsibility c. 340 BC · Free will c. 397 · Consequences 1789 · Virtue c. 340 BC · Agency c. 340 BC

Williams and Nagel published papers under this title in 1976; the underlying problem is Kantian, since Kant's insistence that the good will shines like a jewel regardless of outcome is what moral luck contradicts.