source

Rawls, “Two Concepts of Rules”

Act-Utilitarianism: An act is wrong if and only if it would fail to produce as much welfare as
any alternative act open to the agent.
Rule-Utilitarianism: An act is wrong if an only if it would be forbidden by the set of rules
whose (universal? near universal?) adoption would produce the most welfare.
Utilitarianism about Institutions: Our institutions should be designed so that they benefit
society, by producing at least as much welfare as any other design would have produced.
• Which version of utilitarianism is Rawls defending?

Punishment

Two theories of just punishment:
1. Retributivist: Punishment is justified if, only if, and because the person who does wrong
deserves to be punished. A world in which people suffer in proportion to their wrongs is
morally better than one in which people don’t.
Some features of the retributivist view:
• the punishment must be proportional to the crime
• we are never justified in punishing an innocent person
• in evaluating the justifiability of punishment, we should look backwards, to see what the
person to be punished did, not forwards, towards the consequences of our punishment
2. Utilitarian: Punishment is justified if, only if, and because punishing someone will lead to at
least as much welfare as not punishing. The severity of the punishment should be
determined by what level of punishment would have the best effects.

Some features of the utilitarian view:
• because punishment causes (at least temporary) reduction in the welfare of the person
being punished, we should always choose the minimal punishment necessary to achieving
the good effects
• in evaluating the justifiability of punishment, we should look forwards, towards the
consequences of our punishment, not backwards, to the crimes committed
An apparent problem for the utilitarian view:
• might it justify punishing an innocent person? (See Carritt’s example, p. 10)
2
Rawls’ suggestion: The utilitarian approach to punishment is appropriate when
evaluating/justifying the institution or practice of punishment; the retributivist approach is
appropriate when evaluating/justifying a particular decision within the practice. In other words,
utilitarian arguments justify adopting a practice of punishment that is governed by retributivist
rules.
• Question: Would utilitarian considerations lead us to adopt an institution of punishment
governed by retributivist rules? Would utilitarian considerations definitely lead us to
reject “telishment”? Would an institution of punishment designed on the basis of
utilitarian considerations look any different from the retributivist model? (If they look
the same, might they do so only contingently?)
• Question: Would it be possible/appropriate to justify the institution of punishment on
purely retributivist grounds? (The proportionality rule underdetermines the severity of
punishment, and “and eye for an eye” seems a bit barbaric.)
• Rawls suggests that different roles for the utilitarian and retributivist justifications for
punishment can be brought out by considering the difference between two kinds of
questions:
(i) Q: Why was J thrown in jail? (What justified throwing J in jail?)
A: He robbed a bank. (Backward-looking, retributivist justification)
(ii) Q: Why throw bank robbers in jail?
A: Doing so deters future crimes, and so is for the good of society. (Forwardlooking,
utilitarian justification)
Are these really two different kinds of questions, demanding two different kinds of
answers? Mightn’t we ask the first, even if we knew J robbed a bank?
Promising
The problem for utilitarians: It seems like a utilitarian will have difficulty accounting for our
intuition that we are not justified in breaking a promise just because doing so will have the best
effects overall (even taking into account the negative effects on the practice of promising),
because the act-utilitarian principle, applied to individual decisions about when to keep or break
a promise, would tell us to break it when doing so would have the best effects overall. (Consider
the death-bed promise case.)
Rawls’ suggestion: utilitarian considerations should be appealed to to justify the practice of
promising, but the rules of a practice justified on utilitarian grounds would not include a rule
permitting us to break our promises when doing so would be best on the whole, because such a
practice would be much less useful than our actual practice. So utilitarian considerations may
not be appealed to to justify particular decisions falling under the practice.

Again, compare two questions:

(i) Q: Why did J do that? (Why should J do that?)
A: He promised he would. (Backward-looking)
(ii) Q: Why do what you promised?

A: Because it leads to the best consequences. (Forward-looking)
But again, we might ask, are these really different kinds of questions, demanding
different kinds of answers?
Practice Rules v. Summary Rules
Summary Rule: a “rule of thumb” or rule of convenience; it represents a kind of “summary” of
past decisions arrived at by direct application of more basic reasoning (e.g., the utilitarian
principle)
• useful to us because similar cases tend to reoccur
• decisions made on particular cases are logically prior to the rules: “the performance of
the action to which the rule refers does not require the stage-setting of a practice of which
this rule is a part”
• in principle, we’re “always entitled to reconsider the correctness of a rule and to question
whether or not it is proper to follow it in a particular case (But: can a rule of thumb serve
a useful purpose if we can question it every time?)
• employing a summary rule is justified if and because applying it will lead us to be more
likely to make an independently correct decision then we would by direct application of
the more basic reasoning; but what the right thing to do is in this case is independent of
what the rule tells us…
• Examples? Don’t tell a fatally ill person that he’s dying; Look both ways before you
cross the street…
Practice Rule: Rule setting up offices, or defining certain actions and specifying when they are
appropriate, establishing penalties, etc.
• Rules of practice are logically prior to particular cases of actions governed by those rules
– e.g. striking out, stealing a base, making a promise: there is no way to strike out
without following the rules of baseball.
• Someone who wants to perform an action governed by a practice rule can’t meaningfully
ask whether or not he should follow the rule in this case (consider “can I have a fourth
strike?” – such a person would be most charitably interpreted as asking what the rules
were)
• Questioning whether to follow a rule, Rawls suggests, must take the form of questioning
the rule itself – that is, questioning whether the practice is designed as well as it might be
(see footnote 25, pp. 28-9, for an example)
• Defenses of particular actions falling under practices must take the form of appeals to the
rules of the practice, and then defenses of the practice as a whole. (Remember our two
questions…) Actions governed by practice rules are correct or incorrect independently of
the place-setting of the practice. (But: can’t I respond by saying, I know J robbed a
bank, and I know there are good justifications for a (general) practice of punishing bank
robbers by imprisoning them, but ought we to imprison J? After all, imprisoning J is an
action that’s not defined by the practice, even if punishment is…)
• “One can be as radical as one likes but in the case of actions specified by practices the
objects of one’s radicalism must be the social practices and people’s acceptance of them.”
(p. 32)

Question: Is it never appropriate to ask whether a particular act mandated by a practice is
justified, if we’re agreed that the practice itself is justified? (I might not be able to ask
whether, say, a fourth strike is justified, but I can ask whether my throwing the kid the
ball again is justified…)