Rawls Two Concepts of Rules and the Time Inconsistency Problem

Rawls’ Two Concepts of Rules and the Time Inconsistency Problem

In his paper, “Two Concepts of Rules” Rawls reexamines utilitarianism, and makes the important

distinction between justifying a practice and justifying a particular action that falls under that broader

practice in order to then defend utilitarianism against objections that have been made against it with

regards to punishment. The utilitarian view on punishment in terms of maximizing total welfare runs into

a contradiction between the fact that we think the state ought never to punish a person it knows to be

innocent, but sometimes maximizing total welfare may require the punishment of an innocent person. In

order to reconcile this problem, Rawls proposes that we examine rules as falling under two different

categories, summary rules and practice rules. What differentiates practice rules from summary rules is

that practice rules can be thought of as all jointly contributing or pertaining to a single goal or purpose,

which is the practice. Rawls believes that punishment can be thought of as exactly such a practice, and

argues that we can still use the utilitarian approach to analyze the practice of the institution of

punishment, but once we have done this, it is no longer appropriate to judge individual acts within the

practice outside of the context of the practice. Finally, I will expand on Rawls’ ideas by making the

utilitarian claim for the practice of punishment where only the guilty can be punished through the

framework of the time-inconsistency problem.

In the traditional utilitarian view, punishment is only justified if it leads to an increase in total

welfare within society. If the choice of punishing a person results in lower social welfare than the choice

not to punish, then the punishment would not be justified. This definition of just punishment runs into

some difficulties when the question arises of whether it can ever be just to punish an innocent person.

Our intuitive feeling is that punishing an innocent man or woman would not be just, but the traditional

utilitarian argument for why this is the case is not very strong. Bentham argues that the harm felt by the

innocent person being punished would be so great, that it would always outweigh any potential benefits.

He believes that the principle of utility would never justify punishing a person who has committed no

harm.

But this argument isn’t very satisfying, because it doesn’t create an absolute rule against

punishing innocents, but instead leaves it up an empirical calculation.

There are hypothetical examples we can create in which punishing an innocent person would

actually lead to higher levels of total welfare, the classic one being the case of punishing someone to

make an example of them in order to prevent a riot, which would cause much greater violence and

destruction than the act of punishment. The traditional utilitarian explanation does not give us an

absolute rule against punishing innocents, and it does not tell us how to deal with a situation in which

maximizing total welfare might require the state to punish a person it knows to be innocent. Rawls uses

the distinction between practices and actions within those practices to argue on behalf of utilitarianism

against these types of criticisms. Rawls makes the argument that utilitarianism can in fact be consistent

with the position that the state should never punish a person it knows to be innocent by appealing to the

separation between the practice of punishment and the act of punishment.

Bentham, 675

In doing so, he begins by defining the difference between practice rules and summary rules.

Summary rules can considered as “rules of thumb” or rules of convenience; they represent a type of

summary of past decisions arrived at the direct application of basic reasoning. An example of a summary

rule would be the idea that we should look both ways before crossing the street or that we should knock

before entering a room. In the case of summary rules, we can always reconsider the correctness of

applicability of the rule in particular cases. We adopt summary rules because, on the whole, following

the rule leads us to be better off than we would be if we were to rely on basic reasoning every time. This

does not preclude us from using our own reasoning prior to acting, and deciding that, in this particular

case, it might be better to not apply the rule.

Rawls also defines a second type of rules, and these are practice rules. Practice rules are logically

prior to the particular actions governed by the rules. Practice rules set up offices, define specific actions,

and specify and establish appropriateness and consequences of actions. To explain this relationship

between practice rules and actions, Rawls uses the metaphor of baseball. Baseball involves certain

actions: throwing the ball, swinging the bat, or running to the base, but in order for to say that we are

playing the game of baseball, the rules of baseball must exist prior to us engaging in these actions. We

can imagine that before baseball was invented, people could still perform the actions of throwing a ball

and hitting it with a stick, but they wouldn’t be playing baseball until the rules were created which set up

and defined the relationships in the game.

In addition, particular actions within the practice can only

make sense in the context of that practice. For example, in baseball, it is impossible to strike out or steal

a base outside of the context of following the rules of baseball. These actions are specific to the practice,

and do not exist outside of the game of baseball.

