John Rawls, "Two Concepts of Rules," The Philosophical Review 64 (1955): 3-32.

In this paper I want to show the importance of the distinction between justifying a practice1 and justifying a particular action falling under it, and I want to explain the logical basis of this distinction and how it is possible to miss its significance. While the distinction has frequently been made,2 and is now becoming commonplace, there remains the task of explaining the tendency either to overlook it altogether, or to fail to appreciate its importance.

To show the importance of the distinction I am going to defend utilitarianism against those objections which have traditionally been made against it in connection with punishment and the obligation to keep promises. I hope to show that if one uses the distinction in question then one can state utilitarianism in a way which makes it a much better explication of our considered moral judgments than these traditional objections would seem to admit.3 Thus the importance of the distinction is shown by the way it strengthens the utilitarian view regardless of whether that view is completely defensible or not.

To explain how the significance of the distinction may be overlooked, I am going to discuss two conceptions of rules. One of these conceptions conceals the importance of distinguishing between the justification of a rule or practice and the justification of a particular action falling under it. The other conception makes it clear why this distinction must be made and what is its logical basis.

I

The subject of punishment, in the sense of attaching legal penalties to the violation of legal rules, has always been a troubling moral question.4 The trouble about it has not been that people disagree as to whether or not punishment is justifiable. Most people have held that, freed from certain abuses, it is an acceptable institution. Only a few have rejected punishment entirely, which is rather surprising when one considers all that can be said against it. The difficulty is with the justification of punishment: various arguments for it have been given by moral philosophers, but so far none of them has won any sort of general acceptance; no justification is without those who detest it. I hope to show that the use of the aforementioned distinction enables one to state the utilitarian view in a way which allows for the sound points of its critics.

For our purposes we may say that there are two justifications of punishment. What we may call the retributive view is that punishment is justified on the grounds that wrongdoing merits punishment. It is morally fitting that a person who does wrong should suffer in proportion to his wrongdoing. That a criminal should be punished follows from his guilt, and the severity of the appropriate punishment depends on the depravity of his act. The state of affairs where a wrongdoer suffers punishment is morally better than the state of affairs where he does not, and it is better irrespective of any of the consequences of punishing him.

What we may call the utilitarian view holds that on the principle that bygones are bygones and that only future consequences are material to present decisions, punishment is justifiable only by reference to the probable consequences of maintaining it as one of the devices of the social order. Wrongs committed in the past are, as such, not relevant considerations for deciding what to do. If punishment can be shown to promote effectively the interest of society it is justifiable, otherwise it is not.

I have stated these two competing views very roughly to make one feel the conflict between them: one feels the force of both arguments and one wonders how they can be reconciled. From my introductory remarks it is obvious that the resolution which I am going to propose is that in this case one must distinguish between justifying a practice as a system of rules to be applied and enforced, and justifying a particular action which falls under these rules; utilitarian arguments are appropriate with regard to questions about practices, while retributive arguments fit the application of particular rules to particular cases.

We might try to get clear about this distinction by imagining how a father might answer the question of his son. Suppose the son asks, "Why was J put in jail yesterday?" The father answers, "Because he robbed the bank at B. He was duly tried and found guilty. That's why he was put in jail yesterday." But suppose the son had asked a different question, namely, "Why do people put other people in jail?" Then the father might answer, "To protect good people from bad people" or "To stop people from doing things that would make it uneasy for all of us; for otherwise we wouldn't be able to go to bed at night and sleep in peace." There are two very different questions here. One question emphasizes the proper name: It asks why J was punished rather than someone else, or it asks what he was punished for. The other question asks why we have the institution of punishment: Why do people punish one another rather than, say, always forgiving one another?

Thus the father says in effect that a particular man is punished, rather than some other man, because he is guilty, and he is guilty because he broke the law (past tense). In his case the law looks back, the judge looks back, the jury looks back, and a penalty is visited upon him for something he did. That a man is to be punished, and what his punishment is to be, is settled by its being shown that he broke the law and that the law assigns that penalty for the violation of it.

On the other hand we have the institution of punishment itself, and recommend and accept various changes in it, because it is thought by the (ideal) legislator and by those to whom the law applies that, as a part of a system of law impartially applied from case to case arising under it, it will have the consequence, in the long run, of furthering the interests of society.

