Thomas Nagel, Moral Luck1

[Immanuel] Kant believed that good or bad luck remain, this one seems to lose its footing. So a clear
should influence neither our moral judgment of a absence of control, produced by involuntary
person and his actions, nor his moral assessment of movement, physical force, or ignorance of the
himself. circumstances, excuses what is done from moral
The good will is not good because of what it judgment. But what we do depends in many more
effects or accomplishes or because of its adequacy ways than these on what is not under our control--
to achieve some proposed end; it is good only
because of its willing, i.e., it is good of itself And, what is not produced by a good or a bad will, in
regarded for itself, it is to be esteemed Kant's phrase. And external influences in this
incomparably higher than anything which could be broader range are not usually thought to excuse
brought about by it in favor of any inclination or what is done from moral judgment, positive or
even of the sum total of all inclinations. Even if it negative.
should happen that, by a particularly unfortunate
fate or by the niggardly provision of a step Let me give a few examples, beginning with
motherly nature, this will should be wholly lacking the type of case Kant has in mind. Whether we
in power to accomplish its purpose, and if even the succeed or fail in what we try to do nearly always
greatest effort should not avail it to achieve depends to some extent on factors beyond our
anything of its end, and if there remained only the control. This is true of murder, altruism, revolution,
good will (not as a mere wish but as the summoning
of all the means in our power), it would sparkle the sacrifice of certain interests for the sake of
like a jewel in its own right, as something that had others -- almost any morally important act. What
its full worth in itself Usefulness or fruitlessness can has been done, and what is morally judged, is
neither diminish nor augment this worth.2 partly determined by external factors. However
He would presumably have said the same jewel-like the good will may be in its own right,
about a bad will: whether it accomplishes its evil there is a morally significant difference between
purposes is morally irrelevant. And a course of rescuing someone from a burning building and
action that would be condemned if it had a bad dropping him from a twelfth-storey window while
outcome cannot be vindicated if by luck it turns out trying to rescue him. Similarly, there is a morally
well. There cannot be moral risk. This view seems significant difference between reckless driving
to be wrong, but it arises in response to a and manslaughter. But whether a reckless driver
fundamental problem about moral responsibility hits a pedestrian depends on the presence of the
to which we possess no satisfactory solution. pedestrian at the point where he recklessly passes
The problem develops out of the ordinary a red light. What we do is also limited by the
conditions of moral judgment. Prior to reflection it opportunities and choices with which we are faced,
is intuitively plausible that people cannot be and these are largely determined by factors
morally assessed for what is not their fault, or for beyond our control. Someone who was an officer
what is due to factors beyond their control. Such in a concentration camp might have led a quiet
judgment is different from the evaluation of and harmless life if the Nazis had never come to
something as a good or bad thing, or state of power in Germany. And someone who led a quiet
affairs. The latter may be present in addition to and harmless life in Argentina might have become
moral judgment, but when we blame someone for an officer in a concentration camp if he had not
his actions we are not merely saying it is bad that left Germany for business reasons in 1930.
they happened, or bad that he exists: we are I shall say more later about these and other
judging him, saying he is bad, which is different examples. I introduce them here to illustrate a
from his being a bad thing. This kind of judgment general point. Where a significant aspect of what
takes only a certain kind of object. Without being someone does depends on factors beyond his
able to explain exactly why, we feel that the control, yet we continue to treat him in that respect
appropriateness of moral assessment is easily as an object of moral judgment, it can be called
undermined by the discovery that the act or moral luck. Such luck can be good or bad. And the
attribute, no matter how good or bad, is not under problem posed by this phenomenon, which led
the person's control. While other evaluations
1 Thomas Nagel, “Moral Luck,” Proceedings of the Aristotelian Society , 50 (1976), 137–55.
2 Foundations of the Metaphysics of Morals, first section, third paragraph.
Kant to deny its possibility, is that the broad range intuitively acceptable conditions of moral
of external influences here identified seems on judgment threaten to undermine it all.
close examination to undermine moral assessment It resembles the situation in another area of
as surely as does the narrower range of familiar philosophy, the theory of knowledge. There too
excusing conditions. If the condition of control is conditions which seem perfectly natural, and which
consistently applied, it threatens to erode most of grow out of the ordinary procedures for
the moral assessments we find it natural to make. challenging and defending claims to knowledge,
The things for which people are morally judged threaten to undermine all such claims if consistently
are determined in more ways than we at first applied. Most skeptical arguments have this
realize by what is beyond their control. And when quality: they do not depend on the imposition of
the seemingly natural requirement of fault or arbitrarily stringent standards of knowledge,
responsibility is applied in light of these facts, it arrived at by misunderstanding, but appear to
leaves few pre-reflective moral judgments intact. grow inevitably from the consistent application of
Ultimately, nothing or almost nothing about what a ordinary standards.3 There is a substantive
person does seems to be under his control. parallel as well, for epistemological skepticism
Why not conclude, then, that the condition of arises from consideration of the respects in which
control is false--that it is an initially plausible our beliefs and their relation to reality depend on
hypothesis refuted by clear counter-examples? factors beyond our control. External and internal
One could in that case look instead for a more causes produce our beliefs. We may subject these
refined condition which picked out the kinds of lack processes to scrutiny in an effort to avoid error,
of control that really undermine certain moral but our conclusions at this next level also result, in
judgments, without yielding the unacceptable part, from influences which we do not control
conclusion derived from the broader condition, directly. The same will be true no matter how far
that most or all ordinary moral judgments are we carry the investigation. Our beliefs are always,
illegitimate. ultimately, due to factors outside our control, and
What rules out this escape is that we are the impossibility of encompassing those factors
dealing not with a theoretical conjecture but with without being at the mercy of others leads us to
a philosophical problem. The condition of control doubt whether we know anything. It looks as
does not suggest itself merely as a generalization though, if any of our beliefs are true, it is pure
from certain clear cases. It seems correct in the biological luck rather than knowledge.
further cases to which it is extended beyond the Moral luck is like this because while there are
original set. When we undermine moral assessment various respects in which the natural objects of
by considering new ways in which control is absent, moral assessment are out of our control or
we are not just discovering what would follow influenced by what is out of our control, we cannot
given the general hypothesis, but are actually reflect on these facts without losing our grip on the
being persuaded that in itself the absence of judgments…They are all opposed by the idea that
control is relevant in these cases too. The erosion one cannot be more culpable or estimable for
of moral judgment emerges not as the absurd anything than one is for that fraction of it which is
consequence of an over-simple theory, but as a under one's control. It seems irrational to take or
natural consequence of the ordinary idea of moral dispense credit or blame for matters over which a
assessment, when it is applied in view of a more person has no control, or for their influence on
complete and precise account of the facts. It would results over which he has partial control. Such
therefore be a mistake to argue from the things may create the conditions for action, but
unacceptability of the conclusions to the need for action can be judged only to the extent that it goes
a different account of the conditions of moral beyond these conditions and does not just result
responsibility. The view that moral luck is from them.
paradoxical is not a mistake, ethical or logical, but Let us first consider luck, good and bad, in the
a perception of one of the ways in which the way things turn out. Kant, in the above-quoted
passage, has one example of this in mind, but the

