The Significance of Choice: T T L H V

The Significance of Choice
T . M . SCANLON, JR.
T HE T ANNER LECTURES ON H UMAN VALUES
Delivered at
Brasenose College, Oxford University
May 16, 23, and 28, 1986

T. M. SCANLON is Professor of Philosophy at Harvard
University. He was educated at Princeton, Brasenose
College, Oxford, and Harvard, and taught philosophy at
Princeton from 1966 until 1984. Professor Scanlon is the
author of a number of articles in moral and political
philosophy and was one of the founding editors of
Philosophy and Public Af fairs.
Lecture 1
1. INTRODUCTION
Choice has obvious and immediate moral significance. The
fact that a certain action or outcome resulted from an agents
choice can make a crucial difference both to our moral appraisal
of that agent and to our assessment of the rights and obligations
of the agent and others after the action has been performed. My
aim in these lectures is to investigate the nature and basis of this
significance. The explanation which I will offer will be based
upon a contractualist account of morality that is, a theory
according to which an act is right if it would be required or
allowed by principles which no one, suitably motivated, could
reasonably reject as a basis for informed, unforced general
agreement.1
I believe that it is possible within this general theory of
morality to explain the significance of various familiar moral
notions such as rights, welfare, and responsibility in a way that
preserves their apparent independence rather than reducing all of
them to one master concept such as utility. The present lectures
are an attempt to carry out this project for the notions of responsi-
bility and choice.
This is a revised version of three lectures presented at Brasenose College,
Oxford, on May 16, 23, and 28, 1986. I am grateful to the participants in the
seminars following those lectures for their challenging and instructive comments.
These lectures are the descendants of a paper, entitled Freedom of the Will in
Political Theory, which I delivered at a meeting of the Washington, D.C., Area
Philosophy Club in November 1977. Since that time I have presented many inter-
vening versions to various audiences. I am indebted to members of those audiences
and to numerous other friends for comments, criticism, and helpful suggestions.
1I have set out my version of contractualism in Contractualism and Utili-
tarianism, in Amartya Sen and Bernard Williams, eds., Utilitarianism and Beyond
(Cambridge: Cambridge University Press, 1982), pp. 103-28. What follows can be
seen as an attempt to fulfill, for the case of choice, the promissory remarks made at
the end of section III of that paper.
[151]
152 The Tanner Lectures on Human Values
2. THE PROBLEMS OF FREE WILL

Quite apart from this general theoretical project, however,
there is another, more familiar reason for inquiring into the basis
of the moral significance of choice. This is the desire to under-
stand and respond to the challenge to that significance which has
gone under the heading of the problem of free will. This problem
has a number of forms. One form identifies free will with a per-
sons freedom to act otherwise than he or she in fact did or will.
The problem, on this view, is the threat to this freedom posed by
deterministic conceptions of the universe. A second, related prob-
lem is whether determinism, if true, would deprive us of the kind
of freedom, whatever it may be, which is presupposed by moral
praise and blame. This version of the problem is closer to my
present concern in that it has an explicitly moral dimension. In
order to address it one needs to find out what the relevant kind
of freedom is, and this question can be approached by asking what
gives free choice and free action their special moral significance.
Given an answer to this question, which is the one I am primarily
concerned with, we can then ask how the lack of freedom would
threaten this significance and what kinds of unfreedom would
do so.
The challenge I have in mind, however, is not posed by deter-
minism but by what I call the Causal Thesis. This is the thesis
that the events which are human actions, thoughts, and decisions
are linked to antecedent events by causal laws as deterministic as
those governing other goings-on in the universe. According to
this thesis, given antecedent conditions and the laws of nature,
the occurrence of an act of a specific kind follows, either with
certainty or with a certain degree of probability, the indeterminacy
being due to chance factors of the sort involved in other natural
processes. I am concerned with this thesis rather than with deter-
minism because it seems to me that the space opened up by the
falsity of determinism would be relevant to morality only if it
[S CANLON] The Significance of Choice 153
were filled by something other than the cumulative effects of

indeterministic physical processes, If the actions we perform result
from the fact that we have a certain physical constitution and
have been subjected to certain outside influences, then an apparent
threat to morality remains, even if the links between these causes
and their effects are not deterministic.
The idea that there is such a threat is sometimes supported by
thought experiments such as the following: Suppose you were to
learn that someones present state of mind, intentions, and actions
were produced in him or her a few minutes ago by the action of
outside forces, for example by electrical stimulation of the nervous
system. You would not think it appropriate to blame that person
for what he or she does under such conditions. But if the Causal
Thesis is true then all of our actions are like this. The only dif-
ferences are in the form of outside intervention and the span of
time over which it occurs, but surely these are not essential to the
freedom of the agent.
How might this challenge be answered? One strategy would
be to argue that there are mistakes in the loose and naive idea of
causality to which the challenge appeals or in the assumptions it
makes about the relation between mental and physical events.
There is obviously much to be said on both of these topics. I pro-
pose, however, to follow a different (but equally familiar) line.
Leaving the concepts of cause and action more or less unanalyzed,
I will argue that the apparent force of the challenge rests on mis-
taken ideas about the nature of moral blame and responsibility.2
2In his admirably clear and detailed defense of incompatibilism, Peter van
Inwagen observes that if one accepts the premises of his argument for the incom-
patibility of determinism and free will (in the sense required for moral responsi-
bility) then it is puzzling how people could have the kind of freedom required
for moral responsibility even under indeterministic universal causation. (See An
Essay on Free Will [Oxford: Oxford University Press, 1983], pp. 149-50.) O n the
other hand, he takes it to be not merely puzzling but inconceivable that free will
should be impossible or that the premises of his arguments for incompatibilism
should be false or that the rules of inference which these arguments employ should
be invalid. This leads him, after some further argument, to reject determinism:
If incompatibilism is true, then either determinism or the free-will thesis is false.
It has sometimes been maintained that even if the Causal

Thesis holds, this does not represent the kind of unfreedom that
excuses agents from moral blame. That kind of unfreedom, it is
sometimes said, is specified simply by the excusing conditions
which we generally recognize: a person is acting unfreely in the
relevant sense only if he or she is acting under posthypnotic sug-
gestion, or under duress, is insane, or falls under some other gen-
erally recognized excusing condition. Since the Causal Thesis does
not imply that people are always acting under one or another of
these conditions, it does not imply that moral praise and blame
are generally inapplicable.
I am inclined to think that there is something right about this
reaffirmation of common sense. But in this simple form it has
been rightly rejected as question begging. It begs the question
because it does not take account of the claim that commonsense
morality itself holds that people cannot be blamed for what they
do when their behavior is the result of outside causes, a claim
which is supported by our reactions to imaginary cases like the
thought experiment mentioned above and by more general reflec-
tion on what a world of universal causality would be like.
In order to show that moral praise and blame are compatible
with the Causal Thesis, it is necessary to rebut this claim. The
most promising strategy for doing so is to look for a general
account of the moral significance of choice, an account which, on
To deny the free-will thesis is to deny the existence of moral responsibility, which
would be absurd. Moreover, there seems to be no good reason to accept determinism
(which, it should be recalled, is not the same as the Principle of Universal Causa-
tion). Therefore, we should reject determinism (p. 2 2 3 ) .
My response is somewhat different. Determinism is a very general empirical
thesis. Our convictions about moral responsibility seem to me an odd basis for
drawing a conclusion one way or the other about such a claim. In addition, what-
ever one may decide about determinism, it remains puzzling how moral responsi-
bility could be compatible with Universal Causation. I am thus led to wonder
whether our initial assumptions about the kind of freedom required by moral
responsibility might not be mistaken. Rather than starting with a reinterpretation
of the principle of alternative possibilities (along the lines of the conditional analy-
sis), my strategy is to ask first, Why does the fact of choice matter morally? and
then, What kind of freedom is relevant to mattering in that way?
[SCANLON] The Significance of Choice 155
the one hand, explains why the significance of choice is under-

mined both by commonly recognized excusing conditions and by
factors such as those imagined to be at work in the thought experi-
ment described above and, on the other hand, explains why the
moral significance of choice will not be undermined everywhere
if the Causal Thesis is true. Such an account, if convincing, would
provide a basis for arguing that our initial response to the Causal
Thesis was mistaken. At the very least, it would shift the burden
of argument to the incompatibilist, who would need to explain
why the proffered account of the moral significance of choice was
inadequate. Before beginning my search for an account of the
significance of choice, however, I will take a moment to examine
some other forms of the free-will problem.
The problem of free will is most often discussed as a problem
about moral responsibility, but essentially the same problem arises
in other forms as well. It arises in political philosophy, for ex-
ample, as a problem about the significance of choice as a legitimating
condition. We generally think that the fact that the affected
parties chose or assented to an outcome is an important factor in
making that outcome legitimate. But we also recognize that there
are conditions under which acquiescence does not have this legiti-
mating force. These include conditions like those listed above:
hypnosis, brain stimulation, mental incapacity, brainwashing, and
so on. To many, at least, it seems plausible to maintain that these
conditions deprive choice of its moral significance because they are
conditions under which the agents action is the result of outside
causes. But if the Causal Thesis holds, this is true of all actions,
and it would follow that choice never has moral significance as a
legitimating factor.
Let me turn to a different example, drawn from John Rawlss
book, A Theory of Justice. 3 (I believe the example involves a
misinterpretation of Rawls, albeit a fairly natural one, but I will
3A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1971),
pp. 72-74, 104.

try to correct that later.) Replying to an argument for the justice

of a purely laissez-faire economy, Rawls observes that in such a
system economic rewards would be unacceptably dependent on
factors such as innate talents and fortunate family circumstances,
which are, as he puts it, arbitrary from a moral point of view.
In particular, he says that even such factors as willingness to exert
oneself will depend to a large extent on family circumstances and
upbringing. Therefore we cannot say, of those who might have
improved their economic position if they had exerted themselves,
that because their predicament is their own doing they have no
legitimate complaint. Their lack of exertion has no legitimating
force because it is the result of arbitrary factors.
But this argument, if successful, would seem to prove too
much. Consider a society satisfying Rawlss Difference Principle.
This principle permits some inequalities, such as those resulting
from incentives which improve productivity enough to make every-
one better off. When such inequalities exist, they will be due to
the fact that some people have responded to these incentives while
others have not. If the Causal Thesis is correct, however, there
will be some causal explanation of these differences in behavior.
They will not be due to gross differences in economic status, since,
by hypothesis, these do not exist. But they must be due to some-
thing, and it seems clear that the factors responsible, whatever
they are, are likely to be as morally arbitrary in at least one
sense of that phrase as the factors at work in the case of the laissez-
faire society to which Rawls was objecting. To sustain Rawlss
argument, then, we need a better explanation of how morally
arbitrary background conditions can undermine the legitimating
force of choice, an explanation which will not deprive all choice
of moral force if the Causal Thesis is correct.
Let me mention a further, slightly different case. We think it
important that a political system should, as we say, leave people
free to make up their own minds, especially about important
political questions and questions of personal values. We regard
certain conditions as incompatible with this important freedom

and therefore to be avoided. Brainwashing is one extreme ex-
ample, but there are also more moderate, and more common, forms
of manipulation, such as strict control of sources of information,
bombardment with one-sided information, and the creation of an
environment in which people are distracted from certain questions
by fear or other competing stimuli. What is it that is bad about
these conditions? If they count as conditions of unfreedom simply
because they are conditions under which peoples opinions are
causal products of outside factors, then there is no such thing as
freedom of thought if the Causal Thesis is correct. It would
follow that defenders of freedom of thought who accept the
Causal Thesis could rightly be accused of ideological blindness :
what they advocate as freedom is really just determination by a
different set of outside factors, factors which are less rational and
no more benign than those to which they object. There may be
good reasons to favor some determining factors over others, but
the issue cannot be one of freedom. Here again, then, the prob-
lem is to show that determination by outside causes is not a
sufficient condition for unfreedom. To do this we need to come
up with some other explanation of what is bad about the condi-
tions which supporters of freedom of thought condemn.4
These are versions of what I will call the political problem of
free will. As I have said, they have much the same structure as the
more frequently discussed problem about moral praise and blame.
In addition to these problems there is what might be called the
personal problem of free will. If I were to learn that one of my
past actions was the result of hypnosis or brain stimulation, I
would feel alienated from this act: manipulated, trapped, reduced
to the status of a puppet. But why, if the Causal Thesis is correct,
should we not feel this way about all of our acts? Why should
4I have said more about this version of the problem in section IIB of Free-
dom of Expression and Categories of Expression, University of Pittsburgh Law

Review 40 (1979).
we not feel trapped all the time? This is like the other problems
in that what we need in order to answer it is a better explanation
of why it is proper to feel trapped and alienated from our own
actions in cases like hypnosis, an explanation which goes beyond
the mere fact of determination by outside factors. But while this
problem is like the others in its form, it differs from them in not
being specifically a problem about morality: the significance with
which it deals is not moral significance. This makes it a particu-
larly difficult problem, much of the difficulty being that of explain-
ing what the desired but threatened form of significance is sup-
posed to be. Since my concern is with moral theory I will not
address this problem directly, though the discussion of the value
of choice in lecture 2 will have some bearing on it.
I will be concerned in these lectures with the first two of these
problems and with the relation between them: to what degree can
the better explanation that each calls for be provided within the
compass of a single, reasonably unified theory? My strategy is to
put forward two theories which attempt to explain why the con-
ditions which we commonly recognize as undermining the moral
significance of choice in various contexts should have this effect.
These theories, which I will refer to as the Quality of Will theory
and the Value of Choice theory, are similar to the theories put
forward in two famous articles, P. F. Strawsons Freedom and
Resentment,5 and H. L. A. Harts Legal Responsibility and
Excuses.6 My aim is to see whether versions of these two
approaches extended in some respects and modified in others
to fit within the contractualist theory I espouse can be put
together into a single coherent account. We can then see how far
this combined theory takes us toward providing a satisfactory
account of the moral significance of choice across the range of
cases I have listed above.
5In Strawson, ed., Studies in the Philosophy of Thought and Action (Oxford:
Oxford University Press, 1968), pp. 71-96.

