In contemporary philosophy, philosophers once again are taking

advertisement
Free Will Denial, the Original Position, and Punishing the Innocent
In contemporary free will theory, a significant number of philosophers are once again
taking seriously the possibility that we do not have free will, and that we are therefore not
morally responsible for our actions. In this talk, I will not be offering a new argument for
denying free will and moral responsibility. I think there are enough in circulation to support the
view that the position is tenable, and if this is right, its implications are worth exploring in their
own right. In this talk I will consider a problem which arises in the theory of punishment when
we reject moral responsibility. I will next offer a solution to it based on Rawls’ idea of the
original position, and then I will reflect on the broader ramifications of this approach for
reconciling free will denial and desert. I will conclude with a brief critique of Saul Smilansky’s
approach to reconciling free will denial and desert.
Giving up moral responsibility would require us to reject retributivism about punishment,
which has been and continues to be the dominant theory of punishment. Retributivism, as it will
be understood here, is the view that agents are morally responsible for their actions, and because
they are morally responsible for their actions, they deserve to be punished for them when they fit
certain criteria, such as being crimes. This means that if we are not morally responsible for our
actions, retributivism must be rejected.
This is a problem for philosophers who reject moral responsibility, because it seems
likely that without punishment as a check on criminal behavior, society could not function. But
if we cannot justify punishment, yet we go on punishing people anyway in order to keep society
functioning, then social functioning comes at the price of accepting a fundamentally immoral
practice. I will argue that punishment can in fact be justified without the concept of moral
responsibility. (Derk Pereboom has also discussed this issue, in Living Without Free Will, but he
takes a different approach and comes to different conclusions, and therefore I will not be
addressing his approach.)
The first step in justifying punishment is to adopt a theory the basic thrust of which is
consequentialist, rather than retributive. According to consequentialism about punishment,
punishment is justified because of its good consequences, not because criminals are morally
responsible for their actions and therefore deserve punishment. In other words, if we are
consequentialists about punishment, the reason we punish criminals is because punishing them
produces a good outcome. If we imprison a criminal, he suffers, and according to
consequentialism, his suffering is as morally bad as anyone else’s. But the consequences of
imprisoning him are typically such as to prevent significantly more suffering than the criminal
experiences: his imprisonment prevents him from continuing to commit crime, deters him to
some degree from repeating his crime upon his release, and deters other potential criminals from
committing the crime he committed. Since consequentialism about punishment does not rely on
the idea that the criminal is morally responsible for his crime, we can endorse it even if we reject
moral responsibility.
But as it stands, consequentialism about punishment faces a deep problem which has
been noted by many. Suppose an official could perfectly frame a celebrity so that everyone but
the official and the celebrity would believe the celebrity committed the crime, and then have the
celebrity punished instead of the criminal who actually committed the crime. And suppose the
punishment of the celebrity would be covered in every media outlet, while the punishment of the
actual criminal would not strike anyone as interesting enough to report at all. Suppose that
because of its thorough coverage in the media, the punishment of the celebrity would be a very
2
effective deterrent, and would therefore have much better consequences than the punishment of
the actual criminal. If we are consequentialists about punishment, how can we justify punishing
the real criminal instead of the perfectly framed celebrity?
It might be supposed that giving up moral responsibility means that it is in fact
appropriate to punish the framed celebrity instead of the real criminal, because the source of the
intuition that it is wrong to punish the framed celebrity is just the intuition that the actual
criminal is morally responsible for the crime. But I think this is a mistake. As I will go on to
argue, even according to everyday moral intuitions, there are kinds of desert-claims we can
legitimately make just because we are persons, even if we are not morally responsible for our
actions and we therefore cannot legitimately make any desert-claims on the basis of our actions.
Desert-claims based on the sheer fact of personhood are not jeopardized when we reject the
concept of moral responsibility, because it would be mistaken to suppose that we could be
morally responsible for being persons. It is simply because we are persons that we deserve not to
be punished for crimes we did not commit. This point allows us to justify punishing the actual
criminal despite rejecting the concept of moral responsibility. I will explain this in more detail in
terms of Rawls’ idea of the original position, which offers an approach to choosing the basic
principles of society in which only the sheer fact of personhood is relevant.