3 See Thompson Clarke, 'The Legacy of Skepticism', The

Journal of Philosophy, LXIX, no. 20 (November 9, 1972),
754--69.
category covers a wide range. It includes the truck assessment of the choice is possible in advance, but
driver who accidentally runs over a child, the artist another kind must await the outcome, because the
who abandons his wife and five children to devote outcome determines what has been done. The
himself to painting,4 and other cases in which the same degree of culpability or estimability in
possibilities of success and failure are even intention, motive, or concern is compatible with a
greater. The driver, if he is entirely without fault, wide range of judgments, positive or negative,
will feel terrible about his role in the event, but will depending on what happened beyond the point
not have to reproach himself Therefore this of decision. The mens rea which could have existed
example of agent-regret5 is not yet a case of in the absence of any consequences does not
moral bad luck. However, if the driver was guilty exhaust the grounds of moral judgment. Actual
of even a minor degree of negligence -- failing to results influence culpability or esteem in a large
have his brakes checked recently, for example -- class of unquestionably ethical cases ranging from
then if that negligence contributes to the death of negligence through political choice. That these are
the child, he will not merely feel terrible. He will genuine moral judgments rather than expressions
blame himself for the death. And what makes this of temporary attitude is evident from the fact that
an example of moral luck is that he would have to one can say in advance how the moral verdict will
blame himself only slightly for the negligence itself depend on the results… Perhaps it is true that what
if no situation arose which required him to brake is done depends on more than the agent's state of
suddenly and violently to avoid hitting a child. Yet mind or intention. The problem then is, why is it not
the negligence is the same in both cases, and the irrational to base moral assessment on what
driver has no control over whether a child will run people do, in this broad sense? It amounts to
into his path. holding them responsible for the contributions of
The same is true at higher levels of negligence. fate as well as for their own - provided they have
If someone has had too much to drink and his car made some contribution to begin with. If we look
swerves on to the sidewalk, he can count himself at cases of negligence or attempt, the pattern
morally lucky if there are no pedestrians in its seems to be that overall culpability corresponds to
path. If there were, he would be to blame for their the product of mental or intentional fault and the
deaths, and would probably be prosecuted for seriousness of the outcome. Cases of decision under
manslaughter. But if he hurts no one, although his uncertainty are less easily explained in this way,
recklessness is exactly the same, he is guilty of a for it seems that the overall judgment can even
far less serious legal offence and will certainly shift from positive to negative depending on the
reproach himself and be reproached by others outcome. But here too it seems rational to subtract
much less severely. To take another legal example, the effects of occurrences subsequent to the choice,
the penalty for attempted murder is less than that that were merely possible at the time, and
for successful murder - however similar the concentrate moral assessment on the actual
intentions and motives of the assailant may be in decision in light of the probabilities. If the object
the two cases. His degree of culpability can of moral judgment is the person, then to hold him
depend, it would seem, on whether the victim accountable for what he has done in the broader
happened to be wearing a bullet-proof vest, or sense is akin to strict liability, which may have its
whether a bird flew into the path of the bullet - legal uses but seems irrational as a moral position.
matters beyond his control… The result of such a line of thought is to pare
In many cases of difficult choice the outcome down each act to its morally essential core, an
cannot be foreseen with certainty. One kind of

