Moral Luck
Moral luck occurs when an agent is legitimately held morally responsible for something that turns in significant part on factors outside her control (see the SEP entry). The term was coined (paradoxically) by Bernard Williams and pressed philosophically by Thomas Nagel in a pair of 1976 essays.
A pure moral-luck move
X culture won Y war because they were morally superior to the losers.
That sentence, probably false, always suspicious, treats the outcome of contingent events as evidence of moral desert.
Why is this paradoxical?
If we hold people responsible for what they didn’t control, the Control Principle is wrong; if we don’t, half our moral practice is confused. The paradox is that we do both.
Hold both of these and you have a problem:
- Control Principle: people are morally assessable only for what’s within their control
- Practice: we actually judge people harder for worse outcomes, for being in testing circumstances, for temperaments they were born with, for all sorts of things they didn’t choose
Either the Control Principle is wrong, or half our moral practice is confused.
Nagel’s four kinds
- Resultant luck: how things turn out. Two drunk drivers weave home; one hits a child, one doesn’t. We blame them wildly differently. Their agency was identical
- Circumstantial luck: the situations you’re placed in. The Nazi collaborator in 1939 Berlin condemned himself through actions he’d never have been tested on in Buenos Aires. His counterpart there may have been morally identical but got away clean
- Constitutive luck: who you are. Your temperament, tendencies, patience, courage, intelligence, mostly products of genes, upbringing, environment. Yet we blame the coward for cowardly acts
- Causal luck: the free will problem, re-labeled. If your actions flow from prior causes you didn’t control, all moral assessment is luck-sensitive. This is the compatibilism question in another form
Williams's Gauguin case
A painter abandons his family for Tahiti. Whether his choice was justified turns, in part, on whether the resulting art is great. If he’d painted garbage, the decision was monstrous; if he painted masterpieces, something like vindication. The outcome is partly what settles the moral question, and the outcome is partly luck.
Williams introduced the term to attack a Kantian picture on which morality is supposed to be the one domain immune to luck, the place where the agent’s worth is sealed off from contingency. Williams thinks this picture is a fantasy. Actual ethical life is riddled with luck, and pretending otherwise is a kind of philosophical squeamishness.
Three responses (per SEP):
- Deny moral luck (Kantians, rationalists): stick with the Control Principle. Where our practices seem to track outcomes, reinterpret them: we’re using outcomes as evidence of intent, not as independent grounds for blame. The two drunk drivers are equally culpable; we just have more evidence on one
- Accept moral luck (Williams, Nagel): luck genuinely bears on moral assessment; the Control Principle is too strong. This is revisionary; we should stop being surprised that we judge outcomes, and drop the pretense that morality is contingency-proof
- Narrow the domain: distinguish moral responsibility (maybe luck-free) from broader moral assessment (luck-involved). Let the Control Principle govern the narrow thing and let everything else be ethical evaluation in Williams’s sense
My take
This is the problem that pushes hardest against my usual moves, because I default to the denial response: “same intent, same moral status; different outcomes are epistemic noise.” That’s the clean position and it’s how I’d want courts to work.
But Nagel’s constitutive-luck cut is the one I can’t shake. I didn’t choose my temperament. I didn’t choose to find certain things easy. The dispositions I credit myself for (persistence, curiosity, whatever) are constitutively lucky. If I strip away everything that wasn’t in my control, there’s almost nothing left to assess. The Control Principle, taken seriously, eats the entire subject of moral assessment.
So I end up roughly where Nagel does: the Control Principle is obviously right in the abstract, obviously wrong in practice, and both facts are stable. That’s not a solution, it’s a tension to live with. It does, I think, give a real argument for epistemic humility about judging other people: most of what I’d be judging is luck I happened to win, and if I’d lost it, I’d be the one being judged.
Closes the loop with compatibilism via causal luck: if determinism means my character was installed by prior causes, moral luck says that installation is morally load-bearing even though I didn’t do it. Both point at the same uncomfortable fact: most of what counts as “you” for moral purposes wasn’t chosen by you.