One can say, then, that the judge and the legislator stand in different positions and look in different directions: one to the past, the other to the future. The justification of what the judge does, qua judge, sounds like the retributive view; the justification of what the (ideal) legislator does, qua legislator, sounds like the utilitarian view. Thus both views have a point (this is as it should be since intelligent and sensitive persons have been on both sides of the argument); and one's initial confusion disappears once one sees that these views apply to persons holding different offices with different duties, and situated differently with respect to the system of rules that make up the criminal law.5

One might say, however, that the utilitarian view is more fundamental since it applies to a more fundamental office, for the judge carries out the legislator's will so far as he can determine it. Once the legislator decides to have laws and to assign penalties for their violation (as things are there must be both the law and the penalty) an institution is set up which involves a retributive conception of particular cases. It is part of the concept of the criminal law as a system of rules that the application and enforcement of these rules in particular cases should be justifiable by arguments of a retributive character. The decision whether or not to use law rather than some other mechanism of social control, and the decision as to what laws to have and what penalties to assign, may be settled by utilitarian arguments; but if one decides to have laws then one has decided on something whose working in particular cases is retributive in form.6

The answer, then, to the confusion engendered by the two views of punishment is quite simple: One distinguishes two offices, that of the judge and that of the legislator, and one distinguishes their different stations with respect to the system of rules which make up the law; and then one notes that the different sorts of considerations which would usually be offered as reasons for what is done under the cover of these offices can be paired off with the competing justifications of punishment. One reconciles the two views by the time-honored device of making them apply to different situations.

But can it really be this simple? Well, this answer allows for the apparent intent of each side. Does a person who advocates the retributive view necessarily advocate, as an institution, legal machinery whose essential purpose is to set up and preserve a correspondence between moral turpitude and suffering? Surely not.7 What retributionists have rightly insisted upon is that no man can be punished unless he is guilty, that is, unless he has broken the law. Their fundamental criticism of the utilitarian account is that, as they interpret it, it sanctions an innocent person's being punished (if one may call it that) for the benefit of society.

On the other hand, utilitarians agree that punishment is to be inflicted only for the violation of law. They regard this much as understood from the concept of punishment itself.8 The point of the utilitarian account concerns the institution as a system of rules: utilitarianism seeks to limit its use by declaring it justifiable only if it can be shown to foster effectively the good of society. Historically it is a protest against the indiscriminate and ineffective use of the criminal law.9 It seeks to dissuade us from assigning to penal institutions the improper, if not sacrilegious, task of matching suffering with moral turpitude. Like others, utilitarians want penal institutions designed so that, as far as humanly possible, only those who break the law run afoul of it. They hold that no official should have discretionary power to inflict penalties whenever he thinks it for the benefit of society; for on utilitarian grounds an institution granting such power could not be justified.10

The suggested way of reconciling the retributive and the utilitarian justifications of punishment seems to account for what both sides have wanted to say. There are, however, two further questions which arise, and I shall devote the remainder of this section to them.

First, will not a difference of opinion as to the proper criterion of just law make the proposed reconciliation unacceptable to retributionists? Will they not question whether, if the utilitarian principle is used as the criterion, it follows that those who have broken the law are guilty in a way which satisfies their demand that those punished deserve to be punished? To answer this difficulty, suppose that the rules of the criminal law are justified on utilitarian grounds (it is only for laws that meet his criterion that the utilitarian can be held responsible). Then it follows that the actions which the criminal law specifies as offenses are such that, if they were tolerated, terror and alarm would spread in society. Consequently, retributionists can only deny that those who are punished deserve to be punished if they deny that such actions are wrong. This they will not want to do.

The second question is whether utilitarianism doesn't justify too much. One pictures it as an engine of justification which, if consistently adopted, could be used to justify cruel and arbitrary institutions. Retributionists may be supposed to concede that utilitarians intend to reform the law and to make it more humane; that utilitarians do not wish to justify any such thing as punishment of the innocent; and that utilitarians may appeal to the fact that punishment presupposes guilt in the sense that by punishment one understands an institution attaching penalties to the infraction of legal rules, and therefore that it is logically absurd to suppose that utilitarians in justifying punishment might also have justified punishment (if we may call it that) of the innocent. The real question, however, is whether the utilitarian, in justifying punishment, hasn't used arguments which commit him to accepting the infliction of suffering on innocent persons if it is for the good of society (whether or not one calls this punishment). More generally, isn't the utilitarian committed in principle to accepting many practices which he, as a morally sensitive person, wouldn't want to accept? Retributionists are inclined to hold that there is no way to stop the utilitarian principle from justifying too much except by adding to it a principle which distributes certain rights to individuals. Then the amended criterion is not the greatest benefit of society simpliciter, but the greatest benefit of society subject to the constraint that no one's rights may be violated. Now while I think that the classical utilitarians proposed a criterion of this more complicated sort, I do not want to argue that point here.11 What I want to show is that there is another way of preventing the utilitarian principle from justifying too much, or at least of making it much less likely to do so: namely, by stating utilitarianism in a way which accounts for the distinction between the justification of an institution and the justification of a particular action failing under it.

I begin by defining the institution of punishment as follows: a person is said to suffer punishment whenever he is legally deprived of some of the normal rights of a citizen on the ground that he has violated a rule of law, the violation having been established by trial according to the due process of law, provided that the deprivation is carried out by the recognized legal authorities of the state, that the rule of law clearly specifies both the offense and the attached penalty, that the courts construe statutes strictly, and that the statute was on the books prior to the time of the offense.12 This definition specifies what I shall understand by punishment. The question is whether utilitarian arguments may be found to justify institutions widely different from this and such as one would find cruel and arbitrary.