Before I explain the original position strategy in detail, however, I want to mention why
I’m not using a strategy commonly claimed to solve problems of this sort in consequentialism,
that is, the distinction between act- and rule-consequentialism. Rule consequentialism tries to
make room for absolute moral rules which should be followed even when the consequences of
following them are worse than the consequences of breaking them. Rule-consequentialists hold
that rules are tested by appeal to their consequences, but actions are not tested by appeal to their
3
consequences, only by appeal to the rules. Act-consequentialists, on the other hand, think
actions are tested by appeal to their consequences. Act-consequentialists could not accept a
moral rule unless it was certain that the consequences of following it would always be better than
the consequences of breaking it. With respect to punishment, rule consequentialists would point
out that punishment can only produce the good consequence of deterrence if it is generally
believed that the person punished is the person who actually committed the crime. The rule for
punishment which is most likely to secure this general belief is to punish only the person who
has actually committed the crime. Since rule-consequentialists hold that only rules are tested by
appeal to consequences, they will hold that the good consequences of punishing the celebrity are
not morally relevant. All that is morally relevant to punishing the celebrity is that it breaks the
rule that only actual criminals are to be punished, so it is immoral.
If the rule-consequentialist solution were acceptable, our task would be done. But it is
not acceptable. Certainly, in some cases, we do not have time to evaluate the consequences of
particular actions, and we are forced to use heuristic rules, i.e. rules of thumb. If we are unsure
about the consequences, the safest bet is to punish the actual criminal. But in cases where we do
have time to evaluate the consequences of a particular action, as the official contemplating
punishment of the celebrity does, there is no need to appeal to heuristic rules to expedite
decisions. Also, since the celebrity can be perfectly framed, there is no chance that punishing the
celebrity will undermine the public’s belief that only actual criminals are punished. In such a
case, why should we accept the idea that we test only rules, not actions, by appeal to their
consequences? It is arbitrary to simply stipulate that this is right. Rule-consequentialists could
offer more than stipulation if they infused some sort of deontological authority into their rules.
Then they could hold the following: even though the only justification for moral rules is that they
4
have better consequences than any other rules, we have a moral duty to follow them even when
the consequences of breaking them are better than the consequences of following them. But
absurdity would seem to threaten here, and in any case the deontological element makes it
unclear whether this theory is a really a form of consequentialism. I of course do not wish to
claim to have refuted rule-consequentialism with such a short argument. But it seems fair to say
that the response rule-consequentialism offers to this problem about punishment is not so strong
that alternatives are not worth considering.
For these reasons, I will simply accept that there must be limits on consequentialism in an
acceptable theory of punishment, and I will look to Rawls’ idea of the original position for help
in describing those limits. The original position is an idealized standpoint for choosing the
principles of a just social arrangement. The original position is formed by drawing what Rawls
calls a “veil of ignorance” between the persons who make up a society and all the particular
characteristics of those persons which determine their social roles. In the original position, we
cannot know whether we are among the best off or worst off, what our religion, ethnicity, or
gender are, or—and this is crucial for the issues being considered in this paper—what character
traits and patterns of action we exhibit, for example, whether we are industrious or lazy. Rawls
holds that since people are risk-averse in conditions of extreme uncertainty, deliberators in the
original position will choose principles which make the situation of the worst-off member of
society the best possible.
Rawls designs the original position to screen out knowledge of all particular
characteristics of persons because he thinks they are morally irrelevant to choosing principles of
justice. As he argues in section 17 of A Theory of Justice, these characteristics are morally
irrelevant because we do not deserve to have them, and therefore we cannot deserve to have any
5
advantages they might confer upon us, so we do not deserve to have a social arrangement in
which they confer advantages upon us. Rawls presents this as continuous with everyday moral
intuitions. But the fact that he extends his critique of desert so far as to deny even that we
deserve the fruits of our industriousness would seem to place him closer to free will rejectionists
than to proponents of everyday moral intuitions. This suggests that the original position and the
conclusions about ethics to which it leads should be of interest to free will rejectionists. Rawls
does not reject all desert-claims, however. He argues that
given a just system of cooperation as a scheme of public rules and the
expectations set up by it, those who…have done what the system announces that
it will reward are entitled to their advantages. In this sense the more fortunate
have a claim to their better situation; their claims are legitimate expectations
established by social institutions…But this sense of desert presupposes the
existence of the cooperative scheme[.]
Rawls’ description of the foundation of desert in his philosophy in this passage is a bit odd,
however. He makes it sound as though individual persons considered apart from social
institutions do not deserve anything, and that desert is instead an emergent property of just social
institutions. But this cannot be Rawls’ overall view, because he is clearly committed to the view
that we can make desert-claims apart from actually being members of just social institutions.
That is, it seems clear that he thinks we deserve to be treated in accordance with the principles
that would be chosen in the original position. The reason we can come to deserve things within
just social institutions is that they are just, not that they are social institutions, and what makes
them just is that they implement the principles that we would choose in the original position.