4 Such a case, modelled on the life of Gauguin, is Gauguin to justify himself to others, but still determines his
discussed by Bernard Williams in 'Moral Luck' Proceediugs most basic feelings, that shows only that his most basic
of the Aristotelian Society, supplementary vol. L (1976), feelings need not be moral. It does not show that morality is
115-35 (to which the original version of this essay was a subject to luck. If the retrospective judgment were moral, it
reply). He points out that though success or failure cannot be would imply the truth of a hypothetical judgment made in
predicted in advance, Gauguin's most basic retrospective advance, of the form 'If I leave my family and become a
feelings about the decision will be determined by the great painter, I will be justified by success; if I don't become
development of his talent. My disagreement with Williams is a great painter, the act will be unforgivable.'
that his account fails to explain why such retrospective 5 Williams' term (ibid.).

attitudes can be called moral. If success does not permit

inner act of pure will assessed by motive and how it can seem necessary to subtract from it
intention… anything that merely happens - even though the
The area of genuine agency, and therefore of ultimate consequence of such subtraction is that
legitimate moral judgment, seems to shrink under nothing remains. And something in the ordinary
this scrutiny to an extensionless point. Everything idea of knowledge must explain why it seems to
seems to result from the combined influence of be undermined by any influences on belief not
factors, antecedent and posterior to action, that within the control of the subject - so that knowledge
are not within the agent's control. Since he cannot seems impossible without an impossible foundation
be responsible for them, he cannot be responsible in autonomous reason. But let us leave
for their results -- though it may remain possible to epistemology aside and concentrate on action,
take up the aesthetic or other evaluative character, and moral assessment.
anacogues of the moral attitudes that are thus The problem arises, I believe, because the self
displaced. which acts and is the object of moral judgment is
It is also possible, of course, to brazen it out threatened with dissolution by the absorption of its
and refuse to accept the results, which indeed acts and impulses into the class of events. Moral
seem unacceptable as soon as we stop thinking judgment of a person is judgment not of what
about the arguments. Admittedly, if certain happens to him, but of him. It does not say merely
surrounding circumstances had been different, then that a certain event or state of affairs is fortunate
no unfortunate consequences would have followed or unfortunate or even terrible. It is not an
from a wicked intention, and no seriously culpable evaluation of a state of the world, or of an
act would have been performed; but since the individual as part of the world. We are not
circumstances were not different, and the agent in thinking just that it would be better if he were
fact succeeded in perpetrating a particularly cruel different, or did not exist, or had not done some
murder, that is what he did, and that is what he is of the things he has done. We are judging him,
responsible for. Similarly, we may admit that if rather than his existence or characteristics. The
certain antecedent circumstances had been effect of concentrating on the influence of what is
different, the agent would never have developed not under his control is to make this responsible self
into the sort of person who would do such a thing; seem to disappear, swallowed up by the order of
but since he did develop (as the inevitable result mere events.
of those antecedent circumstances) into the Sort of What, however, do we have in mind that a
swine he is, and into the person who committed such person must be to be the object of these moral
a murder, that is what he is blameable for. In both attitudes? While the concept of agency is easily
cases one is responsible for what one actually does undermined, it is very difficult to give it a positive
- even if what one actually does depends in characterization. This is familiar from the literature
important ways on what is not within one's control. on free will.
This compatibilist account of Our moral judgments I believe that in a sense the problem has no
would leave room for the ordinary conditions of solution, because something in the idea of agency
responsibility - the absence of coercion, ignorance, is incompatible with actions being events, or
or involuntary movement - as part of the people being things. But as the external
determination of what someone has done - but it is determinants of what someone has done are
understood not to exclude the influence of a great gradually exposed, in their effect on
deal that he has not done.6 consequences, character, and choice itself, it
The only thing wrong with this solution is its becomes gradually clear that actions are events
failure to explain how skeptical problems arise. and people things. Eventually nothing remains
For they arise not from the imposition of an which can be ascribed to the responsible self, and
arbitrary external requirement, but from the we are left with nothing but a portion of the larger
nature of moral judgment itself. Something in the sequence of events, which can be deplored or
ordinary idea of what someone does must explain celebrated, but not blamed or praised.