This question is best answered, I think, by taking up a particular accusation. Consider the following from Carritt:

. . . the utilitarian must hold that we are justified in inflicting pain always and only to prevent worse pain or bring about greater happiness. This, then, is all we need to consider in so-called punishment, which must be purely preventive. But if some kind of very cruel crime becomes common, and none of the criminals can be caught, it might be highly expedient, as an example, to hang an innocent man, if a charge against him could be so framed that he were universally thought guilty; indeed this would only fail to be an ideal instance of utilitarian 'punishment' because the victim himself would not have been so likely as a real felon to commit such a crime in the future; in all other respects it would be perfectly deterrent and therefore felicific.13
Carritt is trying to show that there are occasions when a utilitarian argument would justify taking an action which would be generally condemned; and thus that utilitarianism justifies too much. But the failure of Carritt's argument lies in the fact that he makes no distinction between the justification of the general system of rules which constitutes penal institutions and the justification of particular applications of these rules to particular cases by the various officials whose job it is to administer them. This becomes perfectly clear when one asks who the "we" are of whom Carritt speaks. Who is this who has a sort of absolute authority on particular occasions to decide that an innocent man shall be "punished" if everyone can be convinced that he is guilty? Is this person the legislator, or the judge, or the body of private citizens, or what? It is utterly crucial to know who is to decide such matters, and by what authority, for all of this must be written into the rules of the institution. Until one knows these things one doesn't know what the institution is whose justification is being challenged; and as the utilitarian principle applies to the institution one doesn't know whether it is justifiable on utilitarian grounds or not.

Once this is understood it is clear what the countermove to Carritt's argument is. One must describe more carefully what the institution is which his example suggests, and then ask oneself whether or not it is likely that having this institution would be for the benefit of society in the long run. One must not content oneself with the vague thought that, when it's a question of this case, it would be a good thing if somebody did something even if an innocent person were to suffer.

Try to imagine, then, an institution (which we may call "telishment") which is such that the officials set up by it have authority to arrange a trial for the condemnation of an innocent man whenever they are of the opinion that doing so would be in the best interests of society. The discretion of officials is limited, however, by the rule that they may not condemn an innocent man to undergo such an ordeal unless there is, at the time, a wave of offenses similar to that with which they charge him and telish him for. We may imagine that the officials having the discretionary authority are the judges of the higher courts in consultation with the chief of police, the minister of justice, and a committee of the legislature.

Once one realizes that one is involved in setting up an institution, one sees that the hazards are very great. For example, what check is there on the officials? How is one to tell whether or not their actions are authorized? How is one to limit the risks involved in allowing such systematic deception? How is one to avoid giving anything short of complete discretion to the authorities to telish anyone they like? In addition to these considerations, it is obvious that people will come to have a very different attitude towards their penal system when telishment is adjoined to it. They will be uncertain as to whether a convicted man has been punished or telished. They will wonder whether or not they should feel sorry for him. They will wonder whether the same fate won't at any time fall on them. If one pictures how such an institution would actually work, and the enormous risks involved in it, it seems clear that it would serve no useful purpose. A utilitarian justification for this institution is most unlikely.

It happens in general that as one drops off the defining features of punishment one ends up with an institution whose utilitarian justification is highly doubtful. One reason for this is that punishment works like a kind of price system: by altering the prices one has to pay for the performance of actions, it supplies a motive for avoiding some actions and doing others. The defining features are essential if punishment is to work in this way; so that an institution which lacks these features, for example, an institution which is set up to "punish" the innocent, is likely to have about as much point as a price system (if one may call it that) where the prices of things change at random from day to day and one learns the price of something after one has agreed to buy it.14

If one is careful to apply the utilitarian principle to the institution which is to authorize particular actions, then there is less danger of its justifying too much. Carritt's example gains plausibility by its indefiniteness and by its concentration on the particular case. His argument will only hold if it can be shown that there are utilitarian arguments which justify an institution whose publicly ascertainable offices and powers are such as to permit officials to exercise that kind of discretion in particular cases. But the requirement of having to build the arbitrary features of the particular decision into the institutional practice makes the justification much less likely to go through.

II

I shall now consider the question of promises. The objection to utilitarianism in connection with promises seems to be this: it is believed that on the utilitarian view when a person makes a promise the only ground upon which he should keep it, if he should keep it, is that by keeping it he will realize the most good on the whole. So that if one asks the question "Why should I keep my promise?" the utilitarian answer is understood to be that doing so in this case will have the best consequences. And this answer is said, quite rightly, to conflict with the way in which the obligation to keep promises is regarded.