6 Chapter 2 of Hart, Punishment and Responsibility (Oxford: Oxford Univer-
sity Press, 1968).
3. THE INFLUENCEABILITY THEORY

Before presenting the Quality of Will theory, it will be help-
ful to consider briefly an older view which serves as a useful
benchmark. This view, which I will call the Influenceability theory,
employs a familiar strategy for explaining conditions which excuse
a person from moral blame.7 This strategy is first to identify the
purpose or rationale of moral praise and blame and then to show
that this rationale fails when the standard excusing conditions are
present. According to the Influenceability theory, the purpose of
moral praise and blame is to influence peoples behavior. There is
thus no point in praising or blaming agents who are not (or were
not) susceptible to being influenced by moral suasion, and it is
this fact which is reflected in the commonly recognized excusing
conditions.
The difficulties with this theory are, I think, well known.8 I
will not go into them here except to make two brief points. The
first is that the theory appears to conflate the question of whether
moral judgment is applicable and the question of whether it
should be expressed (in particular, expressed to the agent). The
second point is that difficulties arise for the theory when it is
asked whether what matters is influenceability at or shortly before
the time of action or influenceability at the (later) time when
moral judgment is being expressed. The utilitarian rationale
for praise and blame supports the latter interpretation, but it is
the former which retains a tie with commonsense notions of
responsibility.
7See J. J. C. Smart, Freewill, Praise, and Blame, Mind 70 (1961) : 291-
306; reprinted in G. Dworkin, ed., Determinism, Free Will, and Moral Responsi-
bility (Englewood Cliffs, N.J.: Prentice-Hall, 1970; page references will be to this
edition). The theory was stated earlier by Moritz Schlick in chapter 7 of The Prob -
lems o f Ethics, trans. D. Rynin (New York: Prentice-Hall, 1939), reprinted as
When Is a Man Responsible? in B. Berofsky, ed., Free Will and Determinism
(New York: Harper and Row, 1966; page references will be to this edition).
8Some are set forth by Jonathan Bennett in section 6 of Accountability, in
Zak van Staaten, ed., Philosophical Subjects (Oxford: Oxford University Press,
1980).
The Influenceability theory might explain why a utilitarian

system of behavior control would include something like what we
now recognize as excusing conditions. What some proponents of
the theory have had in mind is that commonsense notions of re-
sponsibility should be given up and replaced by such a utilitarian
practice. Whatever the merits of this proposal, however, it is
clear that the Influenceability theory does not provide a satis-
factory account of the notions of moral praiseworthiness and
blameworthiness as we now understand them. The usefulness
of administering praise or blame depends on too many factors
other than the nature of the act in question for there ever to be a
good fit between the idea of influenceability and the idea of
responsibility which we now employ.9
4. QUALITY OF WILL: STRAWSON'S ACCOUNT

The view which Strawson presents in Freedom and Resent-
ment is clearly superior to the Influenceability theory. Like that
theory, however, it focuses less on the cognitive content of moral
judgments than on what people are doing in making them. The
centerpiece of Strawsons analysis is the idea of a reactive attitude.
It is the nature of these attitudes that they are reactions not simply
to what happens to us or to others but rather to the attitudes
toward ourselves or others which are revealed in an agents
actions. For example, when you tread on my blistered toes, I may
feel excruciating pain and greatly regret that my toes were stepped
on. In addition, however, I am likely to resent the malevolence
or callousness or indifference to my pain which your action indi-
cates. This resentment is what Strawson calls a personal reactive
attitude: it is my attitudinal reaction to the attitude toward me
which is revealed in your action. Moral indignation, on the other
9Broadening the theory to take into account the possibility of influencing
people other than the agent will produce a better fit in some cases, but at the price of
introducing even more considerations which are intuitively irrelevant to the ques-
tion of responsibility.
hand, is what he calls a vicarious attitude: a reaction to the

attitude toward others in general (e.g., lack of concern about
their pain) which your action shows you to have. All of these
are what Strawson calls participant attitudes. They belong
to involvement or participation with others in inter-personal
human relationships.10 This is in contrast to objective atti-
tudes, which involve seeing a person as an object of social
policy; as an object for what in a wide range of senses might be
called treatment; as something certainly to be taken account, per-
haps precautionary account, of; to be managed or handled or cured
or trained.11
It follows from this characterization that the discovery of new
facts about an action or an agent can lead to the modification or
withdrawal of a reactive attitude in at least three ways: (a) by
showing that the action was not, after all, indicative of the agents
attitude toward ourselves or others; (b) by showing that the atti-
tude indicated in the act was not one which makes a certain reac-
tive attitude appropriate; (c) by leading us to see the agent as
someone toward whom objective, rather than participant, atti-
tudes are appropriate.
Commonly recognized excusing conditions work in these ways.
The most extreme excusing conditions sever any connection be-
tween an action (or movement) and the attitudes of the agent.
If your stepping on my toes was a mere bodily movement resulting
from an epileptic seizure, then it shows nothing at all about your
concern or lack of concern about my pain. It would therefore be
inappropriate for me to resent your action or for someone else,
taking a more impartial view, to feel moral disapproval of you on
that account.
Other excusing conditions have the less extreme effect of
modifying the quality of will which an action can be taken to
indicate, thus modifying the reactive attitudes which are appropri-
10Strawson, Freedom and Resentment, p. 79.
11
Ibid.
ate. If I learn, for example, that you stepped on my foot by acci-

dent, then I can no longer resent your callousness or malevolence,
but I may still, if conditions are right, resent your carelessness. If
I learn that you (reasonably) believed that the toy spider on my
boot was real, and that you were saving my life by killing it before
it could bite me, then I can no longer resent your action at all,
although it remains indicative of a particular quality of will on
your part.
Actions produced by posthypnotic suggestion are a less clear
case. Much depends on what we take the hypnosis to do. Hypno-
sis might lead you to perform the intentional act of stamping your
foot on mine but without any malice or even any thought that you
are causing me harm. In this case a criticizable attitude is indi-
cated by your act: a kind of complacency toward touching other
peoples bodies in ways that you have reason to believe are un-
wanted. But this attitude is not really attributable to you. You
may not lack any inhibition in this regard: it is just that your
normal inhibition has been inhibited by the hypnotist. The case is
similar if the hypnotist implants in you a passing hatred for me
and a fleeting but intense desire to cause me pain. Here again
there is a criticizable attitude more serious this time but it is
not yours. It is just visiting, so to speak.
Strawsons account of why conditions such as insanity and
extreme immaturity excuse people from moral blame is less satis-
factory. The central idea is that these conditions lead us to take an
objective attitude toward a person rather than to see him or her
as a participant in those interpersonal human relationships of
which the reactive attitudes are a part. Strawsons claim here can
be understood on two levels. On the one hand there is the empiri-
cal claim that when we see someone as warped or deranged,
neurotic or just a child . . . all our reactive attitudes tend to be pro-
foundly modified.l2 In addition to this, however, there is the
12Ibid. My appreciation of this straightforwardly factual reading of Strawsons
argument was aided by Jonathan Bennetts perceptive analysis in Accountability.

suggestion that these factors render reactive attitudes such as

resentment and indignation inappropriate. But Strawsons theory
does not explain the grounds of this form of inappropriateness as
clearly as it explained the grounds of the other excusing condi-
tions. In fact, aside from the references to interpersonal relation-
ships, which are left unspecified, nothing is said on this point.
In other cases, however, Strawsons theory succeeds in giving
a better explanation of commonly recognized excusing conditions
than that offered by the idea that a person is not to be blamed for
an action which is the result of outside causes. The mere fact of
causal determination seems to have little to do with the most
common forms of excuse, such as accident and mistake of fact.
It is a distinct advantage of Strawsons analysis that it accounts
for the force of more extreme excuses such as hypnosis and brain
stimulation in a way that is continuous with a natural explanation
of these less extreme cases as well. Moreover, his theory can ex-
plain the relevance of inability to do otherwise in several senses
of that phrase. Sometimes, as in the case of brain stimulation, the
factors which underlie this inability sever any connection between
an action and the agents attitudes. In other cases, inability to do
otherwise in the different sense of lack of eligible alternatives
can modify the quality of will indicated by an agents willingness
to choose a particular course of action. For example, if you stamp
on my toes because my archenemy, who is holding your child
hostage next door, has ordered you to do so, this does not make
you less responsible for your act. The act is still fully yours, but the
quality of will which it indicates on your part is not blameworthy.
As Strawson observes, these appeals to inability to do other-
wise do not generalize. The truth of the Causal Thesis would
not mean that either of these forms of inability obtained generally
or that actions never indicated the presence in the agent of those
attitudes or qualities of will which make resentment or moral
indignation appropriate.
Like the unsuccessful defense of common sense mentioned

above, Strawsons analysis is internal to our moral concepts as we
now understand them. Its explanation of the conditions which
negate or modify moral responsibility rests on a claim that, given
the kind of thing that moral indignation is, it is an appropriate
response only to actions which manifest certain attitudes on the
part of the agent. This internal character may be thought to be a
weakness in Strawsons account, and he himself considers an
objection of this sort. The objection might be put as follows:
You have shown what is and is not appropriate given the moral
notions we now have; but the question is whether, if the Causal
Thesis is correct, it would not be irrational to go on using those
concepts and holding the attitudes they describe. Strawsons direct
response to this objection is to say that the change proposed is
practically inconceivable.
The human commitment to participation in ordinary interper-
sonal relationships is, I think, too thoroughgoing and deeply
rooted for us to take seriously the thought that a general con-
viction might so change our world that, in it, there were no
longer any such things as inter-personal relationships as we
normally understand them; and being involved in inter-
personal relationships as we normally understand them pre-
cisely is being exposed to the range of reactive attitudes and
feelings that is in question.13
But there is another reply which is suggested by something

that Strawson goes on to say and which seems to me much
stronger.14 This reply points out that the principle If your action
was a causal consequence of prior factors outside your control
then you cannot properly be praised or blamed for performing it
derives its strength from its claim to be supported by commonsense
morality. Consequently, if an analysis such as Strawsons succeeds
13 Strawson, Freedom and Resentment, p. 82.

14 Ibid., p. 83,
in giving a convincing account of the requirements of freedom

implicit in our ordinary moral views in particular, giving a
systematic explanation of why commonly recognized excusing
conditions should excuse then this is success enough. Succeed-
ing this far undermines the incompatibilist challenge by striking
at its supposed basis in everyday moral thought.I5
Plausible and appealing though it is, there are several respects
in which Strawsons analysis is not fully satisfactory. One of these
has already been mentioned in connection with insanity. Straw-
son suggests that the attitudes which moral judgments express are
appropriately held only toward people who are participants in
certain interpersonal relationships and that these attitudes are
therefore inhibited when we become aware of conditions which
render a person unfit for these relationships. But one needs to
know more about what these relationships are, about why moral
reactive attitudes depend on them, and about how these relation-
ships are undermined or ruled out by factors such as insanity.
A second problem is more general. Strawson explains why
certain kinds of unfreedom make moral praise and blame in-
applicable by appealing to a fact about interpersonal reactive atti-
tudes in general (and moral ones in particular), namely the fact
that they are attitudes toward the attitudes of others, as mani-
fested in their actions. But one may wonder whether anything
further can be said about why attitudes of moral approval and dis-
approval are of this general type. Moreover, it is not clear that
moral judgments need always involve the expression of any par-
15Compare Thomas Nagels comments on Strawsons theory in The View from
Nowhere (Oxford: Oxford University Press, 1986), pp, 124-26. The response I
am advocating here does not deny the possibility of what Nagel has called ex-
ternal criticism of our practices of moral evaluation. It tries only to deny the
incompatibilist critique a foothold in our ordinary ideas of moral responsibility. It
claims that a commitment to freedom which is incompatible with the Causal Thesis
is not embedded in our ordinary moral practices in the way in which a commitment
to objectivity which outruns our experience is embedded in the content of our ordi-
nary empirical beliefs. The incompatibilist response, obviously, is to deny this claim.
My point is that the ensuing argument, which I am trying to advance one side of,
is internal to the system of our ordinary moral beliefs.
ticular reactive attitude. For example, I may believe that an action

of a friend, to whom many horrible things have recently hap-
pened, is morally blameworthy. But need this belief, or its expres-
sion, involve a feeling or expression of moral indignation or dis-
approval on my part? Might I not agree that what he did was
wrong but be incapable of feeling disapproval toward him?
Here Strawsons analysis faces a version of one of the objec-
tions to the Influenceability theory: it links the content of a moral
judgment too closely to one of the things that may be done in
expressing that judgment. Of course, Strawson need not claim
that moral judgment always involves the expression of a reactive
attitude. It would be enough to say that such a judgment always
makes some attitude (e.g., disapproval) appropriate. But then
one wonders what the content of this underlying judgment is and
whether the requirement of freedom is not to be explained by
appeal to this content rather than to the attitudes which it makes
appropriate.
In order to answer these questions one needs a more complete
account of moral blameworthiness. A number of different moral
theories might be called upon for this purpose, but what I will do
is to sketch briefly how a Quality of Will theory might be based
on a contractualist account of moral judgment.
5. QUALITY OF WILL: A CONTRACTUALIST ANALYSIS