Furthermore, because the rules are logically prior to the actions, unlike in the case of summary

rules, it is impossible logically ask whether or not the rule should apply in a particular case or instance.

In baseball, the practice rules dictate that every batter get three strikes. If someone were to ask for a

fourth strike, the other players would say: no, we are playing baseball; the rules of baseball only allow

three strikes. It would not make sense for the first person to then say: yes I agree that we are playing

baseball, and that there should be 3 strikes, but I think that in this case, I should get a fourth. Once we

have established and accepted the rules of a practice, it is irrelevant to talk about the merits of an action

or rule outside of the context of its role within the practice, since once we no longer follow the rule, we

are no longer participating in the practice. If we begin giving some batters three strikes, and some batters

four, we are no longer playing baseball. We cannot simultaneously participate in the practice of baseball

and also give batters different numbers of strikes.

In some cases, however, we may feel that it would be acceptable to make individual or case

specific exceptions. What if you are playing baseball with young children and one of them hasn’t been

able to get a hit the whole game. You know she will be very upset if she strikes out again, so you make an

exception and allow her to take a couple more tries. You know that in baseball, the rules say that there

are only three strikes allowed, but in this case your ultimate goal is that all the kids have fun rather than

Rawls, 25

Rawls, 26

to adhere strictly to all the rules of baseball, so it would actually be better not to follow the rule.

However, given this difference in the ultimate goal here in this example, a supporter of Rawls could

respond that the practice of baseball could actually be divided into two sub-practices. There is official

baseball, where the rules must be followed strictly in order to ensure a fair game and winner, and there

is also recreational baseball, where the main objective is to have fun; and in recreational baseball, the

rules may actually allow for a change in the number of strikes for the sake of making sure that everyone

is enjoying themselves. If prior to beginning the game, it is clearly established that recreational baseball

is being played as opposed to official baseball, the decision to allow the child to take an extra strike

would no longer be problematic.

The above claim could imply that in any case where you think there may be a legitimate

exception to a practice rule, it shouldn’t be considered as breaking the rule, but instead as creating a

sub-practice where the rules are actually slightly different. But isn’t this problematic for cheating,

someone might ask. Every time you want to break the rules, you could simply say that you are not

actually participating in the original practice, but in a new sub-practice where your actions are not

actually against the rules and that you have no obligation to the rules of the original practice. Here it is

important to remember the definition of practice rules: practice rules are set up prior to actions.

So in order for a cheater to make an appeal to a sub-practice, the sub-practice must already exist

and have rules that correspond exactly to the way that the cheater is behaving. If this is indeed that case,

then cheating can be prevented by clearly establishing consent for the practice in which everyone is

participating before anyone has committed to any particular actions. If, before the game, the players

come to a consensus that they are playing official baseball, thereby giving their implied consent to be

bound by the rules of official baseball, then any attempt to sneak in an additional strike cannot be

attributed to the belief of the cheater that they were actually playing recreational baseball, and even if it

was truly a misunderstanding, it still would not be an exculpatory defense for breaking the rules.

Knowingly participating in a particular practice implies that you have consented to following the rules of

that practice. This is akin to establishing jurisdictions for various laws if we were compare the different

sub-practices to different geographical areas with different laws. Maybe jaywalking is illegal in New York,

but not in New Jersey. However, once you enter New York, you have implicitly consented to follow the

New York state laws, and arguing a jay walking fine by saying that jaywalking is perfectly legal in New

Jersey will not be an acceptable defense.

So, if we can’t question whether to perform a particular action in the case of practice rules, what

is the correct way to exercise our judgment about whether we should be following a rule? Rawls argues

that the questioning whether to follow a practice rule must take the form of questioning the rule itself,

not its implications in a particular circumstance. In order to question a practice rule, you must question

whether the practice itself is effectively designed.

Going back to the baseball example, if you wanted to

challenge the practice rule about strikes, you would have to challenge the entire practice of allowing only

3 strikes. You can argue that the rules of official baseball should be changed to allow four strikes, but you

cannot argue that the rules of baseball say there are only three strikes, but in this particular

Rawls, 16

circumstance, you think you should be allowed a fourth. If you want to defend or appeal a particular

action within a practice, the argument must take the form of appeals or defenses of the entire practice

as a whole. It does not make sense to argue about the correctness of an action outside of the context of

its role or function within a practice.