Now of course critics of utiliarianism are not unaware that one defense sometimes attributed to utilitarians is the consideration involving the practice of promise-keeping.15 In this connection they are supposed to argue something like this: it must be admitted that we feel strictly about keeping promises, more strictly than it might seem our view can account for. But when we consider the matter carefully it is always necessary to take into account the effect which our action will have on the practice of making promises. The promisor must weigh, not only the effects of breaking his promise on the particular case, but also the effect which his breaking his promise will have on the practice itself. Since the practice is of great utilitarian value, and since breaking one's promise always seriously damages it, one will seldom be justified in breaking one's promise. If we view our individual promises in the wider context of the practice of promising itself we can account for the strictness of the obligation to keep promises. There is always one very strong utilitarian consideration in favor of keeping them, and this will insure that when the question arises as to whether or not to keep a promise it will usually turn out that one should, even where the facts of the particular case taken by itself would seem to justify one's breaking it. In this way the strictness with which we view the obligation to keep promises is accounted for.

Ross has criticized this defense as follows:16 however great the value of the practice of promising, on utilitarian grounds, there must be some value which is greater, and one can imagine it to be obtainable by breaking a promise. Therefore there might be a case where the promisor could argue that breaking his promise was justified as leading to a better state of affairs on the whole. And the promisor could argue in this way no matter how slight the advantage won by breaking the promise. If one were to challenge the promisor his defense would be that what he did was best on the whole in view of all the utilitarian considerations, which in this case include the importance of the practice. Ross feels that such a defense would be unacceptable. I think he is right insofar as he is protesting against the appeal to consequences in general and without further explanation. Yet it is extremely difficult to weigh the force of Ross's argument. The kind of case imagined seems unrealistic and one feels that it needs to be described. One is inclined to think that it would either turn out that such a case came under an exception defined by the practice itself, in which case there would not be an appeal to consequences in general on the particular case, or it would happen that the circumstances were so peculiar that the conditions which the practice presupposes no longer obtained. But certainly Ross is right in thinking that it strikes us as wrong for a person to defend breaking a promise by a general appeal to consequences. For a general utilitarian defense is not open to the promisor: it is not one of the defenses allowed by the practice of making promises.

Ross gives two further counterarguments:17 First, he holds that it overestimates the damage done to the practice of promising by a failure to keep a promise. One who breaks a promise harms his own name certainly, but it isn't clear that a broken promise always damages the practice itself sufficiently to account for the strictness of the obligation. Second, and more important, I think, he raises the question of what one is to say of a promise which isn't known to have been made except to the promisor and the promisee, as in the case of a promise a son makes to his dying father concerning the handling of the estate.18 In this sort of case the consideration relating to the practice doesn't weigh on the promisor at all, and yet one feels that this sort of promise is as binding as other promises. The question of the effect which breaking it has on the practice seems irrelevant. The only consequence seems to be that one can break the promise without running any risk of being censured; but the obligation itself seems not the least weakened. Hence it is doubtful whether the effect on the practice ever weighs in the particular case; certainly it cannot account for the strictness of the obligation where it fails to obtain. It seems to follow that a utilitarian account of the obligation to keep promises cannot be successfully carried out.

From what I have said in connection with punishment, one can foresee what I am going to say about these arguments and counterarguments. They fail to make the distinction between the justification of a practice and the justification of a particular action falling under it, and therefore they fall into the mistake of taking it for granted that the promisor, like Carritt's official, is entitled without restriction to bring utilitarian considerations to bear in deciding whether to keep his promise. But if one considers what the practice of promising is one will see, I think, that it is such as not to allow this sort of general discretion to the promisor. Indeed, the point of the practice is to abdicate one's title to act in accordance with utilitarian and prudential considerations in order that the future may be tied down and plans coordinated in advance. There are obvious utilitarian advantages in having a practice which denies to the promisor, as a defense, any general appeal to the utilitarian principle in accordance with which the practice itself may be justified. There is nothing contradictory, or surprising, in this: utilitarian (or aesthetic) reasons might properly be given in arguing that the game of chess, or baseball, is satisfactory just as it is, or in arguing that it should be changed in various respects, but a player in a game cannot properly appeal to such considerations as reasons for his making one move rather than another. It is a mistake to think that if the practice is justified on utilitarian grounds then the promisor must have complete liberty to use utilitarian arguments to decide whether or not to keep his promise. The practice forbids this general defense; and it is a purpose of the practice to do this. Therefore what the above arguments presuppose -- the idea that if the utilitarian view is accepted then the promisor is bound if, and only if, the application of the utilitarian principle to his own case shows that keeping it is best on the whole -- is false. The promisor is bound because he promised: weighing the case on its merits is not open to him.19

Is this to say that in particular cases one cannot deliberate whether or not to keep one's promise? Of course not. But to do so is to deliberate whether the various excuses, exceptions and defenses, which are understood by, and which constitute an important part of, the practice, apply to one's own case.20 Various defenses for not keeping one's promise are allowed, but among them there isn't the one that, on general utilitarian grounds, the promisor (truly) thought his action best on the whole, even though there may be the defense that the consequences of keeping one's promise would have been extremely severe. While there are too many complexities here to consider all the necessary details, one can see that the general defense isn't allowed if one asks the following question: what would one say of someone who, when asked why he broke his promise, replied simply that breaking it was best on the whole? Assuming that his reply is sincere, and that his belief was reasonable (i.e., one need not consider the possibility that he was mistaken), I think that one would question whether or not he knows what it means to say "I promise" (in the appropriate circumstances). It would be said of someone who used this excuse without further explanation that he didn't understand what defenses the practice, which defines a promise, allows to him. If a child were to use this excuse one would correct him; for it is part of the way one is taught the concept of a promise to be corrected if one uses this excuse. The point of having the practice would be lost if the practice did allow this excuse.