According to contractualism as I understand it, the basic moral
motivation is a desire to regulate ones behavior according to
standards that others could not reasonably reject insofar as they,
too, were looking for a common set of practical principles.
Morality, on this view, is what might be called a system of
co-deliberation. Moral reasoning is an attempt to work out prin-
ciples which each of us could be expected to employ as a basis
for deliberation and to accept as a basis for criticism. To believe
that one is morally at fault is just to believe that one has not
regulated ones behavior in the way that such standards would
require. This can be so either because one has failed to attend

to considerations that such standards would require one to take
account of or because one has consciously acted contrary to what
such standards would require. If one is concerned, as most people
are to at least some extent, to be able to justify ones actions to
others on grounds they could not reasonably reject, then the
realization that one has failed in these ways will normally produce
an attitude of serious self-reproach. But this attitude is distinct
from the belief which may give rise to it. Similarly, to believe that
another persons behavior is morally faulty is, at base, to believe
that there is a divergence of this kind between the way that person
regulated his or her behavior and the kind of self-regulation that
mutually acceptable standards would require. For reasons like
those just mentioned, this belief will normally be the basis for
attitudes of disapproval and indignation. This view of morality
grounds the fact that moral appraisal is essentially concerned with
the quality of an agents will in an account of the nature of
moral reasoning and moral motivation. The analysis of moral
judgment which it supports is essentially cognitivist. It can explain
why moral judgments would normally be accompanied by certain
attitudes, but these attitudes are not the basis of its account of
moral judgment.
Contractualism also gives specific content to the idea, sug-
gested by Strawson, that moral judgments presuppose a form of
interpersonal relationship. On this view, moral judgments apply
to people considered as possible participants in a system of co-
deliberation. Moral praise and blame can thus be rendered inap-
plicable by abnormalities which make this kind of participation
impossible. (The implications of this idea for excusing conditions
such as insanity will be discussed below.)
6. THE SPECIAL FORCE OF MORAL JUDGEMENT

Insofar as it goes beyond Strawsons theory in committing itself
to a fuller account of the nature of moral blameworthiness, the
contractualist view I have described leaves itself open to the objec-

tion that this notion of blameworthiness requires a stronger form
of freedom, a form which may be incompatible with the Causal
Thesis. In order to assess this objection, it will be helpful to com-
pare the contractualist account of blame with what Smart calls
praise and dispraise. According to Smart, we commonly use the
word praise in two different ways.16 On the one hand, praise
is the opposite of blame. These terms apply only to what a person
does or to aspects of a persons character, and they are supposed
to carry a special force of moral approval or condemnation. But
we also praise things other than persons and their character: the
California climate, the flavor of a melon, or the view from a cer-
tain hill. In this sense we also praise features of persons which
we see as gifts beyond their control: their looks, their coordina-
tion, or their mathematical ability. Praise in this sense is not the
opposite of blame, and Smart coins the term dispraise to denote
its negative correlate. Praise and dispraise lack the special force
of moral approval or condemnation which praise and blame are
supposed to have. To praise or dispraise something is simply to
grade it.
Smart takes the view that the kind of moral judgment involved
in praise and blame as these terms are normally used must be
rejected because it presupposes an unacceptable metaphysics of
free will. However, we can praise and dispraise actions and char-
acter just as we can grade eyes and skill and mountain peaks. The
primary function of praise in this grading sense, according to
Smart, is just to tell people what people are like.17 However,
since people like being praised and dislike being dispraised, praise
and dispraise also have the important secondary function of serv-
ing to encourage or discourage classes of actions. Smart suggests
that clear-headed people, insofar as they use the terminology of
praise and blame, will use it only in this grading sense and will
16Smart, Freewill, Praise, and Blame, p. 210.
17Ibid., p. 211.
[SCANLON] The Siginifance of Choice 169
restrict its use to cases in which this important secondary function

can be fulfilled.
Most people would agree that moral praise and blame of the
kind involved when we hold a person responsible have a force
which goes beyond the merely informational function of telling
people what people are like. The problem for a compatibilist
is to show that judgments with this additional force can be
appropriate even if the Causal Thesis is true. The prior problem
for moral theory is to say what this additional force is. What is
it that an account of moral judgment must capture in order to be
successfully compatibilist?
As I have said, Smarts analysis is not compatibilist. His aim
is to replace ordinary moral judgment, not to analyze it. Strawson,
on the other hand, is offering a compatibilist analysis of (at least
some kinds of) moral judgment, and his analysis clearly satisfies
one-half of the compatibilist test. The expression of interpersonal
reactive attitudes is compatible with the Causal Thesis for much
the same reason that Smarts notions of praise and dispraise are.
These attitudes are reactions to what people are like, as this is
shown in their actions. As long as the people in question really
are like this as long, that is, as their actions really do manifest
the attitudes in question these reactive attitudes are appropriate.
Strawsons theory is more appealing than Smarts because it
offers a plausible account of moral judgment as we currently
understand it, an account of how moral judgment goes beyond
merely saying what people are like and of how it differs from
mere attempts to influence behavior. But his theory is like Smarts
in locating the special force of moral judgment in what the
moral judge is doing. The contractualist account I am offering,
on the other hand, locates the origin of this distinctive force in
what is claimed about the person judged. It is quite compatible
with this analysis that moral judgments should often be intended
to influence behavior and that they should often be made as
expressions of reactive attitudes ; but such reforming or expressive
intent is not essential. What is essential, on this account, is that a

judgment of moral blame asserts that the way in which an agent
decided what to do was not in accord with standards which that
agent either accepts or should accept insofar as he or she is con-
cerned to justify his or her actions to others on grounds that they
could not reasonably reject. This is description, but given that
most people care about the justifiability of their actions to others,
it is not mere description.
This account of the special force of moral judgment may still
seem inadequate. Given what I have said it may seem that, on the
contractualist view, this special force lies simply in the fact that
moral judgments attribute to an agent properties which most
people are seriously concerned to have or to avoid. In this respect
moral judgments are like judgments of beauty or intelligence. But
these forms of appraisal, and the pride and shame that can go with
accepting them, involve no attribution of responsibility and hence
raise no question of freedom. To the extent that moral appraisal
is different in this respect, and does raise a special question of free-
dom, it would seem that this difference is yet to be accounted for.
One way in which freedom is relevant to moral appraisal on
the Quality of Will theory (the main way mentioned so far) is
this: insofar as we are talking about praising or blaming a person
on the basis of a particular action, the freedom or unfreedom of
that action is relevant to the question whether the intentions and
attitudes seemingly implicit in it are actually present in the agent.
This evidential relevance of freedom is not peculiar to moral
appraisal, however. Similar questions can arise in regard to assess-
ments of intelligence or skill on the basis of particular pieces of
behavior. (We may ask, for example, whether the occasion was
a fair test of her skill, or whether there were interfering condi-
tions.) The objection just raised does not dispute the ability of
the Quality of Will theory to explain this way in which moral
judgments may depend on questions of freedom, but it suggests
that this is not enough. It assumes that blameworthy intentions
and attitudes are correctly attributed to an agent and then asks

how, on the analysis I have offered, this attribution goes beyond
welcome or unwelcome description. Behind the objection lies the
idea that going beyond description in the relevant sense would
involve holding the agent responsible in a way that people are
not (normally) responsible for being beautiful or intelligent and
that this notion of responsibility brings with it a further condition
of freedom which my discussion of the Quality of Will theory has
so far ignored.
I do not believe that in order to criticize a person for behaving
in a vicious and callous manner we must maintain that he or she
is responsible for becoming vicious and callous. Whether a person
is so responsible is, in my view, a separate question. Leaving this
question aside, however, there is a sense in which we are respon-
sible for or, I would prefer to say, accountable for our inten-
tions and decisions but not for our looks or intelligence. This is
just because, insofar as these intentions and decisions are ours,
it is appropriate to ask us to justify or explain them appropriate,
that is, for someone to ask, Why do you think you can treat me
this way? in a way that it would not be appropriate to ask, in an
accusing tone, Why are you so tall? This is not to say that these
mental states are the kinds of thing which have reasons rather than
causes but only that they are states for which requests for reasons
are in principle relevant.
Moral criticism and moral argument, on the contractualist
view, consist in the exchange of such requests and justifications.
Adverse moral judgment therefore differs from mere unwelcome
description because it calls for particular kinds of response, such
as justification, explanation, or admission of fault. In what way
does it call for these responses? Here let me make three points.
First, the person making an adverse moral judgment is often
literally asking for or demanding an explanation, justification, or
apology. Second, moral criticism concerns features of the agent
for which questions about reasons, raised by the agent him or her-
self, are appropriate. Insofar as I think of a past intention, deci-

sion, or action as mine, I think of it as something which was
sensitive to my assessment, at the time, of relevant reasons. This
makes it appropriate for me to ask myself, Why did I think or do
that? and Do I still take those reasons to be sufficient? Third, the
contractualist account of moral motivation ties these two points
together. A person who is concerned to be able to justify him- or
herself to others will be moved to respond to the kind of demand
I have mentioned, will want to be able to respond positively (i.e.,
with a justification) and will want to carry out the kind of first-
person reflection just described in a way that makes such a response
possible. For such a person, moral blame differs from mere un-
welcome description not only because of its seriousness but also
because it engages in this way with an agents own process of criti-
cal reflection, thus raising the questions Why did I do that? Do I
still endorse those reasons? Can I defend the judgment that they
were adequate grounds for acting?
Whether one accepts this as an adequate account of the spe-
cial force of moral judgments will depend, of course, on what
one thinks that moral judgment in the ordinary sense actually
entails. Some have held that from the fact that a person is morally
blameworthy it follows that it would be a good thing if he or she
were to suffer some harm (or, at least, that this would be less bad
than if some innocent person were to suffer the same harm).I8 I
do not myself regard moral blame as having this implication. So
if a compatibilist account of moral judgment must have this con-
sequence, I am content to be offering a revisionist theory. (The
problem of how the fact of choice may make harmful con-
sequences more justifiable will, however, come up again in
lecture 2.)
18This idea was suggested to me by Derek Parfit in the seminar following the
presentation of this lecture in Oxford.

7. BLAMEWORTHINESS AND FREEDOM

It remains to say something about how this contractualist ver-
sion of the Quality of Will theory handles the difficult question
of moral appraisal of the insane. Discussion of this matter will
also enable me to draw together some of the points that have just
been made and to say more about the kind of freedom which is
presupposed by moral blameworthiness according to the theory
I have been proposing.
As I said earlier, to believe that ones behavior is morally
faulty is to believe either that one has failed to attend to con-
siderations which any standards that others could not reasonably
reject would require one to attend to or that one has knowingly
acted contrary to what such standards would require. Let me focus
for a moment on the first disjunct. Something like this is a neces-
sary part of an account of moral blameworthiness, since failure
to give any thought at all to what is morally required can certainly
be grounds for moral criticism. But the purely negative statement
I have given above is too broad. The class of people who simply
fail to attend to the relevant considerations includes many who do
not seem to be candidates for moral blame: people acting in their
sleep, victims of hypnosis, young children, people suffering from
mental illness, and so on. W e need to find, within the notion of
moral blame itself, some basis for a nonarbitrary qualification of
the purely negative criterion.
According to contractualism, thought about right and wrong is
a search for principles for the regulation of behavior which
others, similarly motivated, have reason to accept. What kind of
regulation is intended here? Not regulation from without
through a system of social sanctions but regulation from within
through critical reflection on ones own conduct under the pressure
provided by the desire to be able to justify one actions to others on
grounds they could not reasonably reject. This idea of regulation
has two components, one specifically moral, the other not. The
specifically moral component is the ability to reason about what

could be justified to others. The nonmoral component is the more
general capacity through which the results of such reasoning make
a difference to what one does. Let me call this the capacity for
critically reflective, rational self-governance critically reflec-
tive because it involves the ability to reflect and pass judgment
upon ones actions and the thought processes leading up to them;
rational in the broad sense of involving sensitivity to reasons
and the ability to weigh them; self-governance because it is a
process which makes a difference to how one acts.
The critical reflection of a person who has this capacity will
have a kind of coherence over time. Conclusions reached at one
time will be seen as relevant to critical reflection at later times
unless specifically overruled. In addition, the results of this reflec-
tion will normally make a difference both in how the person acts
given a certain perception of a situation and in the features of
situations which he or she is on the alert for and tends to notice.
This general capacity for critically reflective, rational self-
governance is not specifically moral, and someone could have it
who was entirely unconcerned with morality. Morality does not
tell one to have this capacity, and failing to have it in general or
on a particular occasion is not a moral fault. Rather, morality is
addressed to people who are assumed to have this general capacity,
and it tells them how the capacity should be exercised. The most
general moral demand is that we exercise our capacity for self-
governance in ways that others could reasonably be expected to
authorize. More specific moral requirements follow from this.
Since moral blameworthiness concerns the exercise of the
general capacity of self-governance, our views about the limits of
moral blame are sensitive to changes in our views about the limits
of this capacity. We normally believe, for example, that very
young children lack this capacity and that it does not govern our
actions while we are asleep. Nor, according to some assumptions
about hypnosis, does it regulate posthypnotic suggestion, and it is
generally believed to be blocked by some forms of mental illness.