The question, then, is what sort of desert-basis is available in the original position? Since
the “veil of ignorance” screens out knowledge of all particular characteristics of persons, we are
left only with the sheer fact of personhood, that is, the universal characteristics of persons, to
serve as a desert basis. We need not attempt to list the universal characteristics of persons for
6
present purposes. Presumably they would include characteristics such as setting and pursuing
ends, having rights and needs, and requiring collaboration with others. One virtue of the original
position approach is that the universal characteristics of persons are implicitly specified, so to
speak, through the procedure of original position deliberation.
I will say more about the sheer fact of personhood as a desert-basis shortly, but at this
point I will turn to the task of applying original position deliberation to punishment. Rawls
himself does not do this. In section 48 of A Theory of Justice he implies that original position
deliberation cannot be applied to punishment, because the purpose of punishment is to sanction
“bad character”. But a criminal no more deserves his criminal character than an industrious
person deserves her industrious character. This asymmetry in Rawls’ treatment of distributive
justice and punishment has been discussed recently in separate papers by Samuel Scheffler,
Jeffrey Moriarity and Eugene Mills. Scheffler give a qualified defense of the asymmetry.
Moriarity and Mills reject it, as I do, but they think this means Rawls’ critique of desert should
be rejected in the case of distributive justice, rather than extended to punishment, as I seek to do
in the present paper. James Sterba has previously discussed original position deliberation about
punishment. But he holds that we should choose principles of punishment in light of the fact that
the criminal “would have been able to avoid his fate if he had chosen to abide by the laws of the
society” (Sterba, “Retributive Justice”, 1977). But of course this is mistaken if we reject free
will, and as just argued, it seems mistaken even according to Rawls’ own critique of desert.
Having acknowledged these other discussions, let us proceed. What sort of institution of
punishment would deliberators in the original position choose? Let us just stipulate that they are
all free will rejectionists, so that we remove any temptation for them to assume with Sterba that
if they turn out to be criminals, they must have freely chosen not to abide by the laws. Since the
7
deliberators are risk-averse, they would choose principles that made the lot of the worst-off
members of the institution the best possible. The worst-off would be the people punished and
the victims of crime. One might think that the best possible lot for the punished would be not to
be punished at all. But we have assumed that punishment is necessary if there is to be a society
at all. Further, deliberators will worry about being victims as well as criminals. So they will
choose to have an institution of punishment. It will be broadly consequentialist, since rejecting
free will means rejecting retributivism. But they will insist on limits on the amount of suffering
that can be imposed by punishment, and they will be willing to forgo the increased deterrence
that could be achieved by severe punishments.
Will the deliberators choose a principle of only punishing actual criminals even when
punishing other people would be a much better deterrent? The answer to this may not be
immediately obvious. Certainly it would be no loss to actual criminals if perfectly framed people
were punished instead. So if the deliberators choose the institution that makes the lot of
criminals the best possible, they will choose the institution that punishes whomever it is most
effective to punish, even if it is not the actual criminal. This is why it is important to specify that
it is the people punished who are intrinsically among the worst off in the institution of
punishment—criminals are not intrinsically among the worst off. So we must compare the
amount of suffering inflicted by punishment under two alternative institutions of punishment,
one that sometimes punishes perfectly framed people, and one that only punishes actual
criminals. We can do this by asking which of the following two punished people are made to
suffer more: an actual criminal who spends five years in a prison cell, or a perfectly framed
person who spends five years in an exactly similar prison cell? The actual criminal only has to
deal with the suffering of imprisonment. The perfectly framed person is worse off, because she
8
must deal with the suffering of imprisonment, as well as the suffering of being manipulated.
That is, deterrence through the punishment of non-criminals requires framing, which is a form of
manipulation, and it should be uncontroversial that the suffering involved in being manipulated
is considerable. So the lot of the punished is better if we choose the principle “only punish
actual criminals”, and therefore this is the principle which would be chosen in the original
position.
At this point, I want to address in a more general way the distinction I have proposed
between action-based desert and personhood-based desert. Our actions can be desert-bases only
if we are morally responsible for our actions. Action-based desert claims include claims about
praise and reward on the one hand, and blame and punishment on the other—i.e. agents are
claimed to deserve praise and reward, or blame and punishment, because they are morally
responsible for having acted in certain ways. Action-based desert claims also include desert
claims about property when they are construed in Lockean terms, i.e. in terms of the idea that we
become entitled to property by mixing our labor with objects. Action-based desert claims also
include desert claims that might initially appear to be character-based, because we only make
desert-claims on the basis of character-traits if we suppose that we have acted in such a way as to
develop or cultivate those traits.
Personhood-based desert claims, on the other hand, are desert-claims that can be made
without reference to any actions at all. The idea of personhood-based desert can seem
mysterious, so examples are useful. Deserving to be equal before the law, deserving to have
one’s rights respected, and deserving to have one’s vote count equally do not depend upon
having acted in any way at all. They depend only upon being a person. Different ethical theories
will explain in different ways why being a person grounds these desert claims, and I will bracket
9
further discussion of this issue for purposes of brevity. It seems safe to say, however, that all
ethical theories will feel pressure to account for this second category of desert claims based
simply on personhood.