It is no doubt part of the utilitarian view that every practice should admit the defense that the consequences of abiding by it would have been extremely severe; and utilitarians would be inclined to hold that some reliance on people's good sense and some concession to hard cases is necessary. They would hold that a practice is justified by serving the interests of those who take part in it; and as with any set of rules there is understood a background of circumstances under which it is expected to be applied and which need not -- indeed which cannot -- be fully stated. Should these circumstances change, then even if there is no rule which provides for the case, it may still be in accordance with the practice that one be released from one's obligation. But this sort of defense allowed by a practice must not be confused with the general option to weigh each particular case on utilitarian grounds which critics of utilitarianism have thought it necessarily to involve.

The concern which utilitarianism raises by its justification of punishment is that it may justify too much. The question in connection with promises is different: it is how utilitarianism can account for the obligation to keep promises at all. One feels that the recognized obligation to keep one's promise and utilitarianism are incompatible. And to be sure, they are incompatible if one interprets the utilitarian view as necessarily holding that each person has complete liberty to weigh every particular action on general utilitarian grounds. But must one interpret utilitarianism in this way? I hope to show that, in the sorts of cases I have discussed, one cannot interpret it in this way.

III

So far I have tried to show the importance of the distinction between the justification of a practice and the justification of a particular action falling under it by indicating how this distinction might be used to defend utilitarianism against two long-standing objections. One might be tempted to close the discussion at this point by saying that utilitarian considerations should be understood as applying to practices in the first instance and not to particular actions falling under them except insofar as the practices admit of it. One might say that in this modified form it is a better account of our considered moral opinions and let it go at that. But to stop here would be to neglect the interesting question as to how one can fail to appreciate the significance of this rather obvious distinction and can take it for granted that utilitarianism has the consequence that particular cases may always be decided on general utilitarian grounds.21 I want to argue that this mistake may be connected with misconceiving the logical status of the rules of practices; and to show this I am going to examine two conceptions of rules, two ways of placing them within the utilitarian theory.

The conception which conceals from us the significance of the distinction I am going to call the summary view. It regards rules in the following way: one supposes that each person decides what he shall do in particular cases by applying the utilitarian principle; one supposes further that different people will decide the same particular case in the same way and that there will be recurrences of cases similar to those previously decided. Thus it will happen that in cases of certain kinds the same decision will be made either by the same person at different times or by different persons at the same time. If a case occurs frequently enough one supposes that a rule is formulated to cover that sort of case. I have called this conception the summary view because rules are pictured as summaries of past decisions arrived at by the direct application of the utilitarian principle to particular cases. Rules are regarded as reports that cases of a certain sort have been found on other grounds to be properly decided in a certain way (although, of course, they do not say this).

There are several things to notice about this way of placing rules within the utilitarian theory.22

1. The point of having rules derives from the fact that similar cases tend to recur and that one can decide cases more quickly if one records past decisions in the form of rules. If similar cases didn't recur, one would be required to apply the utilitarian principle directly, case by case, and rules reporting past decisions would be of no use.

2. The decisions made on particular cases are logically prior to rules. Since rules gain their point from the need to apply the utilitarian principle to many cases, it follows, that a particular case (or several cases similar to it) may exist whether or not there is a rule covering that case. We are pictured as recognizing particular cases prior to there being a rule which covers them, for it is only if we meet with a number of cases of a certain sort that we formulate a rule. Thus we are able to describe a particular case as a particular case of the requisite sort whether there is a rule regarding that sort of case or not. Put another way: what the A's and the B's refer to in rules of the form 'Whenever A do B' may be described as A's and B's whether or not there is the rule 'Whenever A do B', or whether or not there is any body of rules which make up a practice of which that rule is a part.

To illustrate this consider a rule, or maxim, which could arise in this way: suppose that a person is trying to decide whether to tell someone who is fatally ill what his illness is when he has been asked to do so. Suppose the person to reflect and then decide, on utilitarian grounds, that he should not answer truthfully; and suppose that on the basis of this and other like occasions he formulates a rule to the effect that when asked by someone fatally ill what his illness is, one should not tell him. The point to notice is that someone's being fatally ill and asking what his illness is, and someone's telling him, are things that can be described as such whether or not there is this rule. The performance of the action to which the rule refers doesn't require the stage-setting of a practice of which this rule is a part. This is what is meant by saying that on the summary view particular cases are logically prior to rules.