6 The corresponding position in epistemology would be the correctness of these beliefs and the process by which they
that knowledge consists of true beliefs formed in certain are arrived at would therefore be importantly subject to
ways, and that it does not require all aspects of the process luck. The Nobel Prize is not awarded to people who turn out
to be under the knower's control, actually or potentially. Both to be wrong, no matter how brilliant their reasoning.
 Though I cannot define the idea of the active appearance that determinism obliterates
self that is thus undermined, it is possible to say responsibility. Once we see an aspect of what we
something about its sources. There is a close or someone else does as something that happens,
connexion between our feelings about ourselves we lose our grip on the idea that it has been done
and our feelings about others. Guilt and and that we can judge the doer and not just the
indignation, shame and contempt, pride and happening. This explains why the absence of
admiration are internal and external sides of the determinism is no more hospitable to the concept
same moral attitudes. We are unable to view of agency than is its presence -- a point that has
ourselves simply as portions of the world, and from been noticed often. Either way the act is viewed
inside we have a rough idea of the boundary externally, as part of the course of events.
between what is us and what is not, what we do The problem of moral luck cannot be
and what happens to us, what is our personality understood without an account of the internal
and what is an accidental handicap. We apply the conception of agency and its special connection
same essentially internal conception of the self to with the moral attitudes as opposed to other types
others. About ourselves we feel pride, shame, guilt, of value. I do not have such an account. The degree
remorse -- and agent-regret. We do not regard to which the problem has a solution can be
our actions and our characters merely as fortunate determined only by seeing whether in some
or unfortunate episodes -- though they may also degree the incompatibility between this
be that. We cannot simply take an external conception and the various ways in which we do
evaluative view of ourselves --of what we most not control what' we do is only apparent. I have
essentially are and what we do. And this remains nothing to offer on that topic either. But it is not
true even when we have seen that we are not enough to say merely that our basic moral
responsible for our own existence, or our nature, attitudes toward ourselves and others are
or the choices we have to make, or the determined by what is actual; for they are also
circumstances that give our acts the consequences threatened by the sources of that actuality, and by
they have. Those acts remain ours and we remain the external view of action which forces itself on us
ourselves, despite the persuasiveness of the when we see how everything we do belongs to a
reasons that seem to argue us out of existence. world that we have not created.
It is this internal view that we extend to others
in moral judgment - when we judge them rather
than their desirability or utility. We extend to
others the refusal to limit ourselves to external
evaluation, and we accord to themselves like our
own. But in both cases this comes up against the
brutal inclusion of humans and everything about
them in a world from which they cannot be
separated and of which they are nothing but
contents. The external view forces itself on us at
the same time that we resist it. One way this occurs
is through the gradual erosion of what we do by
the subtraction of what happens.7
The inclusion of consequences in the conception
of what we have done is an acknowledgment that
we are parts of the world, but the paradoxical
character of moral luck which emerges from this
acknowledgment shows that we are unable to
operate with such a view, for it leaves us with no
one to be. The same thing is revealed in the

7 See P. F. Strawson's discussion of the conflict between Thought and Action, ed. P. F. Strawson (London: Oxford!
the objective attitude and personal reactive attitudes in University Press, 1968), and in P. F. Strawson, Freedom and
'Freedom and Resentment", Proceedings of the British Resentment and Other Essays (London: Methuen, 1974).
Academy, 1962, reprinted in Studies in the Philosophy of