Free will denial undermines action-based desert claims, but not personhood-based desertclaims. Action-based desert claims involve the belief that we are morally responsible for our
actions, and according to everyday moral intuitions, we can be morally responsible for our
actions because they are up to us in some deep sense, such that we could have acted otherwise.
Free will denial demands that we reject this idea, and it therefore demands that we reject the idea
that actions are desert-bases.
Free will denial does not similarly undermine personhood-based desert claims, however.
Is it any less plausible that we deserve to be equal before the law, to have our rights respected, or
to have our votes count equally if we do not have free will than if we do? It would seem not.
This is because these desert claims depend only upon being a person, and even according to
everyday moral intuitions, we could not be morally responsible for being persons, since no one
could plausibly suppose that we could have done otherwise that be persons. (If one accepts free
will, one might plausibly suppose that we are morally responsible for continuing to be persons,
since that depends on how we answer Hamlet’s question. But that is very different from being
morally responsible for being a person in the first place, and it is only once we are persons that
we can become morally responsible for continuing to be persons.) So personhood remains solid
as a desert-basis even if we reject moral responsibility.
One cannot simply reject action-based desert and stop there, however. One cannot
distinguish between the perfectly framed celebrity and the actual criminal in terms of the sheer
fact of their personhood—the fact that the criminal committed the crime must be part of the
10
explanation of why it is appropriate to punish him and not the celebrity. Also, more generally,
social functioning requires rule-governed ways of determining which persons are entitled to
which things on the basis of their actions. This means we need analogues of action-based desert.
How does this work? Original position deliberation allows us to begin with personhood-based
desert and reconstruct legitimate analogues of action-based desert. In the original position, we
identify the basic principles in accordance with which we deserve to be treated, based on the
sheer fact of our personhood. I have argued that one such principle is that only actual criminals
are to be punished. According to Rawls, another such principle is that economic inequalities are
just if and only if they improve the lot of the worst-off. He calls this the “difference principle”.
It implies that it is just for the industrious to reap advantages from their industriousness if the
economic dynamics this produces improve the lot of the worst off. We deserve to be treated in
accordance with these principles because we are persons. But these principles can only be
implemented in society by reference to our actions. We determine who is to be punished and
who is entitled to the fruits of industry by reference to actions, but the actions themselves are not
desert-bases. Instead, they are merely necessary reference-points for the implementation of
principles whose desert-base is personhood.
This completes the theory I want to propose here. To conclude, I will offer a brief
critique of Saul Smilansky’s theory, which incorporates both hard determinism and
compatibilism. In his paper “Free Will, Fundamental Dualism, and the Centrality of Illusion”,
Smilansky describes a case where an industrious waiter receives bigger tips than a lazy waiter.
Smilansky accepts hard determinism, but he claims that we must nonetheless acknowledge that
the lazy waiter does not deserve to get tips as big as those of the industrious waiter. He thinks
this demands that we hold the lazy waiter morally responsible for his lazy behavior, and he
11
thinks we must therefore accept a compatibilist analysis of such cases. He thinks this leads us to
a not-quite-contradictory embrace of both compatibilism and hard determinism, which we must
maintain by deluding ourselves about the tension between the two. Smilansky’s approach is bold
and interesting, but if there is a simpler way of maintaining hard determinism alongside the
recognition that the industrious waiter is in some sense entitled to his bigger tips, it would surely
be preferable from the standpoint of theory structure. The theory I have just advanced fits this
bill. We need not hold, with Smilansky, that the industrious waiter is morally responsible to hold
that he is entitled to the bigger tips. Instead, we can reject moral responsibility across the board,
but hold that since the industrious waiter is a person, he deserves to be treated in accordance with
the principles that would be chosen in the original position. According to Rawls, the difference
principle would be chosen, i.e. the principle that economic inequalities are just if they improve
the lot of the worst-off. Since we have reason to believe that it increases overall productivity to
allow the industrious to benefit from their industry to some degree, and that increasing overall
productivity tends to improve the lot of the worst-off, we have reason to believe that it is just to
allow the industrious to benefit from their industry to some degree. The industrious waiter
becomes entitled to his bigger tips because he acquires them by following the rules of a just
economic system. His industrious actions do not themselves serve as desert-bases, but are
merely reference-points for the implementation of principles whose desert basis is the sheer fact
of personhood. This approach allows us to resolve the tension between free will denial and
desert without sacrificing the unity, simplicity and consistency of our theoretical framework, and
I think this tells strongly in its favor.
12
Download