3. Each person is in principle 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. As rules are guides and aids, one may ask whether in past decisions there might not have been a mistake in applying the utilitarian principle to get the rule in question, and wonder whether or not it is best in this case. The reason for rules is that people are not able to apply the utilitarian principle effortlessly and flawlessly; there is need to save time and to post a guide. On this view a society of rational utilitarians would be a society without rules in which each person applied the utilitarian principle directly and smoothly, and without error, case by case. On the other hand, ours is a society in which rules are formulated to serve as aids in reaching these ideally rational decisions on particular cases, guides which have been built up and tested by the experience of generations. If one applies this view to rules, one is interpreting them as maxims, as "rules of thumb"; and it is doubtful that anything to which the summary conception did apply would be called a rule. Arguing as if one regarded rules in this way is a mistake one makes while doing philosophy.

4. The concept of a general rule takes the following form. One is pictured as estimating on what percentage of the cases likely to arise a given rule may be relied upon to express the correct decision, that is, the decision that would be arrived at if one were to correctly apply the utilitarian principle case by case. If one estimates that by and large the rule will give the correct decision, or if one estimates that the likelihood of making a mistake by applying the utilitarian principle directly on one's own is greater than the likelihood of making a mistake by following the rule, and if these considerations held of persons generally, then one would be justified in urging its adoption as a general rule. In this way general rules might be accounted for on the summary view. It will still make sense, however, to speak of applying the utilitarian principle case by case, for it was by trying to foresee the results of doing this that one got the initial estimates upon which acceptance of the rule depends. That one is taking a rule in accordance with the summary conception will show itself in the naturalness with which one speaks of the rule as a guide, or as a maxim, or as a generalization from experience, and as something to be laid aside in extraordinary cases where there is no assurance that the generalization will hold and the case must therefore be treated on its merits. Thus there goes with this conception the notion of a particular exception which renders a rule suspect on a particular occasion.

The other conception of rules I will call the practice conception. On this view rules are pictured as defining a practice. Practices are set up for various reasons, but one of them is that in many areas of conduct each person's deciding what to do on utilitarian grounds case by case leads to confusion, and that the attempt to coordinate behavior by trying to foresee how others will act is bound to fail. As an alternative one realizes that what is required is the establishment of a practice, the specification of a new form of activity; and from this one sees that a practice necessarily involves the abdication of full liberty to act on utflitarian and prudential grounds. It is the mark of a practice that being taught how to engage in it involves being instructed in the rules which define it, and that appeal is made to those rules to correct the behavior of those engaged in it. Those engaged in a practice recognize the rules as defining it. The rules cannot be taken as simply describing how those engaged in the practice in fact behave: it is not simply that they act as if they were obeying the rules. Thus it is essential to the notion of a practice that the rules are publicly known and understood as definitive; and it is essential also that the rules of a practice can be taught and can be acted upon to yield a coherent practice. On this conception, then, rules are not generalizations from the decisions of individuals applying the utilitarian principle directly and independently to recurrent particular cases. On the contrary, rules define a practice and are themselves the subject of the utilitarian principle.

To show the important differences between this way of fitting rules into the utilitarian theory and the previous way, I shall consider the differences between the two conceptions on the points previously discussed.

1. In contrast with the summary view, the rules of practices are logically prior to particular cases. This is so because there cannot be a particular case of an action falling under a rule of a practice unless there is the practice. This can be made clearer as follows: in a practice there are rules setting up offices, specifying certain forms of action appropriate to various offices, establishing penalties for the breach of rules, and so on. We may think of the rules of a practice as defining offices, moves, and offenses. Now what is meant by saying that the practice is logically prior to particular cases is this: given any rule which specifies a form of action (a move), a particular action which would be taken as falling under this rule given that there is the practice would not be described as that sort of action unless there was the practice. In the case of actions specified by practices it is logically impossible to perform them outside the stage-setting provided by those practices, for unless there is the practice, and unless the requisite proprieties are fulfilled, whatever one does, whatever movements one makes, will fail to count as a form of action which the practice specifies. What one does will be described in some other way.

One may illustrate this point frorn the game of baseball. Many of the actions one performs in a game of baseball one can do by oneself or with others whether there is the game or not. For example, one can throw a ball, run, or swing a peculiarly shaped piece of wood. But one cannot steal base, or strike out, or draw a walk, or make an error, or balk; although one can do certain things which appear to resemble these actions such as sliding into a bag, missing a grounder and so on. Striking out, stealing a base, balking, etc., are all actions which can only happen in a game. No matter what a person did, what he did would not be described as stealing a base or striking out or drawing a walk unless he could also be described as playing baseball, and for him to be doing this presupposes the rule-like practice which constitutes the game. The practice is logically prior to particular cases: unless there is the practice the terms referring to actions specified by it lack a sense.23

2. The practice view leads to an entirely different conception of the authority which each person has to decide on the propriety of following a rule in particular cases. To engage in a practice, to perform those actions specified by a practice, means to follow the appropriate rules. If one wants to do an action which a certain practice specifies then there is no way to do it except to follow the rules which define it. Therefore, it doesn't make sense for a person to raise the question whether or not a rule of a practice correctly applies to his case where the action he contemplates is a form of action defined by a practice. If someone were to raise such a question, be would simply show that he didn't understand the situation in which he was acting. If one wants to perform an action specified by a practice, the only legitimate question concerns the nature of the practice itself ("How do I go about making a will?").