These assumptions could be wrong, but given that we hold them
it is natural that we do not take people in these categories to be
morally blameworthy for their actions. (Whether we think it is
useful to blame them is of course another question.) It is im-
portant to our reactions in such cases, however, that what is
impaired or suspended is a general capacity for critically reflec-
tive, rational self-governance. If what is lost is more specifically
moral if, for example, a person lacks any concern for the wel-
fare of others then the result begins to look more like a species
of moral fault.
As a higher order capacity, the capacity for critically reflec-
tive, rational self-governance has an obvious similarity to the
capacities for higher-order desires and judgments which figure in
the analyses of personhood and freedom offered by Harry Frank-
furt and others.19 I have been led to this capacity, however, not
through an analysis of general notions of freedom and personhood
but rather through reflection on the nature of moral argument and
moral judgment. Basic to morality as I understand it is an idea of
agreement between individuals qua critics and regulators of their
own actions and deliberative processes. Critically reflective, ra-
tional self-governance is a capacity which is required in order for
that idea not to be an idle one. It follows that moral criticism
is restricted to individuals who have this capacity and to actions
which fall within its scope.20
19See Harry Frankfurt, Freedom of the Will and the Concept of a Person,
Journal o f Philosophy 6 8 (1971): 5-20; Wright Neely, Freedom and Desire,

Philosophical Review 83 (1974) : 32-54; and Gary Watson, Free Agency, Journal
of Philosophy 72 (1975) : 205-20.
20The idea that moral criticism is applicable only to actions which are within
the scope of a capacity of self-governance which normally makes a difference in

what a person does marks a point of tangency between the Influenceability theory
and the analysis I am offering. I am not suggesting, however, that particular acts of
moral criticism are aimed at influencing people or that moral criticism is always
inappropriate when there is no hope of its making any difference to what people
do. Morality as I am describing it is in a general sense action guiding moral
argument concerns principles for the general regulation of behavior. But moral
In Frankfurts terms, these restrictions correspond roughly to a

restriction to persons (as opposed to wantons) and a restriction
to actions which are performed freely. In my view, however, this
last characterization is not entirely apt. Aside from external
impediments to bodily motion, what is required for moral ap-
praisal on the view I am presenting is the freedom, whatever
it may be, which is required by critically reflective, rational self-
governance. But this is less appropriately thought of as a kind of
freedom than as a kind of intrapersonal responsiveness. What is
required is that what we do be importantly dependent on our
process of critical reflection, that that process itself be sensitive to
reasons, and that later stages of the process be importantly de-
pendent on conclusions reached at earlier stages. But there is no
reason, as far as I can see, to require that this process itself not
be a causal product of antecedent events and conditions. Calling
the relevant condition a form of freedom suggests this require-
ment, but this suggestion is undermined by our investigation into
the moral significance of choice.
8. CONCLUSION
The contractualist version of the Quality of Will theory which
I have described seems to me to provide a satisfactory explanation
of the significance of choice for the moral appraisal of agents.

ought judgments need not be intended as action guiding, and insofar as they do
guide action they need not do so by being prescriptive in form. Rather, they guide
action by calling attention to facts about the justifiability of actions facts which
morally concerned agents care about. In these respects my view differs from R. M.
Hares prescriptivism, though we would say some of the same things about free
will. See his Prediction and Moral Appraisal, in P. French, T. Uehling, and
H. Wettstein, eds., Midwest Studies in Philosophy, vol. III (Minneapolis: Uni-
versity of Minnesota Press, 1978), pp. 17-27.
21For more extended discussion of this issue, see Daniel Dennetts Elbow Room
(Cambridge, Mass.: MIT Press, 1984), especially chs. 3-5. I make no claim to be
advancing beyond what other compatibilists have said about the nature of delibera-
tion and action. My concern is with the question of moral responsibility. Here I
differ with Dennett, who goes much further than I would toward accepting the
Influenceability theory. See ch. 7 of Elbow Room and Gary Watsons criticisms of
it in his review in Journal of Philosophy 8 3 (1986): 517-22.
This theory offers a convincing and unified account of familiar

excusing conditions, such as mistake of fact and duress, and
explains our reactions to questions about moral appraisal of very
young children, the insane, and victims of hypnosis. It can explain
the special critical force which moral judgments seem to have, and
it does this without presupposing a form of freedom incompatible
with the Causal Thesis. But the theory applies only to what I
called earlier the moral version of the free-will problem. A
parallel account may, as I will suggest later, have some relevance
to the case of criminal punishment, but it does not offer a promis-
ing approach to the other problems I have mentioned. The sig-
nificance of a persons choices and other subjective responses for
questions of economic justice and freedom of thought may have
something to do with the fact that these responses reflect what
might loosely be called the quality of the persons will, but this
is not because what we are doing in these cases is judging this
quality or expressing attitudes toward it (since this is not what
we are doing.) So, in search of an explanation that might cover
these other cases, I will look in a different direction.
Lecture 2
1. THE VALUE OF CHOICE

It would have been natural to call these lectures an investiga-
tion into the significance of voluntariness. I have spoken of
choice instead because this term applies not only to something
that an agent does as in She made a choice but also to
what an agent is presented with as in She was faced with this
choice. It thus encompasses both an action and a situation within
which such an action determines what will happen: a set of
alternatives, their relative desirabilities, the information available
to the agent, and so on. My main concern in these lectures is with
the significance of choice in the first of these senses: the moral
significance of the choices people make. In this lecture, however,

I will present a theory which exploits the ambiguity just men-
tioned by seeking to explain one kind of moral significance of the
choices people make in terms of the value of the choices they have.
I will call this the Value of Choice theory.22
This theory starts from the idea that it is often a good thing
for a person to have what will happen depend upon how he or
she responds when presented with the alternatives under the right
conditions. To take a banal example, when I go to a restaurant, it
is generally a good thing from my point of view to have what
appears on my plate depend on the way in which I respond when
presented with the menu. The most obvious reason why choice
has value for me in this situation is simply instrumental: I would
like what appears on my plate to conform to my preferences at the
time it appears, and I believe that if what appears then is made to
depend on my response when faced with the menu then the result
is likely to coincide with what I want. This reason for valuing
choice is both conditional and relative. It is conditional in that the
value of my response as a predictor of future satisfaction depends
on the nature of the question and the conditions under which my
response is elicited. It is relative in that it depends on the reli-
ability of the available alternative means for selecting the out-
comes in question. In the restaurant case this value depends on
how much I know about the cuisine in question and on my condi-
tion at the time the menu arrives: on whether I am drunk or over-
eager to impress my companions with my knowledge of French
22As I have said, the basic idea of this theory was presented by Hart in Legal
Responsibility and Excuses. Since Harts article others have written in a similar
vein, although they have been concerned mainly with the theory of punishment. See,
for example, John Mackie, The Grounds of Responsibility, in P. M. S. Hacker
and J. Raz, eds., Law, Morality, and Society: Essays in Honour of H. L. A. Hart,
(Oxford: Oxford University Press, 1977), and C. S. Nino, A Consensual Theory
of Punishment, Philosophy and Public Affairs 1 2 (1983) : 289-306. Like Hart,
Nino links the significance of choice (in his terms, consent) as a condition of just
punishment with its significance elsewhere in the law, e.g., in contracts and torts.
His view of this significance, however, is closer than my own to what I refer to
below as the Forfeiture View.
or my ability to swallow highly seasoned food. Thus the same

interest which sometimes makes choice valuable the desire that
outcomes should coincide with ones preferences can at other
times provide reasons for wanting outcomes to be determined in
some other way, When I go to an exotic restaurant with my
sophisticated friends, the chances of getting a meal that accords
with my preferences may be increased if someone else does the
ordering.
What I have described so far is what might be called the
predictive or instrumental value of choice. In the example
I have given, choice is instrumental to my own future enjoyment,
but the class of states which one might seek to advance by making
outcomes dependent on choices is of course much broader. Aside
from such instrumental values, however, there are other ways in
which having outcomes depend on my choice can have positive
or negative value for me. One of these, which I will call demon-
strative value, can be illustrated as follows. On our anniversary,
I want not only to have a present for my wife but also to have
chosen that present myself. This is not because I think this process
is the one best calculated to produce a present she will like (for
that, it would be better to let her choose the present herself). The
reason, rather, is that the gift will have special meaning if I choose
it if it reflects my feelings about her and my thoughts about the
occasion. On other occasions, for reasons similar in character but
opposite in sign, I might prefer that outcomes not be dependent
on my choices. For example, I might prefer to have the question
of who will get a certain job (my friend or a stranger) not depend
on how I respond when presented with the choice: I want it to be
clear that the outcome need not reflect my judgment of their re-
spective merits or my balancing of the competing claims of merit
and loyalty.
The features of oneself which one may desire to demonstrate
or see realized in action are highly varied. They may include the
value one attaches to various aims and outcomes, ones knowledge,
awareness, or memory, or ones imagination and skill. Many of

these are involved in the example cited: I want to make the choice
myself because the result will then indicate the importance I attach
to the occasion (my willingness to devote time to choosing a gift);
my memory of, attention to, and concern for what she likes; as
well as my imagination and skill in coming up with an unusual
and amusing gift. The desire to see such features of oneself mani-
fested in actions and outcomes is of course not limited to cases in
which ones feelings for another person are at issue. I want to
choose the furniture for my own apartment, pick out the pictures
for the walls, and even write my own lectures despite the fact that
these things might be done better by a decorator, art expert, or
talented graduate student. For better or worse, I want these things
to be produced by and reflect my own taste, imagination, and
powers of discrimination and analysis. I feel the same way, even
more strongly, about important decisions affecting my life in
larger terms: what career to follow, where to work, how to live.
These last examples, however, may involve not only demon-
strative but also what I will call symbolic value. In a situation
in which people are normally expected to determine outcomes of
a certain sort through their own choices unless they are not compe-
tent to do so, I may value having a choice because my not having
it would reflect a judgment on my own or someone elses part that
I fell below the expected standard of competence. Thus, while I
might like to have the advantage of my sophisticated friends
expertise when the menu arrives tonight, I might prefer, all things
considered, to order for myself, in order to avoid public acknowl-
edgment of my relative ignorance of food, wine, and foreign
cultures.
I make no claim that these three categories of value are
mutually exclusive or that, taken together, they exhaust the forms
of value that choice can have. My aim in distinguishing them is
simply to illustrate the value that choice can have and to make
clear that this value is not always merely instrumental: the reasons
[SCANLON] The Significanace of Choice 181
people have for wanting outcomes to be (or sometimes not to be)

dependent on their choices has to do with the significance that
choice itself has for them, not merely with its efficacy in promoting
outcomes which are desired on other grounds.
The three forms of value which I have distinguished (predic-
tive, demonstrative, and symbolic) would all figure in a full
account of the problem of paternalism. Legal restriction of
peoples freedom for their own good is likely to seem justified
where (a) people who make a certain choice are likely to suffer
very serious loss; (b) the instrumental value of choice as a way
of warding off this loss is, given the circumstances under which
that choice would be exercised, seriously undermined; (c) the
demonstrative value that would be lost by being deprived of this
choice is minimal; and (d) the tendency to make the wrong
choice under the circumstances in question is widely shared, so
that no particular group is being held inferior in the argument for
legal regulation. The pejorative ring of paternalism and the
particular bitterness attaching to it stem from cases in which
either the seriousness of the loss in question or the foolishness of
the choice leading to it is a matter of controversy. Those who are
inclined to make a particular choice may not see it as mistaken
and may attach demonstrative value to it. Consequently, they may
resent paternalistic legislation, which brands them as less than
fully competent when, in their view, they merely differ from the
majority in the things they value. But this kind of resentment need
not properly extend to other kinds of legislation sometimes called
paternalistic, such as wage and hour laws. Whether there is any
reason at all for such resentment will depend on the reasons sup-
porting a piece of legislation and also on the reasons people
actually have for valuing freedom of choice which they would
lose.
As controversies about paternalism illustrate, people can dis-
agree sharply about the value of particular choices. They disagree,
for example, about how important it is to have whether one wears
a seat belt depend on how one reacts (in the absence of any
coercion) when setting off in a car. Some regard it as a significant
loss when some form of coercion or even mild duress (the threat
of a fine, or even the monitory presence of a brief buzzer) is
introduced. Others, like me, regard this loss as trivial, and see the
constrained choice as significantly more valuable than the un-
constrained one. This disagreement reflects differences in the in-
strumental, demonstrative, and symbolic value we attach to these
choices.
The existence of such differences raises the question of what
is to count as the value of a choice as I have been using this
phrase. One possibility is what I will call fully individualized
value. This is the value of the choice to a particular individual,
taking into account the importance that individual attaches to hav-
ing particular alternatives available, the difference that it makes
to that individual which of these alternatives actually occurs, the
importance which the individual attaches to having this be deter-
mined by his or her reactions, and the skill and discernment with
which that individual will choose under the conditions in question.
This fully individualized value may not be the same as the value
which the individual actually assigns to the choice in question;
rather, it is the ex ante value which he or she should assign given
his or her values and propensities.
Fully individualized value is not what normally figures in
moral argument, however. Appeals to the value of choice arise in
moral argument chiefly when we are appraising moral principles
or social institutions rather than when we are discussing particular
choices by specific individuals. In these contexts we have to
answer such general questions as How important is it to have the
selection among these alternatives depend on ones choice? How
bad a thing is it to have to choose under these conditions? When
we address these questions, fully individualized values are not
known. W e argue instead in terms of what might be called the
normalized value of a choice: a rough assignment of values to
categories of choice which we take to be a fair starting point for

justification. Thus, for example, we take it as given for purposes
of moral argument that it is very important that what one wears
and whom one lives with be dependent on ones choices and much
less important that one be able to choose what other people wear,
what they eat, and how they live. And we do this despite the fact
that there may be some who would not agree with this assignment
of values.
This phenomenon the use in moral argument of nonunani-
mously held normalized standards of value is familiar and
by no means limited to the case of choice. The status and justifica-
tion of such standards is a difficult problem in moral theory. I
will not address the general question here but will mention briefly
two points about the case of choice. First, giving people the
choice for example, the opportunity to transfer goods through
market trading is one way to deal with the problem of divergent
individual preferences. What has just been indicated, however, is
that it is at best a partial solution. Having a choice among
specified alternatives under specified conditions is itself a good
which individuals may value differently as is having the
choice whether to have the choice and so on.23 Second, differences
in individualized valuations of choices result not only from dif-
ferences in preference but also from differences in the personal
characteristics which make a choice valuable: differences in fore-
sight, in self-control, in self-understanding, and so on. Moral
argument commonly refers to normal levels of these capacities
as well as to normal valuations of outcomes and of demonstra-
tive and symbolic values.
Let me turn now to the question of how the value of choice
is related to the Quality of Will theory, discussed above. Like
23The variability of the value of choice is pointed out clearly by Gerald
Dworkin in Is More Choice Better Than Less? in P. French, T. Uehling, and U