By now I’m sure we are all wondering how this lengthy discussion of practices and rules relates

back to the original issue we wanted to address about how it can be possible to make a utilitarian

argument for never punishing a person that is known to be innocent. Well, Rawls believes that

punishment can be thought of as a practice, and all the rules regarding the appropriateness, nature, and

logistics of punishment can be considered practice rules. In order to create a practice or institution of

punishment, we must clearly define the common goal of punishment which all of its practice rules

contribute to. This common goal, or definition, of punishment is: to legally deprive someone of their

typical rights in the way permitted or described by the law due to that fact that he or she has violated

the law when that violation has been determined to have occurred by a fair and legal trial.

This implies

that punishment is defined by the guilt of the person who is being punished, and that the practice rules

of punishment require that the person who is being punished by the state has been found guilty of a

crime. This means that the practice laws of the institution of punishment require proof of guilt in order

to punish someone.

When we consider punishment as an institution, and accept its existence as practice in the form

of the definition above, it is no longer relevant to question whether it is appropriate to punish an

innocent person in a particular instance. In the same way that the rules of baseball require three strikes

and it is illogical to argue for four strikes under particular circumstances, the rules of punishment require

that the person being punished be guilty of breaking the law, and it is irrelevant to argue that the person

can be innocent under particular circumstances. Punishment is defined as a consequence of lawbreaking,

so once you punish an innocent person, you are no longer punishing them; you are participating in a

completely different practice, which Rawls names “telishment”

The creation of practices and the enforcement of practice rules can be superimposed relatively

closely onto the roles of legislature and judiciary. The legislature has the job of determining whether a

certain practice is worth integrating into law, but once a practice has been accepted into law, the

judiciary cannot make a decision regarding the correctness of an individual action outside of the context

of the practice. We may usefully compare the roles of judge and legislature can be compared to the roles

of the creators of baseball and to the umpire.

Although there is indeed a major difference between the government and the baseball team, I

think they can be reconciled by examining the judiciary-legislature relationship in more detail. Unlike in

case of umpire, the checks and balances system of our government does allow for some discretion and

creation of new laws on the part of judges. However, in instances where judges can use discretion, this

ability to use discretion is built into the practice.

For example, the rules of punishment often say something like: if the defendant is guilty of X, the

punishment is a fine of between 100 and 500 dollars. In this way, judicial discretion is directly built into

the practice of punishment. The judge can say: yes, this person committed X, but she was simply very

careless and did not commit the wrongdoing with scienter, so I will give her a only the $100 fine. But, he

only has the power to use that discretion to the extent that is allowed by the practice. He cannot say that

he thinks that we should not punish the defendant for her actions at all, nor can he say that he thinks the

defendant acted so badly that she should pay 600 dollars in fines since neither of these actions is within

the range of discretion permitted by the practice.

The second way in which discretion is built into punishment is instances where the law is silent.

In these cases the practice laws do not address whether punishment is acceptable in a particular

circumstance or not, so it is unavoidable that the judge make a discretionary decision. If the

judge/umpire chooses not to enforce, then it sets one type of example, and if she does choose to punish,

it sets the opposite example. Given the efficiency constraints, it is not practical for the judge or umpire to

appeal to the legislature or creators of baseball for a decision. The practice dictates that these individuals

are trained in the intentions and purposes of the rule creators so that they can interpret the rules in the

best unbiased and scientific way. Just as rules are not rewritten by umpires, judges do not rewrite rules,

they only interpret them.

The final way that the judiciary can exercise discretion is through an appeal to the Supreme

Court. In this case, I think it is helpful to keep in mind the idea of practices and sub-practices explained

earlier in the paper. The ultimate practice to which the courts are beholden to is upholding the

constitution. In this way, the courts may appeal the correctness of a certain practice by saying that it

violates the goal of this ultimate practice. The practice of upholding the constitution may require

additional sub-practices with their own rules, but ultimately even they are answerable to this ultimate

practice. We have the practice of baseball, which so far includes the sub-practices of official baseball and

recreational baseball. Someone may try to create a new sub-practice of baseball called British Baseball in

which the players use a flat bat and run back and forth instead of around the bases. It can then be

decided that this new sub-practice is sufficiently different from the original category of baseball that it

can no longer be called baseball, but falls into a different practice called cricket. In the same way, it can

be decided that a sub-practice within the law is sufficiently different from the objectives of our legal

system that it no longer falls within the practice agreed to in the constitution.