This point is illustrated by the behavior expected of a player in games. If one wants to play a game, one doesn't treat the rules of the game as guides as to what is best in particular cases. In a game of baseball if a batter were to ask "Can I have four strikes?" it would be assumed that he was asking what the rule was; and if, when told what the rule was, he were to say that he meant that on this occasion he thought it would be best on the whole for him to have four strikes rather than three, this would be most kindly taken as a joke. One might contend that baseball would be a better game if four strikes were allowed instead of three; but one cannot picture the rules as guides to what is best on the whole in particular cases, and question their applicability to particular cases as particular cases.

3 and 4. To complete the four points of comparison with the summary conception, it is clear from what has been said that rules of practices are not guides to help one decide particular cases correctly as judged by some higher ethical principal. And neither the quasi-statistical notion of generality, nor the notion of a particular exception, can apply to the rules of practices. A more or less general rule of a practice must be a rule which according to the structure of the practice applies to more or fewer of the kinds of cases arising under it; or it must be a rule which is more or less basic to the understanding of the practice. Again, a particular case cannot be an exception to a rule of a practice. An exception is rather a qualification or a further specification of the rule.

It follows from what we have said about the practice conception of rules that if a person is engaged in a practice, and if he is asked why he does what he does, or if he is asked to defend what he does, then his explanation, or defense, lies in referring the questioner to the practice. He cannot say of his action, if it is an action specified by a practice, that he does it rather than some other because he thinks it is best on the whole.24 When a man engaged in a practice is queried about his action he must assume that the questioner either doesn't know that he is engaged in it ("Why are you in a hurry to pay him?" "I promised to pay him today") or doesn't know what the practice is. One doesn't so much justify one's particular action as explain, or show, that it is in accordance with the practice. The reason for this is that it is only against the stage-setting of the practice that one's particular action is described as it is. Only by reference to the practice can one say what one is doing. To explain or to defend one's own action, as a particular action, one fits it into the practice which defines it. If this is not accepted it's a sign that a different question is being raised as to whether one is justified in accepting the practice, or in tolerating it. When the challenge is to the practice, citing the rules (saying what the practice is) is naturally to no avail. But when the challenge is to the particular action defined by the practice, there is nothing one can do but refer to the rules. Concerning particular actions there is only a question for one who isn't clear as to what the practice is, or who doesn't know that it is being engaged in. This is to be contrasted with the case of a maxim which may be taken as pointing to the correct decision on the case as decided on other grounds, and so giving a challenge on the case a sense by having it question whether these other grounds really support the decision on this case.

If one compares the two conceptions of rules I have discussed, one can see how the summary conception misses the significance of the distinction between justifying a practice and justifying actions falling under it. On this view rules are regarded as guides whose purpose it is to indicate the ideally rational decision on the given particular case which the flawless application of the utilitarian principle would yield. One has, in principle, full option to use the guides or to discard them as the situation warrants without one's moral office being altered in any way: whether one discards the rules or not, one always holds the office of a rational person seeking case by case to realize the best on the whole. But on the practice conception, if one holds an office defined by a practice then questions regarding one's actions in this office are settled by reference to the rules which define the practice. If one seeks to question these rules, then one's office undergoes a fundamental change: one then assumes the office of one empowered to change and criticize the rules, or the office of a reformer, and so on. The summary conception does away with the distinction of offices and the variour forms of argument appropriate to each. On that conception there is one office and so no offices at all. It therefore obscures the fact that the utilitarian principle must, in the case of actions and offices defined by a practice, apply to the practice, so that general utilitarian arguments are not available to those who act in offices so defined.25

Some qualifications are necessary in what I have said. First, I may have talked of the summary and the practice conceptions of rules as if only one of them could be true of rules, and if true of any rules, then, necessarily true of all rules. I do not, of course, mean this. (It is the critics of utilitarianism who make this mistake insofar as their arguments against utilitarianism presuppose a summary conception of the rules of practices.) Some rules will fit one conception, some rules the other; and so there are rules of practices (rules in the strict sense), and maxims and "rules of thumb."

Secondly, there are further distinctions that can be made in classifying rules, distinctions which should be made if one were considering other questions. The distinctions which I have drawn are those most relevant for the rather special matter I have discussed, and are not intended to be exhaustive.

Finally, there will be many borderline cases about which it will be difficult, if not impossible, to decide which conception of rules is applicable. One expects border-line cases with any concept, and they are especially likely in connection with such involved concepts as those of a practice, institution, game, rule, and so on. Wittgenstein has shown how fluid these notions are.26 What I have done is to emphasize and sharpen two conceptions for the limited purpose of this paper.