H. Wettstein, eds., Midwest Studies in Philosophy, vol. VII (Minneapolis: Univer-
sity of Minnesota Press, 1982), pp. 47-62.
what I have here called predictive and demonstrative value, the

form of appraisal underlying the Quality of Will theory starts
from the obvious fact that subjective responses can indicate or
express continuing features of a person and from the equally
obvious fact that these responses are better indicators under some
conditions than under others. Even in this common starting point,
however, there is a difference: the features of the person with
which the Quality of Will theory is concerned constitute a narrow
subset of those that give choice its value for the agent. For ex-
ample, I want to choose my own food largely because my choices
will be good indicators of what will please me, but my being pleased
more by fish than by liver is not part of the quality of my will
with which moral judgment is concerned.
Where the two theories differ most importantly, however, is
in the way in which they assign moral significance to this indica-
tive aspect of choice. The Quality of Will theory takes the point
of view of the moral judge. Variations in the indicative value of
subjective responses are significant from this point of view because
moral judgment involves an inference from behavior to quality of
will. The Value of Choice theory, on the other hand, begins with
the value for an agent of having outcomes depend (or not de-
pend) on his or her subjective responses under certain conditions.
This (so far purely personal) value takes on moral significance
by being the basis for a claim against social institutions (or against
other individuals). In my view, to show that a social institution
is legitimate one must show that it can be justified to each person
affected by it on grounds which that person could not reasonably
reject. One thing which people may reasonably demand, however,
is the ability to shape their lives and obligations through the exer-
cise of choice under reasonably favorable conditions. Moral prin-
ciples or social institutions which deny such opportunities when
they could easily be provided, or which force one to accept the
consequences of choice under extremely unfavorable conditions
which could be improved without great cost to others, are likely
to be reasonably rejectable for that reason. Let me illustrate by

considering some examples.
2. JUSTICE AND CHOICE
Consider first the economic justice example which I mentioned

earlier, Suppose a society, not marked by significant economic
inequalities, decides that it needs to have a significant proportion
of its workforce work overtime at a particular job. To this end,
a bonus is offered to anyone willing to undertake the work, at an
amount calculated to elicit the required number of volunteers. The
choice between extra pay and extra leisure has obvious instru-
mental value for the people involved, and giving people this
choice makes it overwhelmingly likely that those who prefer addi-
tional income (with additional labor) will get it, while those who
prefer the opposite will get what they prefer. If overtime work
was not made dependent on choice the scheme would be very dif-
ficult to justify; with this feature, justification is much easier.
Nonetheless, whether or not a given worker winds up among
those with extra pay will no doubt depend on some morally arbi-
trary facts about his or her background. Why then is this situa-
tion any better than the one criticized by Rawls?
The difference does not lie in the fact that the choices made
in one case have causal antecedents while those made in the other
case do not. In the egalitarian case, however, we can say that by
placing the people in those circumstances, offering them that
choice, and letting the outcome be determined by the choice they
make under those conditions, we have done as much for them as
could reasonably be required. In the other case it may be argued
that we cannot say this: once the people are placed in disadvan-
tageous circumstances, circumstances which themselves make it
very unlikely that anyone would make the choices necessary to
escape, offering these people the opportunity to exert themselves
does little to improve their position.
The background conditions under which choices are made in

the laissez-faire system are arbitrary from a moral point of view
in this sense: they could be almost anything. All we know is that
they will be conditions which arose from a series of voluntary
transactions, and this does nothing to ensure that they will be
good conditions under which to choose. Consequently, there is no
assurance that these conditions will have the moral property of
being conditions under which choices confer legitimacy on their
outcomes.
This interpretation of Rawlss objection to the laissez-faire
system of natural liberty provides the basis for a reply to one
line of criticism raised by Nozick and others. Nozick interprets
Rawls as arguing that the fact that some people exert themselves,
take risks, and excel while others do not do so cannot by itself
justify different economic rewards for the two groups because
these differences in motivation may be the result of causal factors
outside the control of the agents themselves. He goes on to object
that
this line of argument can succeed in blocking the introduction
of a persons autonomous choices and actions (and their re-
sults) only by attributing everything noteworthy about the per-
son completely to certain sorts of external factors. So
denigrating a persons autonomy and prime responsibility for
his actions is a risky line to take for a theory that founds so
much (including a theory of the good) upon persons choices.24
The problem which Nozick raises here is a version of the

political problem of free will as I presented it in my first lec-
ture. My reply (I do not claim that this was also Rawlss inten-
tion) is that it is not mere attributability to external factors
that undermines the legitimating force of the choices in a system
of natural liberty. The problem, rather, is that such a system
24Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974),
p. 214.
provides no assurance that these factors will not be ones which

undermine the value of choice for many people in the society.
Suppose that I exert myself to develop my talents and become
wealthy. You, on the other hand, suffering the psychological
effects of your unfortunate starting position, fail to exert yourself,
and as a consequence remain poor. Can I claim credit for my
initiative and perseverance, given that they resulted from fortu-
nate family and social circumstances for which [I] can claim no
credit?25 If to claim credit means simply to consider these
traits and actions mine in the sense required in order to take
pride in them, then the answer is clearly yes. My accomplishments
reflect personal qualities which I really do have. If, however, what
is meant is that these differences in our behavior can be taken to
justify my having more income and your having less, then the
answer may be no. This is not because my actions, being caused
by outside factors, are not mine, or because your actions, simi-
larly caused by other factors, are therefore not yours, but rather
because presenting a person with a choice of the kind you had is
not doing enough for that person.
Of course, Rawls and Nozick disagree over what constitutes
doing enough for a person. For Nozick, one has done enough
as long as the persons Lockean rights have not been violated;
for Rawls, the standard is set by the principles which would be
accepted behind the Veil of Ignorance. As a result, Rawlss re-
marks about factors arbitrary from a moral point of view, as I
have interpreted them, may seem not to advance his argument
against Notick but merely to restate the disagreement between
them. But this restatement seems to me to have several virtues.
First, it locates the disagreement in what seems, intuitively, to
be the right place in a question of justice rather than in a sepa-
rate (and I believe spurious) question of causal determination.
Second, framing the argument in terms of the value of choice has
25Rawls, A Theory of Justice, p. 104; quoted by Nozick, Anarchy, State, and
Utopia, p. 214.
the effect of disentangling the idea of individual liberty from

Nozicks particular system of Lockean rights. This allows oppo-
nents of that system to make clear that they, too, value individual
choice and liberty and gives them a chance to put forward their
alternative interpretations of these values. The argument can then
proceed as a debate about the merits of competing interpretations
of the moral significance of liberty and choice rather than as a
clash between defenders of liberty and proponents of equality or
some other pattern of distribution.
The Value of Choice theory represents a general philosophical
strategy which is common to Harts analysis of punishment and
Rawlss theory of distributive justice as I have just interpreted it.
In approaching the problems of justifying both penal and eco-
nomic institutions we begin with strong pretheoretical intuitions
about the significance of choice : voluntary and intentional com-
mission of a criminal act is a necessary condition of just punish-
ment, and voluntary economic contribution can make an economic
reward just and its denial unjust. One way to account for these
intuitions is by appeal to a preinstitutional notion of desert: cer-
tain acts deserve punishment, certain contributions merit rewards,
and institutions are just if they distribute benefits and burdens in
accord with these forms of desert.
The strategy I am describing makes a point of avoiding any
such appeal. The only notions of desert which it recognizes are
internal to institutions and dependent upon a prior notion of
justice: if institutions are just then people deserve the rewards and
punishments which those institutions assign them. In the justifica-
tion of institutions, the notion of desert is replaced by an inde-
pendent notion of justice; in the justification of specific actions
and outcomes it is replaced by the idea of legitimate (institu-
tionally defined) expectations.26
In order for this strategy to succeed, the conception of justice
by which institutions are to be judged must adequately represent
26Rawls, A Theory of Justice, p. 313.
[ SCANLON] The Significance of Choice 189
our intuitions about the significance of choice without falling back

on a preinstitutional concept of desert. This is where the idea of
the value of choice comes in. Just institutions must make out-
comes depend on individuals choices because of the importance
which individuals reasonably attach to this dependence. But there
is a serious question whether this strategy can account for the dis-
tinctive importance which choice appears to have. Insofar as
choice-dependence is merely one form of individual good among
others, it may seem that the Value of Choice theory will be unable
to explain our intuition that the moral requirement that certain
outcomes be made dependent on peoples choices is not to be sacri-
ficed for the sake of increases in efficiency, security, or other
benefits.
Several defenses can be offered against this charge. The first
is to point out the distinctiveness of the value of choice as com-
pared with other elements in a persons welfare. As I have indi-
cated above, the value of choice is not a purely instrumental value.
People reasonably attach intrinsic significance to having outcomes
depend on their choices. In addition, the moral requirements
which this value gives rise to within a contractualist moral theory
are not corollaries of a more general duty to look out for peoples
welfare. In fact, the demand to make outcomes depend on peoples
choices and the demand to promote their welfare are quite inde-
pendent, and they can often pull in opposite directions.
A second defense parallel to Rawlss argument for the
priority of liberty is to argue that in appraising social institu-
tions people would reasonably set a particularly high value on
having certain kinds of outcomes be dependent on their choices.27
A third, more pragmatic defense is to argue that the distinctive
significance which choice appears to have is in part an artifact of
the position from which we typically view it. This is a position
internal to institutions, and one in which choices have special
salience because they are the last justifying elements to enter the
See section 82 of A Theory of Justice.
27
picture. When the relevant background is in place when condi-

tions are right, necessary safeguards have been provided, and so
on the fact that a person chooses a certain outcome may make
that outcome one that he or she cannot reasonably complain of.
But choice has this effect only when these other factors are present.
Because they are relatively fixed features of the environment, these
background conditions are less noticeable than the actions of the
main actors in the drama, but this does not mean that they are less
impor tan t.
These defenses are most convincing in those cases in which the
first argument is strongest that is, in cases like the economic
justice example just discussed, in which peoples desire to shape
their own lives gives choice an important, positive value. The
Value of Choice theory looks weaker in cases where the only rea-
son for wanting to have a choice is that it makes certain unwanted
outcomes (such as punishment) less likely. Here choice has no
positive value rather than have the choice, one would prefer to
eliminate these outcomes altogether if that were possible yet
the fact of choice seems to retain its special significance as a
justifying condition. Let me turn, then, to an example of this kind.
3. CHOICE AND PROTECTION
Suppose that we, the officials of a town, must remove and

dispose of some hazardous waste. W e need to dig it up from the
illegal dump near a residential area where it has lain for years and
move it to a safer spot some distance away. Digging it up and
moving it will inevitably release dangerous chemicals into the
atmosphere, but this is better than leaving it in its present loca-
tion, where it will in the long run seep into the water supply.
Obviously we must take precautions to minimize the risks involved
in this operation. W e need to find a safe disposal site, far away
from where people normally have to go. We should build a high
fence around the new site, and another around the old one where
the excavation is to be done, both of them with large signs warn-

ing of the danger. W e should also arrange for the removal and
transportation to be carried out at times when few people are
around, in order to minimize the number potentially exposed, and
we must be sure to have the material wetted down and transported
in covered trucks to minimize the amount of chemicals released
into the air. Inevitably, however, enough chemicals will escape to
cause lung damage to those who are directly exposed if, because of
past exposure or genetic predisposition, they happen to be par-
ticularly sensitive, but not enough to pose a threat to anyone who
stays indoors and away from the excavation site. Given that this is
so, we should be careful to warn people, especially those who
know that they are at risk, to stay indoors and away from the
relevant area while the chemicals are being moved.
Suppose that we do all of these things but that nonetheless
some people are exposed. A few of these, who did not know that
they were particularly sensitive to the chemical, suffer lung dam-
age. Let me stipulate that with respect to all of these people we
did all that we could reasonably be expected to do to warn and
protect them. So in that sense they cant complain about what
happened. The question which concerns me, however, is what role
the signs and warnings play in making this the case. These are the
factors which make outcomes depend on peoples choices. Are
they, like the fences, the careful removal techniques, and the
remote location of the new site, just further means through which
the likelihood of someones being injured is reduced? This is what
the Value of Choice theory seems to imply. For after all, since no
one wants to have the opportunity to be exposed to this chemical,
the only value which choice can have in this case is that of making
exposure less likely. This may be an adequate explanation of why
we would want to be warned and hence given the choice
whether to be exposed or not. But it may not account for the full
moral significance of the fact that those who were injured knew
what they were getting into. Consider the following two cases.
Suppose that one person was exposed because, despite the

newspaper stories, mailings, posted signs, radio and television
announcements, and sound trucks, he never heard about the
danger. He simply failed to get the word. So he went for his
usual walk with no idea what was going on. A second person, let
us suppose, heard the warnings but did not take them seriously.
Curious to see how the task was being done, she sneaked past the
guards and climbed the fence to get a better look.
There seems to be a clear difference between these two cases.
In the first, we have done enough to protect the person simply
because, given what we have done, it was extremely unlikely that
anyone would be directly exposed to contamination, and we could
not have made this even more unlikely without inordinate ex-
pense. There is, after all, a limit to the lengths to which we must
go to protect others. The second person, on the other hand, bears
the responsibility for her own injury, and it is this fact, rather than
any consideration of the cost to us of doing more, which makes
it the case that she has no claim against us. By choosing, in the
face of all our warnings, to go to the excavation site, she laid
down her right to complain of the harm she suffered as a result.
4.T HE FORFEITURE VIEW