Alright, so we have an idea of practices, and we have established two important conclusions:

first, that a practice which allows for punishing innocent as well as the guilty is a different from a practice

in which punishment is only applied to the guilty, and second, that we cannot question the righteousness

of actions within practices without questioning the practice itself. But, the question we really want to

answer is: why do we find that the practice which includes a rule allowing us to punish the innocent is

unjust while the practice which only allows us to punish the guilty is just? Well, in this case, utilitarianism

can be applied relatively straightforwardly in order to answer this question.

Utilitarian considerations would not allow for a practice which allows us to punish the innocent

since such a practice would undermine the entire deterrent benefit of punishment. Decisions about

punishment have an inherent time-inconsistency problem, which is characterized by circumstances

under which the best course of action at a particular time may not be the best course of action over all.

In the case of punishment, the best course of action over all is an institution of punishment

where those that are guilty of breaking the law face certain consequences for their disobedience, so that

people are deterred from breaking the law in the future. However, in every particular instance, there is

an incentive for the state to punish an innocent person. Punishing an innocent person would be less

costly than finding the guilty party and going through the trouble of proving their guilt, but would serve

the same purpose of setting an example for those considering breaking the law. However, because

people are forward looking, once they can anticipate that the state is no longer adhering to this promise

to punish only the guilty, they no longer have any incentive to follow the law since their likelihood of

punishment will be same regardless of whether they are guilty or innocent.

The utility gain of punishing an innocent can only be realized if the people still believe that

punishment is only for the guilty. So once you have an established practice of punishing the innocent, the

arbitrage opportunity created by the discrepancy between the state’s actions and the public’s beliefs

about those actions no longer exists, and the system of punishment will collapse.

The classic example of this type of problem is when you have a professor with a class full of

students. The professor really wants everyone in the class to study and learn the material, but he doesn't

necessary want to give students exams since exams take away class time and he doesn't want to grade

them. So the professor tells the class that at the end of the semester there will be a lengthy, challenging,

cumulative final exam. However, once the end of the semester approaches, the professor realizes that he

really doesn't want to spend the first few weeks of June grading all of these exams; so the logical thing

for him to do would be to cancel the exam. The system was based on the promise that there would be

this tough exam at the end of the semester, and as long as the students believe this promise and study

for the exam, then the professor does not need to keep his promise. But, just like people evaluating the

consequences of criminal actions, the students are smart and forward looking, they can anticipate that

it's not really in the professor's interest to have to proctor and grade all of the exams. And once they no

longer believe that the professor will keep his promise, then they no longer have an incentive to study,

and the system ceases to function.

The final exam example contains the same problem as the punishment example since the

summation of individual utility maximizing decisions over a course of time does not actually produce the

overall utility maximizing decision over the course of multiple years or repetitions. In each individual

instance, it will be in the interest of the state to punish the innocent man as opposed to the guilty man,

but over time, this practice will produce a state where people no longer deterred by punishment and no

longer trust their government. In order to preserve a practice which will ultimately be the most welfare

maximizing action in the long run, we may have to consistently make a series of individual decisions in

which the highest welfare option for each particular instance is not chosen.

Strotz, 165

I think Rawls’ use of a distinction between practices and individual actions characterized by those

practices was very enlightening for the understanding of the justification for never punishing the

innocent. Separating the practice of punishment from the individual actions which make up the process

of punishing someone allows us to contain utilitarian considerations to only the practice in general,

rather than applying them to individual cases. Furthermore, I think the theory of the time-inconsistency

problem provides additional insight into how this distinction can be understood and supported from the

utilitarian perspective. Utilitarianism calls for welfare maximization, but it does not often specify a time

frame. In cases where certain decisions produce different welfare outcomes in the short and long run,

practices can be created precisely to preserve certain long run utilitarian benefits in the presence of

short run incentives to defect.