IV

What I have tried to show by distinguishing between two conceptions of rules is that there is a way of regarding rules which allows the option to consider particular cases on general utilitarian grounds; whereas there is another conception which does not admit of such discretion except insofar as the rules themselves authorize it. I want to suggest that the tendency while doing philosophy to picture rules in accordance with the summary conception is what may have blinded philosophers to the significance of the distinction between justifying a practice and justifying a particular action falling under it; and it does so by misrepresenting the logical force of the reference to the rules in the case of a challenge to a particular action falling under a practice, and by obscuring the fact that where there is a practice, it is the practice itself that must be the subject of the utilitarian principle.

It is surely no accident that two of the traditional test cases of utilitarianism, punishment and promises, are clear cases of practices. Under the influence of the summary conception it is natural to suppose that the officials of a penal system, and one who has made a promise, may decide what to do in particular cases on utilitarian grounds. One fails to see that a general discretion to decide particular cases on utilitarian grounds is incompatible with the concept of a practice; and that what discretion one does have is itself defined by the practice (e.g., a judge may have discretion to determine the penalty within certain limits). The traditional objections to utilitarianism which I have discussed presuppose the attribution to judges, and to those who have made promises, of a plenitude of moral authority to decide particular cases on utilitarian grounds. But once one fits utilitarianism together with the notion of a practice, and notes that punishment and promising are practices, then one sees that this attribution is logically precluded.

That punishment and promising are practices is beyond question. In the case of promising this is shown by the fact that the form of words "I promise" is a performative utterance which presupposes the stage-setting of the practice and the proprieties defined by it. Saying the words "I promise" will only be promising given the existence of the practice. It would be absurd to interpret the rules about promising in accordance with the summary conception. It is absurd to say, for example, that the rule that promises should be kept could have arisen from its being found in past cases to be best on the whole to keep one's promise; for unless there were already the understanding that one keeps one's promises as part of the practice itself there couldn't have been any cases of promising.

It must, of course, be granted that the rules defining promising are not codified, and that one's conception of what they are necessarily depends on one's moral training. Therefore it is likely that there is considerable variation in the way people understand the practice, and room for argument as to how it is best set up. For example, differences as to how strictly various defenses are to be taken, or just what defenses are available, are likely to arise amongst persons with different backgrounds. But irrespective of these variations it belongs to the concept of the practice of promising that the general utilitarian defense is not available to the promisor. That this is so accounts for the force of the traditional objection which I have discussed. And the point I wish to make is that when one fits the utilitarian view together with the practice conception of rules, as one must in the appropriate cases, then there is nothing in that view which entails that there must be such a defense, either in the practice of promising, or in any other practice.

Punishment is also a clear case. There are many actions in the sequence of events which constitute someone's being punished which presuppose a practice. One can see this by considering the definition of punishment which I gave when discussing Carritt's criticism of utilitarianism. The definition there stated refers to such things as the normal rights of a citizen, rules of law, due process of law, trials and courts of law, statutes, etc., none of which can exist outside the elaborate stage-setting of a legal system. It is also the case that many of the actions for which people are punished presuppose practices. For example, one is punished for stealing, for trespassing, and the like, which presuppose the institution of property. It is impossible to say what punishment is, or to describe a particular instance of it, without referring to offices, actions, and offenses specified by practices. Punishment is a move in an elaborate legal game and presupposes the complex of practices which make up the legal order. The same thing is true of the less formal sorts of punishment: a parent or guardian or someone in proper authority may punish a child, but no one else can.

There is one mistaken interpretation of what I have been saying which it is worthwhile to warn against. One might think that the use I am making of the distinction between justifying a practice and justifying the particular actions falling under it involves one in a definite social and political attitude in that it leads to a kind of conservatism. It might seem that I am saying that for each person the social practices of his society provide the standard of justification for his actions; therefore let each person abide by them and his conduct will be justified.

This interpretation is entirely wrong. The point I have been making is rather a logical point. To be sure, it has consequences in matters of ethical theory; but in itself it leads to no particular social or political attitude. It is simply that where a form of action is specified by a practice there is no justification possible of the particular action of a particular person save by reference to the practice. In such cases the action is what it is in virtue of the practice and to explain it is to refer to the practice. There is no inference whatsoever to be drawn with respect to whether or not one should accept the practices of one's society. 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.

I have tried to show that when we fit the utilitarian view together with the practice conception of rules, where this conception is appropriate,27 we can formulate it in a way which saves it from several traditional objections. I have further tried to show how the logical force of the distinction between justifying a practice and justifying an action falling under it is connected with the practice conception of rules and cannot be understood as long as one regards the rules of practices in accordance with the summary view. Why, when doing philosophy, one may be inclined to so regard them, I have not discussed. The reasons for this are evidently very deep and would require another paper.