This familiar and intuitively powerful idea about the sig-
nificance of choice, which I will call the Forfeiture View, is not
captured by either of the theories I have been considering. It is
distinct from the Value of Choice theory, since on that theory what
matters is the value of the choice a person is presented with: once
a person has been placed in a sufficiently good position, the out-
come which emerges is legitimate however it may have been pro-
duced. On the Forfeiture View, on the other hand, it matters
crucially that an outcome actually resulted from an agents con-
scious choice, the agent having intentionally passed up specific
alternatives. This is why that view accounts so well for our reac-
tion to the person in the second example: not only does she have
no one else to blame for her fate; she has herself to blame. We
could account for this sense of blame by appealing to a prudential
version of the Quality of Will theory: the process of deliberation
leading to a decision to climb over the fence just to see what they
are doing is obviously faulty. But the Quality of Will theory
is an account of the moral appraisal of agents, while what we are
concerned with here is the justification of outcomes. It may be
natural to suppose that a difference in the first translates into or
supports a difference in the second, but on reflection it is by no
means obvious how this is so.
Moreover, the idea of fault is in fact irrelevant here. The
intuition to which the Forfeiture View calls attention concerns
the significance of the fact of choice, not the faultiness of that
choice. We can imagine a person who, unlike the imprudently
curious woman in my example, did not run the risk of contamina-
tion foolishly or thoughtlessly. Suppose this third person found,
just as the excavation was about to begin, that the day was a perfect
one for working on an outdoor project to which she attached great
value. Aware of the danger, she considered the matter carefully
and decided that taking into account her age and condition it was
worth less to her to avoid the risk than to advance her project in
the time she was likely to have remaining. Surely this person is as
fully responsible for her fate as the imprudent woman whom I
originally described. But her decision is not a foolish or mis-
taken one.
This illustrates the fact that what lies behind the Forfeiture
View is not an idea of desert. That is, it is not an idea according
to which certain choices, because they are foolish, immoral, or
otherwise mistaken, positively merit certain outcomes or responses.
The idea is rather that a person to whom a certain outcome was
available, but who knowingly passed it up, cannot complain about
not having it: volenti non fit iniuria.
It is important to remember here that the challenge of the
Forfeiture View lies in the suggestion that the Value of Choice
theory gives an inadequate account of the significance of choice

in the justification of institutions, policies, and specific moral prin-
ciples. Once we have accepted as justified an institution or policy
attaching specific consequences to particular choices, there is no
disagreement about whether these choices have the kind of special
force which the Forfeiture View claims. This force can be accounted
for by appeal to the institutions, principles, or policies in question.
The disagreement concerns the way in which such institutions,
principles, and policies themselves are to be justified. When the
Forfeiture View says that people who make certain choices can-
not complain about the harms they suffer as a result, what is
meant is that these harms lack the force in this process of justifica-
tion which otherwise comparable harms would have.
It may seem that a view of this kind is in fact forced on us by
contractualism. According to contractualism the crucial question
about a proposed moral principle is whether anyone could rea-
sonably reject it. In order for rejecting a principle to be reasonable
it must at least be reasonable from the point of view of the person
doing the rejecting, that is, the person who would bear the burden
of that principle. It may seem, therefore, that a harm which an
agent has the opportunity to avoid (without great sacrifice) could
never serve as a ground for reasonable rejection of a moral prin-
ciple. Consider the following argument. From the point of view
of an agent, an action which he has the choice of performing must
be seen as available to him. Suppose that an agent will run the
risk of suffering a certain harm if he follows one course of action
but that he would avoid this harm if he were to follow an alterna-
tive course which is available to him and does not involve sig-
nificant sacrifice. Given, then, that the harm is from his point of
view costlessly avoidable, how could the agent appeal to this harm
as grounds for objecting, for example, to a principle freeing others
from any duty to prevent such harms from occurring? It would
seem that such harms can have no weight in moral argument.
But this conclusion is not forced on us. In moral argument we

are choosing principles to apply in general to situations in which
we may be involved. Even if we know that actions avoiding a
certain unwanted outcome will be available to us in a given situa-
tion, we also know that our processes of choice are imperfect. We
often choose the worse, sometimes even in the knowledge that it
is the worse. Therefore, even from the point of view of an agent
looking at his own actions over time, situations of choice have to
be evaluated not only for what they make available but for
what they make it likely that one will choose. It is not unrea-
sonable to want to have some protection against the consequences
of ones own mistakes.
5. REJECTING THE FORFEITURE VIEW

The appeal of the Forfeiture View can and should be resisted.
Note, first, that the Value of Choice theory can account for the
apparent difference between the two victims of hazardous waste
removal described above. W e may have done enough to pro-
tect the first person, who failed to hear of the danger, in the sense
that we have gone to as much effort and expense as could be
expected. But because we did not succeed in making him aware of
the danger we did not make what happened depend on his choice.
Given that this kind of choice-dependence is something which
we all would want for ourselves we want such risks to be, as
far as possible, under our control we did not make this per-
son as well off as we would reasonably want to be. The second
person, on the other hand, did have the benefit of having the
choice, even though this turned out to be worth less to her than
it would be to most of us. (There was in this case a divergence
between individualized and normalized value.) Given that
she had the choice, however, and was provided with the other pro-
tections, it was true of her in a way that it was not of the first per-
son that she was placed in as good a position as one could ask for.
From the fact that a person chose, under good conditions, to

take a risk, we may conclude that he alone is responsible for what
happens to him as a result. But this conclusion need not be seen
as a reflection of the special legitimating force of voluntary action.
Rather, the fact that an outcome resulted from a persons choice
under good conditions shows that he was given the choice and
provided with good conditions for making it, and it is these facts
which make it the case that he alone is responsible. A conscious
decision to take the risk is not necessary. Consider, here, the
case of a person who was informed of the risk of contamination
but then simply forgot. As a result, he was out in his yard exercis-
ing, breathing hard, when the trucks went by. If enough was done
to protect and warn him, then this person is responsible for what
happens to him and cannot complain of it even though he made
no conscious decision to take the risk.
The central element of truth in the Forfeiture View is thus a
consequence of the Value of Choice theory rather than an alterna-
tive to it. Putting this truth in terms of the Forfeiture View, how-
ever, has the distorting effect of suggesting that choice has inde-
pendent deontic force in the justification of institutions and prin-
ciples. It also exaggerates the importance of the fact of choice
relative to that of the conditions under which the choice was made.
The Forfeiture View suggests that these conditions are important
only insofar as they bear on the voluntariness of the choice. This
is a mistake. The fact that a choice was voluntary does not always
establish that we did enough for an agent by placing him or her
in the position from which the choice was made. Nor does the
fact that an agent did not voluntarily choose an outcome, or choose
to take a certain risk, establish that what resulted was not his
fault. Giving him the opportunity to choose may have constituted
doing enough to protect him. It is thus an important virtue of
the Value of Choice theory that it gives the conditions of choice
their appropriate independent weight and forces us to keep them
clearly in view.
6. RESPONSIBILITY AND THE MORAL DIVISION

OF LABOR: BEYOND CHOICE
Within the Value of Choice theory, ideas of responsibility

arise as a derived (and often only implicit) moral division of
labor. Because most people take themselves to be more actively
concerned with the promotion of their own safety and well-being
than others are, they want outcomes to be dependent on their
choices even when this has only avoidance value. Given this
concern, giving people the choice under favorable conditions
makes it extremely unlikely that they will suffer easily avoidable
harms. W e do not want the trouble and expense of supervising
others choices more closely, and do not want them to be super-
vising us. Therefore, we take the view that giving people the
opportunity of avoiding a danger, under favorable conditions,
often constitutes doing enough for them: the rest is their re-
sponsibility. So stated, this is not a principle but only a descrip-
tion of a general tendency in our moral thought. In particular,
the idea of favorable conditions, here left vague, must be filled
in before any specific principle of responsibility is obtained, and
this filling in will be done differently in the case of different risks
and dangers.
This general analysis does, however, shed light on appeals to
responsibility in cases in which the notion of choice seems out of
place. The idea of freedom of thought, mentioned in my first lec-
ture, is one such case. Another, which I will discuss briefly here,
is the idea of responsibility for ones preferences.
This idea arises in the context of debates as to whether, for
purposes of assessing claims of justice, peoples welfare should
be measured in terms of preference satisfaction or in terms of
some objective standard of well-being such as what Rawls has
called Primary Social Goods. Objective standards of this kind
may seem unfair, since the same bundle of objective goods can
yield quite different levels of satisfaction for people with different
preferences. Rawls has replied that someone who makes this

objection must argue in addition that it is unreasonable, if not
unjust, to hold such persons responsible for their preferences and
to require them to make out as best they can. To argue this, he
says, seems to presuppose that citizens preferences are beyond
their control as propensities or cravings which simply happen.
The use of an objective standard like primary goods, on the other
hand, relies on a capacity to assume responsibility for our ends.
The conception of justice which Rawls advocates thus
includes what we may call a social division of responsibility:
society, the citizens as a collective body, accepts responsibility
for maintaining the equal basic liberties and fair equality of
opportunity, and for providing a fair share of the other pri-
mary goods for everyone within this framework, while citizens
(as individuals) and associations accept the responsibility for
revising and adjusting their ends and aspirations in view of
the all-purpose means they can expect, given their present and
foreseeable situation. This division of responsibility relies on
the capacity of persons to assume responsibility for their ends
and to moderate the claims they make on their social institu-
tions in accordance with the use of primary goods. Citizens
claims to liberties, opportunities and all-purpose means are
made secure from the unreasonable demands of others.28
I am strongly inclined to agree with Rawls here, and I have

defended a similar position myself.29 Nonetheless, I find this
argument somewhat worrisome, because it is easily misinterpreted
as involving an appeal to the idea of forfeiture which I argued
against above. On this interpretation, the argument is that the
imagined objection to objective measures of welfare overlooks the
fact that peoples preferences are under their control. Given this
28John Rawls, Social Unity and Primary Goods, in Amartya Sen and
Bernard Williams, eds., Utilitarianism and Beyond (Cambridge: Cambridge Uni-

versity Press, 1982), pp. 168, 169, 170.
29In Preference and Urgency, Journal of Philosophy 72 (1975): 655-69.
The following discussion concerns issues dealt with in my reply to the voluntari-
ness objection on pp. 664-66 of that article.
fact, and in view of the basic moral truth that one cannot com-
plain of harms one could have avoided, the objection is no objec-
tion at all: people whose preferences are particularly difficult to
satisfy have only themselves to blame.
There are two difficulties with this argument. First, for rea-
sons I have already discussed, the basic moral truth to which it
appeals seems open to serious doubt. Second, even if this truth
is correct, the argument appears to exaggerate the degree of con-
trol which people have over their preferences. To be sure, the
argument does not suggest that people can alter their preferences
by simply deciding what to prefer; the kind of control which is
envisaged is to be exercised through decisions affecting the devel-
opment of ones preferences over time. Even so, it is questionable
how much control of this kind people can realistically be assumed
to exercise.
This leads me to look for an alternative interpretation under
which the argument avoids these difficulties while still retaining
its force. Following the general strategy which I have been advo-
cating in this lecture, this alternative interpretation takes the idea
of responsibility for ones preferences to be part of the view being
defended rather than an independent moral premise. As Rawls
says, the conception of justice which he is defending includes
what we may call a social division of responsibility. The ques-
tion is how this combination an objective standard of welfare
and the idea of responsibility which it entails can be defended
without appeal to anything like the notion of forfeiture.
The issue here is the choice between two types of public
standards of justice, objective standards of the sort just described,
according to which institutions are judged on the degree to which
they provide their citizens with good objective conditions for the
development and satisfaction of their preferences, and subjective
standards, under which institutions are also judged on the basis
of the levels of preference satisfaction which actually result from
their policies. In our earlier discussion of individual choice, the
argument for a moral division of labor rested on three claims:

the value which we attach to having outcomes depend on our own
choices (even when this is only avoidance value), our reluctance
to have our choices supervised by others, and our reluctance to
bear the costs of protecting others beyond a certain point. The
case for the social division of responsibility entailed by objec-
tive standards of welfare rests on three analogous claims. We rea-
sonably attach a high value to forming our own preferences under
favorable conditions, and one reason for this is our expectation
that we will to some extent be steered away from forming prefer-
ences when we can see that they will be difficult to satisfy and will
lead mainly to frustration. Second, we do not want others to be
taking an active role in determining what we will prefer. And
third, we do not want to be burdened with the costs of satisfying
other peoples preferences when these are much more costly than
our own.
The first of these claims accounts for the (limited) force of
the idea, to which Rawls appeals, that people can to some extent
avoid costly preferences. But it does this without invoking a
preinstitutional notion of forfeiture, and without assuming the
degree of conscious and deliberate control which the Forfeiture
View would require.
The second claim is especially important. Particularly in a
society marked by sharp disagreements about what is worth pre-
ferring, a public standard of justice requiring government policy
to be aimed at raising individual levels of satisfaction is an open
invitation to unwelcome governmental intervention in the forma-
tion of individuals values and preferences. The social division
of responsibility which goes with an objective standard of wel-
fare is therefore an attractive alternative.
The case for an objective standard of welfare is thus largely
defensive. Giving up the claim to a greater share of resources in
the event that ones preferences turn out to be particularly dif-
ficult to satisfy is the price one pays for greater security against
governmental interference and greater freedom from the pos-

sibly burdensome demands of other peoples preferences. The
role of the possibility of modifying ones preferences (or of avoid-
ing the formation of preferences which are difficult to satisfy) is
just to make this price smaller and not, as the Forfeiture View
would have it, to license the result.
7. CONCLUSION
In this lecture I have presented the idea of the Value of Choice
as part of a general strategy explaining the moral significance of
choice in the justification of social institutions and policies. As
compared with its main rival, the Forfeiture View, this strategy
has the advantage of assigning choice an important positive value
without exaggerating its role and significance in justification. It
remains to be seen what kind of freedom the Value of Choice
theory presupposes and how it fits together with the Quality of
Will theory to account for the significance of choice across a range
of cases. These questions will be addressed in my next lecture.
Lecture 3
1. PUNISHMENT AND PROTECTION

Let me begin with a schematic comparison of the institution of
punishment and the policy of hazardous waste disposal which I
discussed in my last lecture. In each case we have the following
elements. First, there is an important social goal: protecting the
water supply in the one case; protecting ourselves and our pos-
sessions in the other. Second, there is a strategy for promoting
that goal which involves the creation of another risk: the risk of
contamination in the one case, the risk of punishment in the other.
Third, the effect of this strategy is to make it the case that there
is, literally or metaphorically, a certain affected area which one
can no longer enter without danger. In the one case this is the
area of excavation, transport, and disposal, in the other the area
of activities which have been declared illegal. Fourth, although
we introduce certain safeguards to reduce exposure to the risk
created, it remains the case that many of those who choose to
enter the affected area, and perhaps a few others, will suffer harm.
Some of these safeguards (such as requirements of due process,
and careful methods of excavation and transport) have the effect
of protecting those who choose to stay out of the affected area.
Other safeguards enhance the value of choice as a protection by
making it less likely that people will choose to enter. In the
hazardous waste case these include signs, warnings, and publicity
to inform people about the nature of the risk, as well as fences,
guards, and the choice of an obscure disposal site where no one
has reason to go. Analogous features in the case of punishment
are education, including moral education, the dissemination of
basic information about the law, and the maintenance of social
and economic conditions which reduce the incentive to commit
crime by offering the possibility of a satisfactory life within the
law. Restrictions on entrapment by law enforcement officers
also belong in this category of safeguards which make it less likely
that one will choose badly. Without such safeguards the value of
choice as a protection would be reduced to an unacceptable level.
In each case, in order to defend the institution in question we
need to claim that the importance of the social goal justifies creat-
ing the risk and making the affected area unusable and that, given
the prevailing conditions and the safeguards we have put in place,
we have done enough to protect people against suffering harm
from the threat that has been created.
Now let me turn to some of the differences between the two
cases. First, insofar as the activities which make up the affected
area in the case of punishment are ones which it is morally wrong
to engage in, being deprived of the ability to enter this area
without risk cannot be counted as a morally cognizable loss. This
makes the task of justification easier than in the example of

hazardous waste.
A second difference makes this task more difficult, however.
In neither case is it our aim that people should suffer the new
harm, though in both cases the possibility of their doing so is
created by our policy. But in the case of punishment this harm,
when it occurs, is intentionally inflicted on particular people. It
is an essential part of that institution that people who run afoul of
the law should be punished; but it is no part of our waste-removal
policy that those who enter the affected area should suffer con-
tamination. If, as I believe, intentionally inflicting harm is in most
cases more difficult to justify than merely failing to prevent harm,
it follows that an institution of punishment carries a heavier
burden of justification.
When such an institution is justified, however, this justification
entails the kind of forfeiture which we looked for but did not
find in the hazardous waste case. A person who intentionally
commits a crime lays down his or her right not to suffer the pre-
scribed punishment. This forfeiture is a consequence of the jus-
tification of the institution of punishment, however, not an ele-
ment in that justification. It is a consequence, specifically, of the
heavier justificatory burden just mentioned: because the institu-
tion assigns punishment to those who fulfill certain conditions,
justifying the institution involves justifying the infliction of these
penalties. If the conditions for punishment include having made
a certain kind of choice, then a justification for the institution
justifies making that choice a necessary and, when the other con-
ditions are fulfilled, sufficient condition for punishment. No such
assignment and hence no such forfeiture is involved in the jus-
tification of the policy of hazardous waste removal. A person who
recklessly chooses to enter the affected area does not lay down a
right to further protection against contamination: she has already
received all the protection she is entitled to. She does not lay
down her right to treatment (or rescue) unless this has been pre-
scribed and the policy including this prescription is justified. For-

feiture, like economic desert, is the creature of particular social
institutions and relatively specific moral principles (such as those
governing promising). It is not a moral feature of choice in gen-
eral. As I argued in my last lecture, the moral aspect of choice
which figures in the justification and criticism of such institutions
and principles is not forfeiture but the less-sharp-edged notion of
the value of choice.
I have been assuming that the affected area is so defined
that one can enter it only by conscious choice. This will be so
if we identify entering that area with committing a crime whose
definition involves conditions of voluntariness and intent. But a
system of criminal law incorporating elements of strict liability
could also fit the abstract model I have described, If a legal penalty
is attached to selling adulterated milk (not merely to doing so
knowingly, recklessly, or negligently), then one enters the
affected area simply by going into the milk business, and if such
a law is justified then doing this involves laying down ones right
not to be penalized if the milk one sells turns out to be impure.
This enlargement of the affected area is one reason (perhaps not
the only one) why such laws are more difficult to justify, especially
since the newly affected area includes activities, such as conscien-
tious engagement in the milk business, which people are morally
entitled to engage in. Having them entail forfeiture of the right
not to be punished is a morally cognizable loss.
2. EXCUSES AND THE VALUE OF CHOICE

I said in my first lecture that an acceptable account of the sig-
nificance of choice should be able to explain standardly recog-
nized excusing conditions in a way that will not generalize to
undermine the moral significance of all choice if the Causal Thesis
is true. Let me now say something about how the Value of Choice
theory fulfills this assignment. My aim here is not to derive par-
ticular excusing conditions or to define the notion of voluntariness
appropriate to particular social institutions and moral principles.

This would be an extremely time-consuming task, since it is rea-
sonable to suppose that these conditions will vary in detail from
case to case. My present purpose is merely to point out in a more
general way how the Value of Choice theory would account for
these conditions and for their variation.
The general point is obvious. If the justification for a prin-
ciple or institution depends in part on the value of the choices it
presents people with, and if the value of these choices in turn can
vary greatly depending on the presence or absence of certain con-
ditions, then in order to be justifiable the institution will have to
qualify the consequences it attaches to choices by explicitly requir-
ing the presence or absence of the most important of these
conditions.
Lack of knowledge of the nature of the alternatives available,
lack of time to consider them, and the disruptive effects of fear or
emotional distress can all weaken the connection between a per-
sons reaction at a given time and his or her more stable prefer-
ences, values, and sensitivities, thus undermining both the predic-
tive and demonstrative value of choice. Coercion and duress can
have similar disrupting effects on the process of choice, but also
and more often they diminish the value of choice simply by con-
tracting or altering the set of alternatives between which one can
choose. Diminishing the set of alternatives or weighting some
with penalties can sometimes increase the value of choice or so
those of us must believe who sign up to give lectures we have not
yet written and buy automobiles with seat belt buzzers. But this is
not usually the case.
Even when duress, false belief, or other conditions clearly
diminish the value of choice, however, it does not immediately
follow that these conditions must be recognized as negating a par-
ticular obligation or liability. Whether it does or not will depend
on, among other things, the costs to others of introducing such an
exception into the principle or institution in question. This is a
further reason why, on the present theory, it is possible for excus-

ing conditions to vary from principle to principle and institution
to institution.
Here there is a clear contrast with the genesis of excusing con-
ditions under the Quality of Will theory. Once we learn that an
agent acted under duress or under the influence of a mistaken
belief, this immediately alters the will attributable to that agent.
There is no need to ask what the effect would be of recognizing
this excuse. Of course, such considerations are relevant to the
further question of which qualities of will should be regarded
as morally deficient. But the Quality of Will theory plays no role
in answering this question; it is an account only of the process of
moral appraisal.
A second contrast between the two theories is this. The Value
of Choice theory treats changes in the set of alternatives available
to a person and changes in the conditions under which he or she
chooses among them as factors contributing to the answer to a
single question: how good or bad a thing is it to be presented with
that choice ? Under the Quality of Will theory, on the other hand,
there is an important difference here. Some conditions affect the
degree to which a will can be imputed to the agent; others
modify the nature of that will. This difference may explain Harts
remark that while continental jurisprudence has traditionally dis-
tinguished between imputability and fault he sees little to be
gained by observing this rigid distinction.30 This difference is to
be expected insofar as Hart is speaking as a Value of Choice
theorist while the continental tradition may be more concerned
with aspects of the law akin to quality of will.
3. THE VALUE OF CHOICE AND THE CAUSAL THESIS

I turn now to the question of whether choice will retain the
moral significance which the Value of Choice theory assigns it if
the Causal Thesis is true. Whether it does so or not will depend on
30Punishment and Responsibility, p. 218.
[SCANLON] The significance of Choice 207
whether choice will retain its value for an individual if the Causal
Thesis is true. This is at least part of what I called in my first lec-
ture the personal problem of free will. So it seems that the
most that the Value of Choice theory could accomplish would
be to reduce the political problem of free will to the personal
problem.
The mere truth of the Causal Thesis would not deprive choice
of its predictive value: a persons choices could remain indicative
of his or her future preferences and satisfactions even if they had
a systematic causal explanation. Nor, it seems to me, need the
demonstrative value of choice be undermined. A persons choices
could still reflect continuing features of his or her personality
such as feelings for others, memory, knowledge, skill, taste, and
discernment.
This is how things seem to me, perhaps because I am in the
grip of a theory. It is difficult to support these intuitions by argu-
ment because it is difficult, for me at least, to identify clearly the
basis of the intuitions which move one toward the opposite con-
clusion. It might be claimed that what I have called the demon-
strative value of choice would be undermined because the feelings,
attitudes, and so on which a persons choices might be taken to
reflect will no longer belong to that person if the Causal
Thesis is true, but it is not clear why this should be the case. It is
easy to see that particular kinds of causal history might make a
belief or desire alien. This would happen when, as in the
implantation examples mentioned above, the special causal
genesis of a belief meant also that it lacked connection with the
persons other conscious states that it was not all dependent on
other beliefs and desires for support and not subject to modifica-
tion through the agents process of critical reflection. But it does
not seem that this kind of loss of connection need hold generally
if the Causal Thesis is correct.
One can certainly imagine a form of causal determination
which would make this kind of alienation hold generally and
would make it inappropriate to speak of a persons holding beliefs

and attitudes at all. A persons conscious states might be caused
to occur in a pattern which made no sense at all from the inside,
following one another in a random and meaningless sequence pre-
serving no continuity of belief or attitude. It might be argued that
the normal case is more like this than we are inclined to sup-
pose: that our idea of the coherence and regularity of our con-
scious life is to a large degree an illusion. This might undermine
the sense of self on which the value of choice depends. But this,
if true, would be the result of a particular substantive claim about
the order and coherence of the events that make up our mental
lives. It would not be a consequence of the bare Causal Thesis
itself.
4. FREEDOM AND OVERDETERMINATION
The kind of freedom required by the Value of Choice theory
is in one respect more extensive than that required for moral
appraisal of the kind discussed in my first lecture. This difference
can be brought out by considering how the ideas of quality of will
and value of choice apply to overdetermination cases of the kind
introduced by Harry Frankfurt.31 Frankfurts central example in-
volves two drug addicts. It is assumed that neither is capable of
resisting the pull of his addiction: both will take the drug when it
is offered, and neither could do otherwise. But while one, the
unwilling addict, would prefer that the desire to take the drug
not be the one which he acts on, the other, the willing addict,
not only has a desire for the drug but also has the second-order
desire to act on that desire. Frankfurt believes that the latter
addict acts freely in the sense required for moral responsibility but
that the former does not. What interests me here is the fact that
the two theories I have presented appear to give different answers
to the question of freedom in cases like that of Frankfurts willing
addict that is to say, cases in which (for reasons which may or
31In Freedom of the Will and the Concept of a Person.
may not be like those in Frankfurts particular example) a person

has no alternative to doing a certain thing but nonetheless gets
what he wants or does what he is inclined to do. If the question
is whether the action reflects the agents quality of will, then cases
like that of Frankfurts willing addict seem to be cases of freedom.
(This answer agrees with Frankfurt, which is not surprising given
that he is concerned specifically with moral responsibility.) If, on
the other hand, the question is whether the agent has been given
a fair chance to make outcomes conform to or exhibit his or her
preferences and abilities, then the answer seems to be no, and the
cases count as instances of unfreedom.
It may seem that this difference is illusory. The question under
the Value of Choice theory is whether there was the right kind of
opportunity for the persons disposition to choose to be discovered
and registered. Insofar as it is predictive value we are concerned
with, the assumption is that we do not generally know in
advance what a persons preference is: we are trying to set up a
social mechanism to discover this and react to it. In Frankfurts
cases, however, it is assumed that we know the addicts (first- and
second-order) preferences. Indeed, we are assumed to know more
about this than agents themselves normally do. The question of
how these preferences might be discovered is not at issue in
Frankfurts discussion. But this question can arise with respect to
moral responsibility. Administering praise and blame is something
we do, and it is relevant to ask whether we have adequate grounds
for doing so: whether it is fair to judge a person on the basis we
have. This is like the question which arose in application of the
Value of Choice theory: whether there was adequate opportunity
for the persons preferences, whatever they may have been, to be
revealed.
This same question of fairness can also be raised when we are
only forming an opinion about an agents blameworthiness, with-
out intending to express it. But the question whether the agent is
blameworthy goes beyond these questions of adequate grounds,
and it is the question which is fundamental: if the persons will in

doing the action was of the appropriate sort, then a certain moral
judgment is in fact applicable, whether or not any particular per-
son is in a position to make it. Insofar as this is the case, the dif-
ference between the two theories that was pointed out above still
stands.
Of course, parallel to the fact that a person really was blame-
worthy in acting a certain way, there is the fact that a person
really did want X, which was what he got, and this too might
be held to be the fundamental fact, on the basis of which we could
ask, How can he complain, since he got what he wanted? But this
fact of preference is not fundamental in the way that the fact of
blameworthiness is: the two facts are differently related to the
moral ideas on which the theories in which they figure are based.
The Quality of Will theory is based on the idea that the applica-
bility of moral praise and blame depends on what the quality of
will expressed in an action actually was. In determining this
quality we may need to know what the agent believed the alterna-
tives to be, but the question of which of these were actually avail-
able is in at least some cases irrelevant. Under the Value of Choice
theory, however, the basic moral idea is not simply that people
should get what they want but that things should be set up so that
outcomes are made dependent on peoples choices. In overdeter-
mination cases this demand may not have been met, even though,
as it happens, the person is in certain respects no worse off as a
result.
5. THE TWO THEORIES COMBINED
I have described two theories and said something about how
they are related to one another. It remains to be seen how these
two theories, when combined, cover the territory. I have so far
employed the Value of Choice theory mainly to give an account
of the significance of choice in political cases, and I have relied
upon the Quality of Will theory in discussing moral responsibility.
But this division of labor is overly simple. In fact, both analyses

are required to account for the significance of choice in morality,
and both are required to explain its force in the law.
Let me take the moral case first. Suppose you think that I
promised on Monday to pick up your child at school on Tuesday
but then failed to do this. There are two ways in which considera-
tions of voluntariness and choice might enter into an assessment
of how blameworthy I am on this account. First, such considera-
tions could undermine my blameworthiness by making it the case
that I had no obligation to pick up your child in the first place. It
could be that I never assented to your request: when I said yes, it
was to something else, and I never heard your request at all.
Or perhaps I did assent to your request but only because you
threatened me or concealed from me the fact that I would have to
wait three hours beyond the normal end of the school day. Factors
such as these could erase or modify my obligation.
On the other hand, it could be that while I did indeed incur an
obligation to you, my not meeting your child was not due to any
failure on my part to take my obligation seriously and try to ful-
fill it. It might be that I was hit over the head and knocked un-
conscious just before I was to leave, or that my car broke down on
the way, leaving me stranded in a deserted spot.
These two kinds of excusing conditions are quite different.
Something like the Value of Choice theory seems to provide the
best explanation of why moral obligations are qualified by restric-
tions of the first sort. As Hart suggested, a system for the making
of binding agreements, whether moral or legal, is defensible only
if it is constrained by restrictions to ensure that the obligations
one acquires are obligations one judges to be worth acquiring.
The assessment of quality of will has at most a secondary role
here.
Things are reversed in a case of involuntary nonfulfillment of
a valid obligation. Here the natural value of choice analysis
(modeled on that analysis of the choice requirement for criminal
punishment) would be that a morality which held agents liable

to blame in such cases would be objectionable because it gave
people insufficient protection against incurring the sanction of
moral blame. This is clearly not the right explanation. It is wrong
because it treats moral blame simply as a sanction which people
would like to avoid, which we attach to certain actions although
it could just as well be attached to others (eg., to things that are
done involuntarily). This ignores the distinctive content of moral
blame, in virtue of which it is not simply another kind of un-
pleasant treatment, like being shunned. Morality is, at base, a
system of mutually authorizable deliberation. To feel oneself sub-
ject to moral blame is to be aware of a gap between the way one
in fact decided what to do and the form of decision which others
could reasonably demand. The absence of such a gap is by itself
a sufficient explanation of why blame is inapplicable in cases like
that of the person who, despite his or her best efforts, fails to pick
up the child. There is no need to refer to the kind of question
which the Value of Choice theory addresses.
This internal connection between the nature of the moral
sanction and the content of morality between the nature of
blame and the things one can be blamed for differentiates
morality from a social institution set up to serve certain extrinsic
purposes. Of course there could be a social practice according to
which people would be subject to scolding and shunning in cases
for actions involving no faulty willing or deliberation, but what
was expressed by this behavior would not be moral blame. Even
without such a practice there is a question, distinct from that of
blameworthiness, of whether one has good reason to engage in
blaming behavior toward a given person on a given occasion.
As I mentioned in my first lecture, even when people are blame-
worthy it might be callous to scold them, and the reverse may also
be true. For example, even though very young children are not
blameworthy it may be important for their moral education to
treat them as if they were.
The issues raised here are similar to those which arise in con-
nection with what Hart called the definitional stop argument
against exemplary or vicarious punishment of persons known to be
innocent of any off ense.32 A utilitarian justification of punish-
ment, insofar as it is a justification of punishment, could not
justify such practices, this argument ran, because these practices
do not count as punishment, which, by definition, must be of an
offender for an offense. The obvious response to this argument is
that it is not important what we call it; the question is why it
would not be permissible to subject people, known to be innocent,
to unpleasant treatment (prison, fines, etc.) as part of a scheme
to intimidate others ,into obeying the law. As I have said above,
I agree with Hart that the Value of Choice theory provides a good
(though perhaps not fully satisfying) answer to this question.
With respect to moral blame, however, I have responded in effect
that it matters a great deal what you call it, because blameworthi-
ness, rather than any form of blaming behavior, is the central
issue. There is also, of course, a question of the desirability and
permissibility of expressing or administering blame in a certain
way, but this is a separate question and a secondary one.
In the case of criminal punishment this emphasis is reversed:
the main question is whether we can justify depriving people of
their property, their liberty, or even their lives.33 Despite the
32Punishment and Responsibility, pp. 5-6.

33In a recent article, R. B. Brandt put forward something like the Quality of
Will theory as a limitation on legal punishment. See A Motivational Theory of
Excuses in the Criminal Law, in J. R. Pennock and J. W. Chapman, eds.,
NOMOS XXVII: Criminal Justice (New York: New York University Press, 1985),
pp. 165-98. Specifically, Brandt defends the principle that a condition should be
recognized as excusing a person from legal blame if the presence of that condition
blocks the normal inference from the fact that the agent performed a certain act
to the conclusion that the agents motivation is defective. His defense of this prin-
ciple appeals to the value of assuring people that if they lack defective motivation
they will almost certainly not be punished. This is reminiscent of Hart and the
Value of Choice theory, but Brandts defense is avowedly rule-utilitarian: he sees
the value in question merely as a contribution to the general welfare, not as ful-
filling a special requirement of fairness to the individual. Moreover, he sees the
requirement of defective motivation as a replacement for Harts notion of capac-
changed emphasis, however, both elements are still present, and

consequently it does matter what you call it even if this con-
sideration does not settle the crucial question of justification. The
law is not just an organized system of threats. It also provides
rules and standards which good citizens are supposed to respect,
that is, to employ as a way of deciding what to do not simply
as a way of avoiding sanctions but as a set of norms which they
accept as reason-giving. This important feature of law offers a
further reason why the Value of Choice theory was not completely
satisfying as an explanation of the choice requirement for criminal
punishment. Insofar as punishment is in part an expression of
legal blame, as Feinberg and others have pointed out,34 there
is a special inappropriateness in having it fall on persons who
have deliberated and acted just as the law says they should. The
Value of Choice theory thus fails to be a complete account of the
significance of choice in the law for much the same reason that it
fails to be a complete account in the case of morality. In each
case there is something to the definitional stop.
Something, perhaps, but in the case of the law, how much?
Pointing out the expressive function of punishment helps us to
understand our reactions to punishing particular kinds of people,
but what role if any does it have in the justification of punish-
ment? It seems to have no positive role in justifying hard treat-
ment of the legally blameworthy. Insofar as expression is our aim,
we could just as well say it with flowers or, perhaps more
appropriately, with weeds. Nor, it seems, is this idea the central
explanation of the apparent wrongfulness of punishing, say, young
children or the mentally ill. Assuming that these people lack the

ity and fair opportunity to avoid punishment (ibid., p. 180). My analysis is simi-
lar to Brandts in a number of respects, but, unlike him, I see quality of will and
the value of choice as two independent (though related) reasons for the limits of
moral and legal blameworthiness. Since they are related, it is not surprising that
these two kinds of reasons often support the same limits. But they do not always
do so.
34Joel Feinberg, The Expressive Function of Punishment, in Doing and
Deserving (Princeton: Princeton University Press, 1970).

capacity for critically reflective, rational self-governance, we could

argue, as we did in the case of morality, that they cannot be legally
blameworthy. But even in the case of morality, the justification of
blaming behavior is a separate issue from that of blameworthi-
ness, and here it is a much weightier one in view of the losses that
the law can inflict.
The Value of Choice theory offers a more plausible explana-
tion. According to that theory the lack of the normal capacity for
critically reflective, rational self-governance is relevant because
people who lack it are so unlikely to be deterred. This may or may
not make punishment pointless for us, but it certainly makes it
unfair to them: we must protect them against punishment just as,
in my other example, we must post barriers or guards to keep
people with Alzheimers disease away from the hazardous waste.
But within the Value of Choice theory the normal capacity for
critically reflective, rational self-governance lacks the distinctive
importance which it has when moral (or legal) blameworthiness
is at issue. There are many people who have this capacity yet will
not be deterred. It is easy to say why they are blameworthy, but
why should we respond differently to their suffering than to that
of the mentally ill? W e can say that, because they have this normal
capacity for self-governance, deterrence is a plausible strategy for
us to use in dealing with them and that the possibility of their
being deterred is, from their point of view, some measure of pro-
tection. If it turns out not to be enough, then the best we can say,
if it is true, is that we did as much as we could be expected to do
to protect them.
At some moments it seems to me that we must be able to say
more that choice has a further significance not captured by
either of the theories I have considered, perhaps something more
like what the Forfeiture View is straining toward. At other times,
however, it seems to me an advantage of the combined theory I
have been defending, and a natural consequence of its aspiration
to be compatible with the Causal Thesis, that it leaves us in this
position: moral and (if there is such a thing) legal indignation

toward lawbreakers is entirely in order, and the sufferings we
inflict upon them may be justified. But in justifying these suffer-
ings, and inflicting them, we have to say not You asked for this
but There but for the grace of God go I.

The Significance of Choice: T T L H V

Uploaded by

Copyright:

Available Formats

The Significance of Choice: T T L H V

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Significance of Choice: T T L H V

Uploaded by

Copyright:

Available Formats

The Significance of Choice

T HE T ANNER LECTURES ON H UMAN VALUES

May 16, 23, and 28, 1986

2. THE PROBLEMS OF FREE WILL

were filled by something other than the cumulative effects of

It has sometimes been maintained that even if the Causal

the one hand, explains why the significance of choice is under-

pp. 72-74, 104.

try to correct that later.) Replying to an argument for the justice

certain conditions as incompatible with this important freedom

dom of Expression and Categories of Expression, University of Pittsburgh Law

Oxford University Press, 1968), pp. 71-96.

3. THE INFLUENCEABILITY THEORY

The Influenceability theory might explain why a utilitarian

4. QUALITY OF WILL: STRAWSON'S ACCOUNT

9Broadening the theory to take into account the possibility of influencing

hand, is what he calls a vicarious attitude: a reaction to the

ate. If I learn, for example, that you stepped on my foot by acci-

argument was aided by Jonathan Bennetts perceptive analysis in Accountability.

suggestion that these factors render reactive attitudes such as

Like the unsuccessful defense of common sense mentioned

But there is another reply which is suggested by something

13 Strawson, Freedom and Resentment, p. 82.

in giving a convincing account of the requirements of freedom

15Compare Thomas Nagels comments on Strawsons theory in The View from

ticular reactive attitude. For example, I may believe that an action

5. QUALITY OF WILL: A CONTRACTUALIST ANALYSIS

require. This can be so either because one has failed to attend

6. THE SPECIAL FORCE OF MORAL JUDGEMENT

contractualist view I have described leaves itself open to the objec-

restrict its use to cases in which this important secondary function

intent is not essential. What is essential, on this account, is that a

and attitudes are correctly attributed to an agent and then asks

self, are appropriate. Insofar as I think of a past intention, deci-

presentation of this lecture in Oxford.

7. BLAMEWORTHINESS AND FREEDOM

specifically moral component is the ability to reason about what

generally believed to be blocked by some forms of mental illness.

Journal o f Philosophy 6 8 (1971): 5-20; Wright Neely, Freedom and Desire,

the scope of a capacity of self-governance which normally makes a difference in

In Frankfurts terms, these restrictions correspond roughly to a

This theory offers a convincing and unified account of familiar

1. THE VALUE OF CHOICE

significance of the choices people make. In this lecture, however,

or my ability to swallow highly seasoned food. Thus the same

awareness, or memory, or ones imagination and skill. Many of

people have for wanting outcomes to be (or sometimes not to be)

categories of choice which we take to be a fair starting point for

23The variability of the value of choice is pointed out clearly by Gerald

Dworkin in Is More Choice Better Than Less? in P. French, T. Uehling, and U

what I have here called predictive and demonstrative value, the

to be reasonably rejectable for that reason. Let me illustrate by

2. JUSTICE AND CHOICE

Consider first the economic justice example which I mentioned

The background conditions under which choices are made in

The problem which Nozick raises here is a version of the

provides no assurance that these factors will not be ones which

the effect of disentangling the idea of individual liberty from